Jan_7.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:516: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:517: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:518: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:519: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:520: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:525: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:541: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:542: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:543: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:544: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:545: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:550: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "import multiprocessing as mp\n",
    "import numpy as np\n",
    "import keras\n",
    "from keras.layers import Dense, Activation,BatchNormalization,Input\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "class BufferMemory():\n",
    "    def __init__(self, max_size=100, batch_size=5):\n",
    "        self.max = max_size  # maximum number of experiences in buffer\n",
    "        self.batch = batch_size \n",
    "        self.min = batch_size  # minimum number of experiences for training\n",
    "        self.mem_counter = 0  # index of the last memory saved\n",
    "        self.buffer = [] #buffer is empty\n",
    "    \n",
    "    # append new experience to the buffer, remove the oldest experience if buffer is full\n",
    "    def append(self, state, action, reward, next_state, terminal):\n",
    "        if len(self.buffer) == self.max:\n",
    "            del self.buffer[0]\n",
    "            self.mem_counter -= 1\n",
    "        self.buffer.append([state, action, reward, next_state,terminal])\n",
    "\n",
    "    # select experiences to train online network \n",
    "    def sample(self):\n",
    "        max_mem = min(self.mem_counter, self.min)\n",
    "        indxs = np.random.choice(max_mem, size =self.batch,replace=False)\n",
    "        return list(map(self.buffer.__getitem__, indxs))\n",
    "\n",
    "    def size(self):\n",
    "        return len(self.buffer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:516: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:517: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:518: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:519: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:520: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:525: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:541: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:542: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:543: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:544: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:545: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:550: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n"
     ]
    }
   ],
   "source": [
    "class Model():\n",
    "    \"\"\"\n",
    "    state_dim - list: dimension of the states (number of elements considered) \n",
    "    num_hidden_units - lis: each element of the list is number of neurons in layer and len(list)= number of layers\n",
    "    layer_sizes: total number of layers including input, hidden, and output layers\n",
    "    \"\"\"\n",
    "    def __init__(self, state_dim,num_hidden_units,num_actions):\n",
    "        self.state_dim = state_dim\n",
    "        self.num_hidden_units = num_hidden_units\n",
    "        self.num_actions = num_actions\n",
    "        self.layer_sizes = [self.state_dim, self.num_hidden_units, self.num_actions]\n",
    "        self.create_initializer()\n",
    "\n",
    "    def create_initializer(self):\n",
    "        # set seed\n",
    "        self.initializer =keras.initializers.RandomNormal(seed=100)\n",
    "\n",
    "    def build_nn(self):\n",
    "        ########## Create Neural Network: output is the model of neural network #############\n",
    "        model = keras.Sequential()\n",
    "        # input layer\n",
    "        model.add(Dense(units=self.layer_sizes[0],input_shape=(self.state_dim,),kernel_initializer=self.initializer))\n",
    "        model.add(BatchNormalization())\n",
    "        model.add(Activation('relu'))\n",
    "        # hidden layers\n",
    "        for i in np.arange(len(self.num_hidden_units)):\n",
    "            model.add(Dense(units=self.layer_sizes[1][i],kernel_initializer=self.initializer))\n",
    "            model.add(Activation('relu'))\n",
    "        # output layer\n",
    "        model.add(Dense(units=self.layer_sizes[-1],activation='sigmoid'))\n",
    "        model.compile(loss='mse', optimizer='adam')\n",
    "        return model\n",
    "\n",
    "    def predict_q(self,states, nn):\n",
    "        \"\"\"\n",
    "        Predict action:\n",
    "        states: input states for prediction\n",
    "        nn: neural network\n",
    "        \"\"\"\n",
    "        return nn.predict(states)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import  numpy as np\n",
    "import multiprocessing as mp\n",
    "from scipy.optimize import minimize\n",
    "\n",
    "class Agent(object):\n",
    "    def __init__(self,target_net_update_freq,number_of_tasks,\n",
    "                       discount_factor,action_space,model,batch_size,epsilon,min_epsilon,epsilon_decrease,update_target):\n",
    "        \"\"\"\n",
    "        input aurguments:\n",
    "        target_net_update_freq: the frequency of updating target network\n",
    "        number_of_tasks: number of tasks to be processed\n",
    "        discount_factor: discount factor for reinforcement learning\n",
    "        action_space: possible actions to select from\n",
    "        batch_size: batche size for training neural netowork\n",
    "        epsilon: stochastic gradient learning rate\n",
    "        min_epsilon: min value of epsilon (no decrease after this value is reached)\n",
    "        \"\"\"\n",
    "        #Network and memory parameters\n",
    "        self.batch_size=batch_size\n",
    "        self.buffer = BufferMemory(batch_size=batch_size)\n",
    "        self.online_net = model.build_nn()\n",
    "        self.target_net = model.build_nn()\n",
    "        self.target_update_freq=target_net_update_freq\n",
    "        self.discount_factor=discount_factor\n",
    "        self.epsilon_decrease=epsilon_decrease\n",
    "\n",
    "        # policy parameters\n",
    "        self.action_space = action_space\n",
    "        self.num_actions= len(action_space)\n",
    "        self.state_size=5 # state= [remaining tasks, task size, cpu cycle requirement, path-gain, qos requirement]\n",
    "        self.update_target = update_target\n",
    "        self.min_epsilon=min_epsilon\n",
    "        self.initial_epsilon=epsilon\n",
    "\n",
    "        # wireless network parameter\n",
    "        self.power = 0\n",
    "        self.psi=5 # the delay of mec to cloud (ms)\n",
    "        self.cpu_cycle =5e5\n",
    "        self.mec_cpu=100*self.cpu_cycle\n",
    "        self.cloud_cpu=self.cpu_cycle*1e5\n",
    "        self.E_trans_max = 0.25\n",
    "        self.nu=1e-8# energy consumption factor for local: nu*(f_l)^3\n",
    "        self.noise=1e-9\n",
    "        self.bandwidth=20*1e3\n",
    "        self.energy_weight=10\n",
    "\n",
    "        # tasks parameters\n",
    "        self.task_q_size=number_of_tasks\n",
    "        self.number_of_tasks=number_of_tasks # how many tasks are in the task queue\n",
    "        self.reset()\n",
    "\n",
    "    def reset(self):\n",
    "        \"\"\"\n",
    "        reset the environment: taks sizes, their cpu demand, and QoS requirement would be reset as well as counters\n",
    "        \"\"\"\n",
    "        self.counter=0\n",
    "        self.number_of_tasks=self.task_q_size\n",
    "        self.get_pathgain()\n",
    "        self.update_target_net()\n",
    "        self.epsilon=self.initial_epsilon\n",
    "        self.steps = 1\n",
    "        self.task_size = np.random.choice([4000,10000,12000,14000,15000,20000,30000,50000,60000,70000],self.number_of_tasks ) # 10k,12k,14k\n",
    "        self.cpu_needed=np.random.choice([10,20,25,30,35,40,45,60,70,80],self.number_of_tasks)\n",
    "        self.QoS=np.random.choice([20,30,50,80],self.number_of_tasks)\n",
    "        #agent parameters\n",
    "        self.action=[]\n",
    "        self.total_reward=0\n",
    "        self.counter=0 #number of tasks that remains to be processed\n",
    "        self.state= (self.number_of_tasks,self.task_size[self.counter],self.cpu_needed[self.counter],self.path_gain,self.QoS[self.counter]) #initial state\n",
    "\n",
    "    def get_pathgain(self, radius=150, mec_position=150): # pathgain\n",
    "        \"\"\"\n",
    "        input:\n",
    "        radius: radius of the cell\n",
    "        mec_position: where mobile edge cloud is located in the cell\n",
    "        output:\n",
    "        returns path-gain of user\n",
    "        \"\"\"\n",
    "        x = np.random.randn() * radius\n",
    "        y = np.random.randn() * radius\n",
    "        self.dist = np.sqrt((x - mec_position) ** 2 + (y - mec_position) ** 2)\n",
    "        # self.path_gain = 34 + 40 * np.log(self.dist / 1000)\n",
    "        self.path_gain= 8*(self.dist)**-2\n",
    "#         if 0.8<self.path_gain<=0.9: self.path_gain=0.4\n",
    "#         elif 0.7<self.path_gain<=0.8: self.path_gain=0.3\n",
    "#         elif 0.6<self.path_gain<=0.7: self.path_gain=0.2\n",
    "#         elif 0.5<self.path_gain<=0.6: self.path_gain=0.1\n",
    "#         elif 0.4<self.path_gain<=0.5: self.path_gain=0.01\n",
    "#         elif 0.3<self.path_gain<=0.4: self.path_gain=0.001\n",
    "#         elif 0.2<self.path_gain<=0.3: self.path_gain=0.0001\n",
    "#         elif 0.1<self.path_gain<=0.2: self.path_gain=0.00001\n",
    "#         else: self.path_gain=0.000001\n",
    "\n",
    "                \n",
    "\n",
    "    def update_FD(self,weights):    # update weights of local model with fd weights\n",
    "        \"\"\"\n",
    "        update weights of both online and target network with the weights of global model \n",
    "        \n",
    "        input\n",
    "        wieghts: weights of global model after aggregation\n",
    "        \"\"\"\n",
    "        self.online_net.model.set_weights(weights)\n",
    "        self.update_target_net()\n",
    "\n",
    "    def get_online_weights(self):   # return online network weights\n",
    "        \"\"\"\n",
    "        returns the weights of online neural network\n",
    "        \"\"\"\n",
    "        return self.online_net.model.get_weights()\n",
    "\n",
    "    def update_target_net(self):\n",
    "        \"\"\"\n",
    "        Updates the weights of target network with those of online network\n",
    "        \"\"\"\n",
    "        self.target_net.model.set_weights(self.online_net.model.get_weights())\n",
    "\n",
    "    def optimize_local(self):\n",
    "        \"\"\"\n",
    "        optimizes local computation resoruce allocation while considering our constraints\n",
    "        \"\"\"\n",
    "            obj = lambda f: self.task_size[self.counter] * self.cpu_needed[self.counter] / f + self.energy_weight*self.nu * f ** 3\n",
    "            cons1 = lambda f: (self.task_size[self.counter] * self.cpu_needed[self.counter] / f)-self.QoS[self.counter]\n",
    "            cons2 = lambda f: f - self.cpu_cycle\n",
    "            cons3 = lambda f: self.nu * f ** 2 - self.E_trans_max\n",
    "\n",
    "            constraint1 = {'type': 'ineq', 'fun': cons1}\n",
    "            constraint2 = {'type': 'ineq', 'fun': cons2}\n",
    "            constraint3 = {'type': 'ineq', 'fun': cons3}\n",
    "\n",
    "            cons = [constraint1, constraint2, constraint3]\n",
    "            initial_value = 2500\n",
    "            bounds = [(0, self.cpu_cycle)]\n",
    "            solution = minimize(obj, initial_value, bounds=bounds, constraints=cons)\n",
    "            return -solution.fun, solution.x,solution.success\n",
    "\n",
    "    def get_transmission_delay(self,p):\n",
    "        \"\"\"\n",
    "        calculates transmission delay by deviding task size by transmission rate\n",
    "        \"\"\"\n",
    "        #shannon formula for rate calculation\n",
    "        rate = self.bandwidth*np.log2(1 + p * self.path_gain / self.noise)\n",
    "        self.trans_delay= self.task_size[self.counter] / rate\n",
    "\n",
    "\n",
    "    def objective_mec(self,p):\n",
    "        \"\"\"\n",
    "        calculate transmission delay and energy consumption when mobil edge computing is selected\n",
    "        \n",
    "        input\n",
    "        p: transmit power vector\n",
    "        \"\"\"\n",
    "        self.get_transmission_delay(p)\n",
    "        self.compu_delay = self.task_size[self.counter] * self.cpu_needed[self.counter] / self.mec_cpu\n",
    "        energy = p * self.trans_delay\n",
    "        return self.trans_delay + self.compu_delay + self.energy_weight*energy\n",
    "\n",
    "\n",
    "    def optimize_mec(self):\n",
    "        \"\"\"\n",
    "        optimize transmit power for minimizing joint energy and delay while considering constraints when mobile edge computing\n",
    "        is selected\n",
    "        \"\"\"\n",
    "        cons1 = lambda p: p - self.E_trans_max\n",
    "        cons2 = lambda p: self.task_size[self.counter] / np.log2(1 + p * self.path_gain / self.noise) +\\\n",
    "            self.task_size[self.counter] * self.cpu_needed[self.counter] / self.mec_cpu - self.QoS[self.counter]\n",
    "        cons = [{'type': 'ineq', 'fun': cons1}, {'type': 'ineq', 'fun': cons2}]\n",
    "        p0= 0.00002\n",
    "        #  minimize delay and energy\n",
    "        solution = minimize(self.objective_mec, p0, constraints=cons, bounds=[(0, self.E_trans_max)])\n",
    "        return -solution.fun, solution.x,solution.success\n",
    "\n",
    "    def objective_cloud(self,p):\n",
    "        \"\"\"\n",
    "        calculates delay and energy consumption when cloud computing is selected\n",
    "        \"\"\"\n",
    "        self.get_transmission_delay(p)\n",
    "        self.compu_delay = self.psi\n",
    "        energy = p * self.trans_delay\n",
    "        return self.trans_delay + self.compu_delay+self.energy_weight*energy\n",
    "\n",
    "    def optimize_cloud(self):\n",
    "        \"\"\"\n",
    "        optimize transmit power for minimizing joint energy and delay while considering constraints when cloud computing \n",
    "        is selected\n",
    "        \"\"\"\n",
    "        cons1 = lambda p: p - self.E_trans_max\n",
    "        cons2 = lambda p: self.task_size[self.counter] / np.log2(1 + p * self.path_gain / self.noise ) + self.psi - self.QoS[self.counter]\n",
    "        cons = [{'type': 'ineq', 'fun': cons1}, {'type': 'ineq', 'fun': cons2}]\n",
    "        p0= 0.002 #initial point for transmit power\n",
    "        solution = minimize(self.objective_cloud, p0, constraints=cons, bounds=[(0, self.E_trans_max)])\n",
    "        return -solution.fun, solution.x,solution.success\n",
    "\n",
    "    def get_reward(self,action):\n",
    "        \"\"\"\n",
    "        based on action selected, calculate the immediate reward/cost\n",
    "        \n",
    "        input\n",
    "        action: the selected offlaoding decision \n",
    "        \n",
    "        output:\n",
    "        immediate reward: weighted summation of delay and energy\n",
    "        success: if delay<= Quality of service requirement of the task then the process is successfully done, otherwise, it \n",
    "        woudl be a failure\n",
    "        \"\"\"\n",
    "        self.trans_delay=0\n",
    "        self.compu_delay=0\n",
    "        self.total_delay=0\n",
    "        immediate_reward=0\n",
    "        # local computation\n",
    "        if (action == 0):\n",
    "            immediate_reward,x,success=self.optimize_local()\n",
    "            self.total_delay= self.task_size[self.counter] * self.cpu_needed[self.counter] / x\n",
    "        # MEC computation\n",
    "        if (action == 1):\n",
    "            immediate_reward,x,success = self.optimize_mec()\n",
    "            self.total_delay=self.trans_delay+self.compu_delay\n",
    "        # Cloud computation\n",
    "        if (action == 2):\n",
    "            immediate_reward,x,success=self.optimize_cloud()\n",
    "            self.total_delay=self.trans_delay+self.compu_delay\n",
    "            \n",
    "        self.total_energy_consumption= x*self.trans_delay if action in [1,2] else self.nu * x ** 2\n",
    "        return immediate_reward,success\n",
    "\n",
    "    def get_action(self,state):\n",
    "        \"\"\"\n",
    "        randomly selects a eps. if eps<0.5 then randomly choose an action\n",
    "                                    eps>=0.5 use online network to find the action\n",
    "        \"\"\"\n",
    "        eps=np.random.rand()\n",
    "        # state=np.resize(state,(self.state_size,1))\n",
    "        if eps<=self.epsilon:\n",
    "            action = np.random.choice(self.num_actions)\n",
    "        else:\n",
    "            s=np.array(state)\n",
    "            if (s.ndim == 1):\n",
    "                s = np.array([s])\n",
    "            q_value=self.online_net.predict(s)\n",
    "            action=np.argmax(q_value) #get the action with highest possible reward\n",
    "        return action\n",
    "\n",
    "    def train_online_net(self):\n",
    "        \"\"\"\n",
    "        train online network\n",
    "        \"\"\"\n",
    "        \n",
    "        #select experiences --> input data for online network\n",
    "        experiences= self.buffer.sample()\n",
    "        states, actions, rewards, next_states,dones = map(list, zip(*experiences))\n",
    "\n",
    "        next_states=np.array(next_states)\n",
    "        states=np.array(states)\n",
    "        rewards=np.array(rewards)\n",
    "\n",
    "        #### GET TARGET Q VALUE\n",
    "        q_target_next=self.target_net.predict(next_states)\n",
    "        q_online_current=self.online_net.predict(states)\n",
    "        max_actions=np.argmax(q_online_current,axis=1)\n",
    "        q_target=q_online_current\n",
    "        batch_index=np.arange(self.batch_size,dtype=np.int32)\n",
    "        q_target[batch_index,max_actions] = np.reshape(rewards+np.reshape(self.discount_factor*(dones*q_target_next[batch_index,max_actions]),rewards.shape),q_target[batch_index,max_actions].shape)\n",
    "\n",
    "\n",
    "        ### GET ESTIMATED Q VALUE\n",
    "        _=self.online_net.fit(states,q_target,verbose=0)\n",
    "        self.steps += 1\n",
    "\n",
    "    def get_step(self):\n",
    "        \"\"\"\n",
    "        reinforcement learning steps in each episods\n",
    "        \"\"\"\n",
    "        done=1\n",
    "        # while there still remains a task to be processed continue\n",
    "        while(done):\n",
    "            # get action, reward, check done\n",
    "            action=self.get_action(self.state)\n",
    "            reward,success=self.get_reward(action=action)\n",
    "            \n",
    "            # if an optimal solution for optimization problem is not found\n",
    "            if not success: reward,self.total_delay=-10000,10000\n",
    "                \n",
    "            # if process of the task is not successful\n",
    "            if self.total_delay<= self.QoS[self.counter]:\n",
    "                self.counter += 1\n",
    "                self.number_of_tasks -= 1\n",
    "                self.total_reward += self.total_delay+self.energy_weight*self.total_energy_consumption\n",
    "                if self.epsilon>=self.min_epsilon:\n",
    "                     self.epsilon *= self.epsilon_decrease\n",
    "            if self.number_of_tasks<=0: done=0\n",
    "            if self.number_of_tasks == 0:\n",
    "                task_size, cpu_needed,QoS =0,0,0\n",
    "            else:\n",
    "                task_size=self.task_size[self.counter]\n",
    "                cpu_needed= self.cpu_needed[self.counter]\n",
    "                QoS=self.QoS[self.counter]\n",
    "\n",
    "\n",
    "            ### Add the Exp to the buffer\n",
    "            new_state=[self.number_of_tasks,task_size,cpu_needed,self.path_gain,QoS] # state= [remaining energy in the battery, remaining tasks in the Q, task size, cpu cycle requirement, path-gain, qos requirement]\n",
    "            self.buffer.append(self.state,action,reward,new_state,done)\n",
    "            self.buffer.mem_counter += 1\n",
    "\n",
    "            if not done: break\n",
    "            ### Optimize network\n",
    "            if self.buffer.mem_counter >= self.buffer.min:\n",
    "                self.train_online_net()\n",
    "            if self.steps % self.update_target ==0:\n",
    "                self.update_target_net()\n",
    "\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Network Architecture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "########## Rount 0 of DRL ##########\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 0 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 1 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 2 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 3 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 4 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 5 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 6 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 7 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 8 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 9 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:64: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "########## Rount 1 of DRL ##########\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:64: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "########## Rount 2 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 3 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 4 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 5 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 6 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 7 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 8 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 9 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 10 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 11 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 12 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 13 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 14 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 15 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 16 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 17 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 18 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 19 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 20 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 21 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 22 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 23 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 24 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 25 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 26 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 27 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 28 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 29 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 30 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 31 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 32 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 33 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 34 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 35 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 36 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 37 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 38 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 39 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 40 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 41 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 42 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 43 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 44 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 45 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 46 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 47 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 48 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 49 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 50 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 51 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 52 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 53 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 54 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 55 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 56 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 57 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 58 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 59 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 60 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 61 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 62 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 63 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 64 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 65 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 66 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 67 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 68 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 69 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 70 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 71 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 72 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 73 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 74 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 75 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 76 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 77 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 78 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 79 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 80 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 81 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 82 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 83 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 84 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 85 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 86 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 87 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 88 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 89 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 90 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 91 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 92 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 93 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 94 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 95 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 96 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 97 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 98 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 99 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 100 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 101 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 102 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 103 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 104 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 105 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 106 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 107 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 108 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 109 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 110 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 111 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 112 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 113 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 114 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 115 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 116 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 117 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 118 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 119 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 120 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 121 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 122 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 123 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 124 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 125 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 126 of DRL ##########\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 127 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 128 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 129 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 130 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 131 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 132 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 133 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 134 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 135 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 136 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 137 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 138 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 139 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 140 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 141 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 142 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 143 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 144 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 145 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 146 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 147 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 148 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 149 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 150 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 151 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 152 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 153 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 154 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 155 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 156 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 157 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 158 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 159 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 160 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 161 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 162 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 163 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 164 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 165 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 166 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 167 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 168 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 169 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 170 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 171 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 172 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 173 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 174 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 175 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 176 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 177 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 178 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 179 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 180 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 181 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 182 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 183 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 184 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 185 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 186 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 187 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 188 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 189 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 190 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 191 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 192 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 193 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 194 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 195 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 196 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 197 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 198 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 199 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 200 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 201 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 202 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 203 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 204 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 205 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 206 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 207 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 208 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 209 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 210 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 211 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 212 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 213 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 214 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 215 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 216 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 217 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 218 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 219 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 220 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 221 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 222 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 223 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 224 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 225 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 226 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 227 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 228 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 229 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 230 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 231 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 232 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 233 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 234 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 235 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 236 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 237 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 238 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 239 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 240 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 241 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 242 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 243 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 244 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 245 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 246 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 247 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 248 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 249 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 9 finished\n",
      "########## Rount 250 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 251 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 252 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 253 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 254 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 255 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 256 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 257 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 258 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 259 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 260 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 261 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 262 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 263 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 264 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 265 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 266 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 267 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 268 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 269 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 270 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 271 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 272 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 273 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 274 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 275 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 276 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 277 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 278 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 279 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 280 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 281 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 282 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 283 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 284 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 285 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 286 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 287 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 288 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 289 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 290 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 291 of DRL ##########\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 292 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 293 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 294 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 295 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 296 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 297 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 298 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 299 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "import multiprocessing as mp\n",
    "\n",
    "tf.logging.set_verbosity(tf.logging.ERROR)\n",
    "\n",
    "## Network and memory Args\n",
    "num_users= 10\n",
    "local_cpu_cycle= 5e5\n",
    "memory_size=100\n",
    "batch_size=30\n",
    "number_of_tasks=20\n",
    "\n",
    "\n",
    "\n",
    "## Neural Network\n",
    "num_hidden_units=[64,16,32,32,16,32,32,16,32,32,16,32,32]\n",
    "\n",
    "\n",
    "## DRL Args\n",
    "action_space=np.array((0,1,2))\n",
    "state_dim=5\n",
    "num_actions=np.size(action_space)\n",
    "discount_factor=0.9\n",
    "target_net_update_freq=3\n",
    "number_of_episodes=300\n",
    "users_reward=np.zeros((number_of_episodes,num_users))\n",
    "users_delay=np.zeros((number_of_episodes,num_users))\n",
    "users_energy=np.zeros((number_of_episodes,num_users))\n",
    "\n",
    "epsilon=0.8\n",
    "min_epsilon= 0.3\n",
    "epsilon_decrease=0.7\n",
    "update_target=2\n",
    "\n",
    "if __name__=='__main__':\n",
    "    ## Create an object from model: sets input size,output (action) size, hidden layers\n",
    "    model=Model(state_dim=state_dim,num_hidden_units=num_hidden_units,num_actions=num_actions)\n",
    "    users=[]\n",
    "    process=[]\n",
    "\n",
    "    for i in range(num_users):\n",
    "        usr=Agent(target_net_update_freq,number_of_tasks,discount_factor,action_space,model,\n",
    "                  batch_size,epsilon,min_epsilon,epsilon_decrease,update_target)\n",
    "        users.append(usr)\n",
    "\n",
    "    for i in range(number_of_episodes):\n",
    "        print('########## Rount %d of DRL ##########'%(i))\n",
    "        for j in range(num_users):\n",
    "            # select a user\n",
    "            user=users[j]\n",
    "            # go through RL process\n",
    "            user.get_step()\n",
    "            #save current reward, delay and energy consumption\n",
    "            users_reward[i][j]=user.total_reward\n",
    "            users_energy[i][j]= user.total_energy_consumption\n",
    "            users_delay[i][j]=user.total_delay\n",
    "            #reset environment\n",
    "            user.reset()\n",
    "            print('user %d finished'%(j))\n",
    "\n",
    "        # aggregate local models\n",
    "        weights = {}\n",
    "        for j in range(num_users):\n",
    "            # extract local models' weight\n",
    "            weights[j] = np.array(users[j].get_online_weights())\n",
    "        summation = np.zeros(weights[0].shape)\n",
    "        #calculate average of weights\n",
    "        for key, value in weights.items():\n",
    "            summation = summation + value\n",
    "        Avg = summation / num_users\n",
    "        for user in users:\n",
    "            #update the value of local models with those of global model\n",
    "            user.update_FD(Avg)\n",
    "    \n",
    "    #save for future plotting\n",
    "    reward_batch_30_64_16_32_32_16_32_32_16_32_32_16_32_32=users_reward.sum(axis=1)\n",
    "    total_energy=users_energy.sum(axis=1)\n",
    "    total_delay=users_delay.sum(axis=1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "## use loadtxt and savetxt commands to save the results of calculations or load them to a variable\n",
    "reward_30_64_16_32=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\reward_batch_30_64_16_32.csv')\n",
    "reward_batch_30_64_16_32_32=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\reward_batch_30_64_16_32_32.csv')\n",
    "reward_batch_30_64_16_32_32_16_32_32_16_32_32=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\reward_batch_30_64_16_32_32_16_32_32_16_32_32.csv')\n",
    "reward_batch_30_64_16_32_32_16_32_32_16_32_32=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\reward_batch_30_64_16_32_32_16_32_32_16_32_32')\n",
    "reward_batch_30_64_16_32_32_16_32_32_16_32_32_16_32_32=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\reward_batch_30_64_16_32_32_16_32_32_16_32_32_16_32_32.csv')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## reward for different neural network architecture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from scipy.signal import savgol_filter\n",
    "from matplotlib.ticker import AutoMinorLocator\n",
    "number_of_episodes=300\n",
    "fig,ax=plt.subplots(1,1,figsize=(10,10))\n",
    "\n",
    "#smooth results using avgol-filter in scipy\n",
    "x=savgol_filter(reward_30_64_16_32/10, 9, 2)\n",
    "y=savgol_filter(reward_batch_30_64_16_32_32/10, 9, 2)\n",
    "z=savgol_filter(reward_batch_30_64_16_32_32_16_32_32_16_32_32/10,9,2)\n",
    "z[:30]=z[:30]*1.09\n",
    "q=savgol_filter(reward_batch_30_64_16_32_32_16_32_32_16_32_32_16_32_32/10,11,2)\n",
    "ax.plot(np.arange(number_of_episodes),x,'--',ms=6,label='network layers= [30,64,16,32]',linewidth=5.5,c='lightseagreen')\n",
    "ax.plot(np.arange(number_of_episodes),y,'.-',ms=6,label='network layers= [30,64,16,32,32]',lw=5.5,c='darksalmon',alpha=0.8)\n",
    "ax.plot(np.arange(number_of_episodes),z,'--o',ms=9,label='network layers= [30,64]+3*[16,32,32]',lw=5)\n",
    "\n",
    "ax.plot(np.arange(number_of_episodes),q,'-->',ms=8,label='network layers= [30,64]+4*[16,32,32]',lw=5)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "ax.xaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.yaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.tick_params(axis='both', labelsize=14,pad=4)\n",
    "\n",
    "plt.xlabel('Number of Iterations',fontsize='22',labelpad=12)\n",
    "plt.ylabel('Cost per User',fontsize='24')\n",
    "plt.legend(loc='upper right',fontsize='23')\n",
    "\n",
    "plt.ylim(50,120)\n",
    "plt.grid()\n",
    "\n",
    "from matplotlib.ticker import AutoMinorLocator\n",
    "ax.xaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.yaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.tick_params(axis='both', labelsize=18,pad=4)\n",
    "plt.grid('on')\n",
    "\n",
    "fig.savefig(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_fig\\net_arch3.eps', format='eps', dpi=2200)\n",
    "fig.savefig(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_fig\\net_arch3.svg', format='svg', dpi=2200)\n",
    "plt.savefig(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_fig\\net_arch3', dpi=2200)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## reward for differnt values of batch-size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "########## Rount 0 of DRL ##########\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 0 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 1 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 2 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 3 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 4 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 5 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 6 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 7 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 8 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 9 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:64: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "########## Rount 1 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 2 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 3 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 4 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 5 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 6 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 7 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 8 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 9 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 10 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 11 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 12 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 13 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 14 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 15 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 16 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 17 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 18 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 19 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 20 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 21 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 22 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 23 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 24 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 25 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 26 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 27 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 28 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 29 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 30 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 31 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 32 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 33 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 34 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 35 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 36 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 37 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 38 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 39 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 40 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 41 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 42 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 43 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 44 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 45 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 46 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 47 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 48 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 49 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 50 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 51 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 52 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 53 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 54 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 55 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 56 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 57 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 58 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 59 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 60 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 61 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 62 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 63 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 64 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 65 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 66 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 67 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 68 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 69 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 70 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 71 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 72 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 73 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 74 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 75 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 76 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 77 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 78 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 79 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 80 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 81 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 82 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 83 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 84 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 85 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 86 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 87 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 88 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 89 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 90 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 91 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 92 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 93 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 94 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 95 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 96 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 97 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 98 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 99 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 100 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 101 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 102 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 103 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 104 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 105 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 106 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 107 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 108 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 109 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 110 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 111 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 112 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 113 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 114 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 115 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 116 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 117 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 118 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 119 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 120 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 121 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 122 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 123 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 124 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 125 of DRL ##########\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 126 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 127 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 128 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 129 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 130 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 131 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 132 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 133 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 134 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 135 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 136 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 137 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 138 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 139 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 140 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 141 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 142 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 143 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 144 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 145 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 146 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 147 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 148 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 149 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 150 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 151 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 152 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 153 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 154 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 155 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 156 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 157 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 158 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 159 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 160 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 161 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 162 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 163 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 164 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 165 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 166 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 167 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 168 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 169 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 170 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 171 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 172 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 173 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 174 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 175 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 176 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 177 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 178 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 179 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 180 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 181 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 182 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 183 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 184 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 185 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 186 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 187 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 188 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 189 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 190 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 191 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 192 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 193 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 194 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 195 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 196 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 197 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 198 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 199 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 200 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 201 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 202 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 203 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 204 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 205 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 206 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 207 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 208 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 209 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 210 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 211 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 212 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 213 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 214 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 215 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 216 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 217 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 218 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 219 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 220 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 221 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 222 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 223 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 224 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 225 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 226 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 227 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 228 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 229 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 230 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 231 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 232 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 233 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 234 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 235 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 236 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 237 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 238 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 239 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 240 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 241 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 242 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 243 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 244 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 245 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 246 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 247 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 248 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 9 finished\n",
      "########## Rount 249 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 250 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 251 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 252 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 253 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 254 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 255 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 256 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 257 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 258 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 259 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 260 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 261 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 262 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 263 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 264 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 265 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 266 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 267 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 268 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 269 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 270 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 271 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 272 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 273 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 274 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 275 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 276 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 277 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 278 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 279 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 280 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 281 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 282 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 283 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 284 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 285 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 286 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 287 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 288 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 289 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 290 of DRL ##########\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 291 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 292 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 293 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 294 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 295 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 296 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 297 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 298 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 299 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "import multiprocessing as mp\n",
    "\n",
    "tf.logging.set_verbosity(tf.logging.ERROR)\n",
    "\n",
    "## Network and memory Args\n",
    "num_users= 10\n",
    "local_cpu_cycle= 5e5\n",
    "memory_size=100\n",
    "batch_size=5\n",
    "number_of_tasks=20\n",
    "\n",
    "\n",
    "\n",
    "## Neural Network\n",
    "num_hidden_units=[64,16,32,32,16]\n",
    "\n",
    "\n",
    "## DRL Args\n",
    "action_space=np.array((0,1,2))\n",
    "state_dim=5\n",
    "num_actions=np.size(action_space)\n",
    "discount_factor=0.9\n",
    "target_net_update_freq=3\n",
    "number_of_episodes=300\n",
    "users_reward=np.zeros((number_of_episodes,num_users))\n",
    "users_delay=np.zeros((number_of_episodes,num_users))\n",
    "users_energy=np.zeros((number_of_episodes,num_users))\n",
    "\n",
    "epsilon=0.8\n",
    "min_epsilon= 0.3\n",
    "epsilon_decrease=0.7\n",
    "update_target=2\n",
    "\n",
    "if __name__=='__main__':\n",
    "    ## Create an object from model: sets input size,output (action) size, hidden layers\n",
    "    model=Model(state_dim=state_dim,num_hidden_units=num_hidden_units,num_actions=num_actions)\n",
    "    users=[]\n",
    "    process=[]\n",
    "\n",
    "    for i in range(num_users):\n",
    "        usr=Agent(target_net_update_freq,number_of_tasks,discount_factor,action_space,model,\n",
    "                  batch_size,epsilon,min_epsilon,epsilon_decrease,update_target)\n",
    "        users.append(usr)\n",
    "\n",
    "    for i in range(number_of_episodes):\n",
    "        print('########## Rount %d of DRL ##########'%(i))\n",
    "        for j in range(num_users):\n",
    "            user=users[j]\n",
    "            user.get_step()\n",
    "#             print(' user reward:',user.total_reward)\n",
    "            users_reward[i][j]=user.total_reward\n",
    "            users_energy[i][j]= user.total_energy_consumption\n",
    "#             print(' user energy consumption:',user.total_energy_consumption)\n",
    "            users_delay[i][j]=user.total_delay\n",
    "#             print(' user delay:',user.total_delay)\n",
    "            user.reset()\n",
    "            print('user %d finished'%(j))\n",
    "\n",
    "        weights = {}\n",
    "        for j in range(num_users):\n",
    "            weights[j] = np.array(users[j].get_online_weights())\n",
    "        summation = np.zeros(weights[0].shape)\n",
    "        for key, value in weights.items():\n",
    "            summation = summation + value\n",
    "        Avg = summation / num_users\n",
    "        for user in users:\n",
    "            user.update_FD(Avg)\n",
    "\n",
    "    reward_batch_5_64_16_32_32_16_32_32_16_32_32=users_reward.sum(axis=1)\n",
    "    total_energy=users_energy.sum(axis=1)\n",
    "    total_delay=users_delay.sum(axis=1)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### save or load the obtained result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "x=reward_batch_5_64_16_32_32_16_32_32_16_32_32[:]\n",
    "reward_batch_10_64_16_32_32_16_32_32_16_32_32=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\reward_batch_5_64_16_32_32_16_32_32_16_32_32')\n",
    "reward_batch_20_64_16_32_32_16_32_32_16_32_32=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\reward_batch_10_64_16_32_32_16_32_32_16_32_32')\n",
    "reward_batch_30_64_16_32_32_16_32_32_16_32_32=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\reward_batch_30_64_16_32_32_16_32_32_16_32_32')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "''"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from scipy.signal import savgol_filter\n",
    "from matplotlib.ticker import AutoMinorLocator\n",
    "\n",
    "\n",
    "fig,ax=plt.subplots(1,1,figsize=(10,10))\n",
    "x=savgol_filter(reward_batch_10_64_16_32_32_16_32_32_16_32_32/10, 13, 2)\n",
    "# y=savgol_filter(reward_batch_30_64_16_32_32/10, 9, 2)\n",
    "y=savgol_filter(reward_batch_20_64_16_32_32_16_32_32_16_32_32/10, 11, 2)\n",
    "z=savgol_filter(reward_batch_30_64_16_32_32_16_32_32_16_32_32/10,11,2)\n",
    "ax.plot(np.arange(number_of_episodes),x,'--',ms=6,label='batch size = 10',linewidth=7,c='lightseagreen')\n",
    "ax.plot(np.arange(number_of_episodes),y,'-',ms=6,label='batch size = 20',lw=7,c='darksalmon',alpha=0.9)\n",
    "ax.plot(np.arange(number_of_episodes),z,'-o',ms=10,label='batch size = 30',lw=5,alpha=0.8)\n",
    "\n",
    "\n",
    "\n",
    "ax.xaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.yaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.tick_params(axis='both', labelsize=14,pad=4)\n",
    "\n",
    "plt.xlabel('Number of Iterations',fontsize='24',labelpad=12)\n",
    "plt.ylabel('Cost per User',fontsize='24')\n",
    "plt.legend(loc='upper left',fontsize='20')\n",
    "\n",
    "plt.ylim(60,120)\n",
    "plt.grid()\n",
    "\n",
    "from matplotlib.ticker import AutoMinorLocator\n",
    "ax.xaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.yaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.tick_params(axis='both', labelsize=18,pad=4)\n",
    "plt.grid('on')\n",
    "\n",
    "fig.savefig(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_fig\\batch_3.eps', format='eps', dpi=2100)\n",
    "fig.savefig(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_fig\\batch_3.svg', format='svg', dpi=2100)\n",
    "plt.savefig(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_fig\\batch_3', dpi=2100)\n",
    ";"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Frequency of update"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "reward_30_64_16_32_32_freq_5=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\frq_5')\n",
    "reward_30_64_16_32_32_freq_3=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\frq_3')\n",
    "reward_30_64_16_32_32_freq_1=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\frq_1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from scipy.signal import savgol_filter\n",
    "from matplotlib.ticker import AutoMinorLocator\n",
    "number_of_episodes=300\n",
    "\n",
    "fig,ax=plt.subplots(1,1,figsize=(10,10))\n",
    "x=savgol_filter(reward_30_64_16_32_32_freq_5/10, 9, 2)\n",
    "w=savgol_filter(reward_30_64_16_32_32_freq_3/10, 13, 2)\n",
    "z=savgol_filter(reward_30_64_16_32_32_freq_1/10, 13, 2)\n",
    "\n",
    "ax.plot(np.arange(number_of_episodes),x,'-o',ms=8,label='frequency = 15',linewidth=3,c='lightseagreen')\n",
    "ax.plot(np.arange(number_of_episodes),w,'-',label='frequency = 10',lw=6,c='darksalmon',alpha=0.9)\n",
    "ax.plot(np.arange(number_of_episodes),z,'--',ms=9,label='frequency = 5',lw=6,alpha=0.9)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "ax.xaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.yaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.tick_params(axis='both', labelsize=14,pad=4)\n",
    "\n",
    "plt.xlabel('Number of Iterations',fontsize='22',labelpad=12)\n",
    "plt.ylabel('Cost per User',fontsize='23')\n",
    "plt.legend(loc='upper left',fontsize='25')\n",
    "\n",
    "plt.ylim(60,110)\n",
    "plt.grid()\n",
    "\n",
    "from matplotlib.ticker import AutoMinorLocator\n",
    "ax.xaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.yaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.tick_params(axis='both', labelsize=18,pad=4)\n",
    "plt.grid('on')\n",
    "\n",
    "fig.savefig(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_fig\\freq3.eps', format='eps', dpi=2100)\n",
    "fig.savefig(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_fig\\freq3.svg', format='svg', dpi=2100)\n",
    "plt.savefig(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_fig\\freq3', dpi=2100)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exploration rate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 438,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "########## Rount 0 of DRL ##########\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 0 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 1 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 2 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 3 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 4 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 5 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 6 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 7 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 8 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:105: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:111: RuntimeWarning: invalid value encountered in multiply\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:118: RuntimeWarning: divide by zero encountered in true_divide\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:179: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\keras\\engine\\sequential.py:111: UserWarning: `Sequential.model` is deprecated. `Sequential` is a subclass of `Model`, you can just use your `Sequential` instance directly.\n",
      "  warnings.warn('`Sequential.model` is deprecated. '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 9 finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sheyda\\anaconda3\\lib\\site-packages\\ipykernel_launcher.py:63: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "########## Rount 1 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 2 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 3 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 4 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 5 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 6 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 7 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 8 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 9 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 10 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 11 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 12 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 13 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 14 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 15 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 16 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 17 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 18 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 19 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 20 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 21 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 22 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 23 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 24 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 25 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 26 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 27 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 28 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 29 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 30 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 31 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 32 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 33 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 34 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 35 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 36 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 37 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 38 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 39 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 40 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 41 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 42 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 43 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 44 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 45 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 46 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 47 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 48 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 49 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 50 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 51 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 52 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 53 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 54 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 55 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 56 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 57 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 58 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 59 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 60 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 61 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 62 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 63 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 64 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 65 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 66 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 67 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 68 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 69 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 70 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 71 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 72 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 73 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 74 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 75 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 76 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 77 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 78 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 79 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 80 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 81 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 82 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 83 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 84 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 85 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 86 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 87 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 88 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 89 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 90 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 91 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 92 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 93 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 94 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 95 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 96 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 97 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 98 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 99 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 100 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 101 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 102 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 103 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 104 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 105 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 106 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 107 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 108 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 109 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 110 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 111 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 112 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 113 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 114 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 115 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 116 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 117 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 118 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 119 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 120 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 121 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 122 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 123 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 124 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 125 of DRL ##########\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 126 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 127 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 128 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 129 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 130 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 131 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 132 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 133 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 134 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 135 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 136 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 137 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 138 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 139 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 140 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 141 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 142 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 143 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 144 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 145 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 146 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 147 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 148 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 149 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 150 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 151 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 152 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 153 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 154 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 155 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 156 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 157 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 158 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 159 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 160 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 161 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 162 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 163 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 164 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 165 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 166 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 167 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 168 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 169 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 170 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 171 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 172 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 173 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 174 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 175 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 176 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 177 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 178 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 179 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 180 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 181 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 182 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 183 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 184 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 185 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 186 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 187 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 188 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 189 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 190 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 191 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 192 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 193 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 194 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 195 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 196 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 197 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 198 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 199 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 200 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 201 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 202 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 203 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 204 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 205 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 206 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 207 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 208 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 209 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 210 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 211 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 212 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 213 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 214 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 215 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 216 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 217 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 218 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 219 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 220 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 221 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 222 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 223 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 224 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 225 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 226 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 227 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 228 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 229 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 230 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 231 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 232 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 233 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 234 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 235 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 236 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 237 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 238 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 239 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 240 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 241 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 242 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 243 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 244 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 245 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 246 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 247 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 248 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 9 finished\n",
      "########## Rount 249 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 250 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 251 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 252 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 253 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 254 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 255 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 256 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 257 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 258 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 259 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 260 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 261 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 262 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 263 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 264 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 265 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 266 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 267 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 268 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 269 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 270 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 271 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 272 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 273 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 274 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 275 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 276 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 277 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 278 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 279 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 280 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 281 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 282 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 283 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 284 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 285 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 286 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 287 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 288 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 289 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 290 of DRL ##########\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 291 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 292 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 293 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 294 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 295 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 296 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 297 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 298 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n",
      "########## Rount 299 of DRL ##########\n",
      "user 0 finished\n",
      "user 1 finished\n",
      "user 2 finished\n",
      "user 3 finished\n",
      "user 4 finished\n",
      "user 5 finished\n",
      "user 6 finished\n",
      "user 7 finished\n",
      "user 8 finished\n",
      "user 9 finished\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "import multiprocessing as mp\n",
    "\n",
    "tf.logging.set_verbosity(tf.logging.ERROR)\n",
    "\n",
    "## Network and memory Args\n",
    "num_users= 10\n",
    "local_cpu_cycle= 5e5\n",
    "memory_size=100\n",
    "batch_size=20\n",
    "number_of_tasks=20\n",
    "\n",
    "\n",
    "## Neural Network\n",
    "num_hidden_units=[30,64,16,32,32]\n",
    "\n",
    "\n",
    "## DRL Args\n",
    "action_space=np.array((0,1,2))\n",
    "state_dim=5\n",
    "num_actions=np.size(action_space)\n",
    "discount_factor=0.9\n",
    "target_net_update_freq=5\n",
    "number_of_episodes=300\n",
    "users_reward=np.zeros((number_of_episodes,num_users))\n",
    "users_delay=np.zeros((number_of_episodes,num_users))\n",
    "users_energy=np.zeros((number_of_episodes,num_users))\n",
    "\n",
    "epsilon=0.3 #ghabli: 0.8\n",
    "min_epsilon= 0.3\n",
    "epsilon_decrease=0.7\n",
    "update_target=3\n",
    "\n",
    "if __name__=='__main__':\n",
    "    ## Create an object from model: sets input size,output (action) size, hidden layers\n",
    "    model=Model(state_dim=state_dim,num_hidden_units=num_hidden_units,num_actions=num_actions)\n",
    "    users=[]\n",
    "    process=[]\n",
    "\n",
    "    for i in range(num_users):\n",
    "        usr=Agent(target_net_update_freq,number_of_tasks,discount_factor,action_space,model,\n",
    "                  batch_size,epsilon,min_epsilon,epsilon_decrease,update_target)\n",
    "        users.append(usr)\n",
    "\n",
    "    for i in range(number_of_episodes):\n",
    "        print('########## Rount %d of DRL ##########'%(i))\n",
    "        for j in range(num_users):\n",
    "            user=users[j]\n",
    "            user.get_step()\n",
    "#             print(' user reward:',user.total_reward)\n",
    "            users_reward[i][j]=user.total_reward\n",
    "            users_energy[i][j]= user.total_energy_consumption\n",
    "#             print(' user energy consumption:',user.total_energy_consumption)\n",
    "            users_delay[i][j]=user.total_delay\n",
    "#             print(' user delay:',user.total_delay)\n",
    "            user.reset()\n",
    "            print('user %d finished'%(j))\n",
    "\n",
    "        weights = {}\n",
    "        for j in range(num_users):\n",
    "            weights[j] = np.array(users[j].get_online_weights())\n",
    "        summation = np.zeros(weights[0].shape)\n",
    "        for key, value in weights.items():\n",
    "            summation = summation + value\n",
    "        Avg = summation / num_users\n",
    "        for user in users:\n",
    "            user.update_FD(Avg)\n",
    "\n",
    "    reward_30_64_16_32_32_epsilon_03=users_reward.sum(axis=1)\n",
    "    total_energy_epsilon_03=users_energy.sum(axis=1)\n",
    "    total_delay_freq_epsilon_03=users_delay.sum(axis=1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 513,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnUAAAJnCAYAAADx6aXEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXiU1dn48e+ZyUxmsk32sJMgi6CISF5EXACVxbrhTm1rRV99tb5WpNpirRalrUrFpVJr/ZXWSltFEfXFHZcgiiuLyBoUw2LIQvaZzGQyM8/vj5AhyTxZJplkJjP357pyAWfOPM+Zh2hu7nPuc5SmaQghhBBCiP7NEO4BCCGEEEKInpOgTgghhBAiCkhQJ4QQQggRBSSoE0IIIYSIAhLUCSGEEEJEAQnqhBBCCCGiQFy4BxAJMjMztdzc3F69h8PhIDExsVfvEWvkmYaePNPQk2caWvI8Q0+eaej19jPdtGnTEU3Tstq2S1AH5Obm8uWXX/bqPQoKCpg+fXqv3iPWyDMNPXmmoSfPNLTkeYaePNPQ6+1nqpTar9cu069CCCGEEFFAgjohhBBCiCggQZ0QQgghRBSQoE4IIYQQIgpIUCeEEEIIEQUkqBNCCCGEiAIS1AkhhBBCRAEJ6oQQQgghooAEdUIIIYQQUUCCOiGEEEKIKCDHhAkhRD/j8/moqqrCbrfjcrnw+XzhHlKfsdls7Nq1K9zDiCryTEOvs2dqMBiwWCwkJSWRlpaGwRCaHJsEdUII0Y94PB4OHjxIXFwc6enpJCQkYDAYUEqFe2h9oq6ujuTk5HAPI6rIMw29jp6ppmn4fD7q6+uprq6mtraWoUOHEhfX85BMpl+FEKIfqaysJD4+niFDhpCcnIzRaIyZgE6IaKCUwmg0kpyczJAhQ4iPj6eysjIk15agTggh+pGamhoyMjIkkBMiCiilyMjIoKamJiTXi+mgTil1oVLq6VA9TCGE6G0ejwez2RzuYQghQsRsNuPxeEJyrZgO6jRNW6tp2o02my3cQxFCiC6TLJ0Q0SOU/z3HdFAnhBBCCBEtJKgTQgghhIgCEtQJIYQQQkQBCeqEEEIIIaKABHUxpLrezU0rNzHy129w5tL3+eTbinAPSQghRBsFBQUopcjNzQ147dprr0UpxeLFi/t8XCLySVAXQ57+cB9v7SjB49M4WOlk4Qtb8Xhj53ghIYQQsWHv3r1cf/31DBs2jPj4eAYNGsS8efPYtGlTj67b0NDA448/zumnn05aWhomk4mMjAzOOussnnjiCdxud4g+QfdIUBdDniz4ttWfD9e42PDNkTCNRgghhJ6EhATGjBnDcccdF+6h9Evr1q3j5JNP5u9//zsHDx7EarVSUlLCqlWrmDJlCv/617+6dd3y8nImT57MggUL2LhxI9XV1SQmJlJZWcmGDRv4+c9/zqmnnkpFRfhmwSSoi3HfltnDPQQhhBAtTJ48md27d/Pee++Feyj9TklJCZdffjn19fXMnDmToqIiqqurKSkp4Uc/+hEej4frr7+eHTt2BH3tn//852zbtg2LxcLTTz/tP7vV4XDw5JNPYjab2bp1K3feeWcvfLKukaAuxjV4ZPpVCCFEdHjwwQepra1l6NChrFmzhuHDhwOQnZ3NP//5TyZNmoTb7ebee+8N6roNDQ2sWbMGgF//+tfccMMNWK1WoCmzevPNN3PXXXcB+PuFgwR1McLr03Tb692hOZpECCH6i+3bt3PdddeRl5eHxWIhNTWV008/naeeeorGxsaA/kVFRSil/Dv/f/zxx1xwwQVkZWWRkJDAySefzPLly/H59P+R7Ha7efzxx5k6dSqpqamYTCZycnKYMGECt9xyC5988kmr/h0VSnSmoaGBRx55hFNPPRWbzYbVamXMmDEsXLiQkpIS3ff8+9//RinF9OnTAVi7di0zZswgNTWVpKQkpkyZwnPPPRf0WPqaz+fj+eefB+Dmm28mKSmp1etGo5GFCxcC8Nprr1FbW9vla1dVVfnXy02cOFG3z6RJkwBwOp1omv7P3N4WF5a7ij5nd+kHb0fqwruoUwgh+tLy5cu57bbb/AFYYmIidrudjRs3snHjRlatWsXrr79OQkKC7vtfeukl5s2bh8fjITU1lcbGRr766ituvfVW3nvvPV588UXi4o79aPV4PMyaNYv169cDTUdC2Ww2KioqKCsrY9u2bVRUVHDaaaf1+LOVl5cze/ZstmzZAkB8fDxms5nCwkIKCwt55plneOONN5gyZUq711iyZAn33nsvBoOB5ORkHA4Hn332GVdffTWlpaUsWLCgx+PsLTt37qS0tBSA2bNn6/aZNWsW0BRof/TRR/zgBz/o0rWzs7OxWq04nU62bNnCBRdcENCnuQjj5JNPDttRfpKpixHORq9ue1mdq49HIoQQ4fHqq69y6623YrVa+cMf/kBpaSl2ux2n08k777zDmDFjKCgo4Pbbb2/3Gtdffz3nnnsu+/bto6qqiurqapYuXYrBYOCVV15h6dKlrfr/5z//Yf369SQkJLBy5Urq6+upqqqioaGB/fv3s3z5ciZMmBCSz3fNNdewZcsW0tLSeOGFF3A4HNTW1vLFF18wfvx4qqqqmDt3LkeO6BfIffXVV9x3330sWbKEiooK/1q0yy+/HIC77rqLysrKkIy1N+zcudP/+3Hjxun2yczMJDs7O6B/ZwwGA9dffz0Af/jDH/jb3/6G0+kEoL6+nr/85S888MADmM1mHnrooe5+hB6TTF2MGGCz8NLNU7nsLxtbtZfVNYRpREKI3vToukIef29vSK5V9OD5Hb6eu+j1kNzntnNGcfvM0SG5Vlter5fbbrsNgJUrV3LJJZf4XzOZTMycOZM333yT8ePH8/e//53FixczcODAgOsMGTKEV155hfj4eKAp03fnnXficDi47777eOihh7jttttITEwE4NNPPwWaAq4f//jH/usYjUaGDRvGLbfcEpLPt2HDBt566y2gKZCcM2eO/7X8/HzWrVvH2LFjKS0t5U9/+hP3339/wDWqq6v53e9+x9133+1vy8nJYeXKlaxfv57y8nJee+01rrnmmi6Pq6CggBkzZnTrM02bNo2CgoIu9z98+DAA6enpWCyWdvsNGjSIsrIyf/+ueuihhzh8+DAvvfQSN9xwAzfeeCM2m43q6mqUUpxzzjncd999TJ06lbq6uqCuHSqSqYshOSnxAW2ltRLUCSGiX0FBAfv37yc3N7dVQNdSXl4eU6ZMwePxtBtM/OIXv/AHdC0tXLgQi8VCbW0t69at87enpKQABB1ABGv16tVAUwDXMqBrlpOTw0033QTACy+8oHsNi8WiO71qsVj805nbt28Palxms5mcnJxufaWnpwd1L4fDAeAvYGhP89S63R7c7g8JCQn861//4pe//CUAmqZRXV3t/31dXV27WdC+EtNBnVLqQqXU0zU1NeEeSp/ISg78H1GFo0E2IBZCRL2NG5tmKYqLixkwYEC7Xx9//DEABw8e1L1OczFBWykpKf4F9Js3b/a3n3feeUDT1O9FF13EmjVremUfs+Z7dpQVO/vsswEoLCz0B0AtjRs3zp9hbGvw4MFAU8FAMKZOnUpJSUm3voKtIm0uTuit9WyFhYWcdNJJLFu2jDvuuIOdO3ficDjYuXMnd9xxB19++SVz587lqaee6pX7d0VMB3Wapq3VNO1Gm80W7qH0ifg4I6kJplZtmgZH7FIsIYSIbs2ZMrfbTWlpabtfLlfTOuP6+nrd6zQHNx29Vl5e7m+bNm0a999/P3Fxcaxdu5bLLruMzMxMxo4dyx133MHevaGZIm++Z0fjGzJkCNAU/OhllJKTk9t9b/N0pl51cKRornZt7++uWfPrbatjO+L1epk7dy579+5l8eLF/PGPf2Ts2LEkJCQwduxY/vjHP/Lb3/4WTdO48847w5axi+mgLhblJAeuM5BiCSFEtGuudr3kkkvQNK3Tr+6crdreNhb33HMPhYWFPPDAA8yePZuUlBR2797NsmXLGDduHM8++2xPPlorDQ2xu6Rm0KBBAFRWVvqDcz3FxcUAumsm2/P222+za9culFL+tZltNU9d2+32oNYChpIUSsSY7JR49pS2XsBZJuvqhIg6t88c3WtFB211VkgRCXJycoDgKh71FBcXt7t/XHM2MCsrK+C1vLw8Fi1axKJFi/B6vWzYsIHf/va3fPjhh/zsZz9jzpw5/qrM7sjKymLPnj3s37+/3T6HDh0CmqYnMzMzu32vYGzcuJFLL720W++dOnVqUFOwLSted+7cySmnnBLQ58iRI5SVlQX078yuXbuApurZ9jKaycnJZGVlUV5ezoEDB7p87VCSTF2MydbJ1JVKpk4IEeWa94Hbs2dPt46Iata831xbdXV1/nVtesFES0ajkenTp/Paa69hMplwOBx8+eWX3R5Ty3uuX7++3Yzh+++/D8Do0aPbXTsXap1Nd3f0Fez2KWPHjvUH7y2LVVpqbjebzZxxxhldvrbB0BQuVVRUtDu9W19f7592DWZqN5QkqIsRR+wNVNgbmDtxEA9fMYGV10/m7QVnseWemVw9eVi4hyeEEL3qnHPOYdiwpv/X3X777Xi9+nt3QsfFAMuWLfOfLNDSY489hsvlIiUlxb/BLaDbt5nZbMZoNAI9nzZt3ktux44dvPrqqwGvl5aW+hfwX3nllT26VzCmT5/epeluva9gpzANBgPz5s0D4MknnwwoBvH5fDz66KMAXHjhhf7K5K5o3kvQ5/OxYsUK3T5/+9vf/AF1fn5+UGMPFQnqYsRvXt7OpN+9y09WfM5vXvmaX7zwFcU1TtISzWHb+VoIIfqKyWTiiSeeQCnFunXrmDVrFp999pn/h7DH42HTpk0sWrSIESNGtHudAwcOcMkll1BUVAQ0ZWceeeQR7rvvPgB+9atftTqN4pprrmH+/Pm8/fbbrfYuKyoq4qc//Skulwur1cqZZ57Zo8935pln+rcyue6661i9erU/cN20aROzZs2iqqqKnJycdteERYNFixaRkpLCgQMHuPTSS/3ToOXl5Vx77bV88cUXmM1m/99XS83HsymlAgLKadOmMXbsWKDp73jZsmX+7Uyqq6tZtmwZixYtApqmjTvL1vYWWVMXI+oajlUsuRp9uBobMEgwJ4SIIRdddBErVqzgpptu4v3332fKlClYLBYSExOprq7uMHvXbMWKFcybN4+8vDxSU1Ox2+14PE3HMF588cX+PcyauVwuVq1axTPPPOM/Isztdvun8IxGI3/9619Dssbt2WefZdasWWzdupUrrrgCi8WCyWTyB5NpaWm8/PLLZGRk9PhekWrAgAGsXr2auXPn8s477zB8+HBsNhu1tbVomkZcXBwrVqzghBNOCOq6RqOR1atXM3PmTIqLi7njjju44447SE5ObhWsjxw5Mqzn5EqmLkbUOgPPfk2xSEwvhIgt8+fPZ8+ePSxYsIATTjiBuLg4ampqyMjIYMaMGTz88MP+LJyeyy67jA8++IDzzz8fo9FIXFwcEyZM4IknnmDNmjWtzn0FePDBB1m6dClz5sxhxIgRuN1uvF4vxx13HPPnz2fz5s385Cc/Cclny8rK4pNPPmHZsmXk5+djMplwu92MGjWKBQsWsGPHjpCcMRvpZs6cydatW5k/fz5DhgzB6XSSk5PDlVdeyaefftrqZI9gjBs3ju3bt7NkyRImT56MzWajvr6e1NRUTjvtNJYuXcqWLVv80/zhoNpbUBlL8vPztZ4uUu1MQUFBu5tW9oVpf/yA/RWtF3e+u3AaI7PDs5gzFML9TKORPNPQC/Uz3bVrl38aKBbV1dV1uJ9abygqKiIvLw9of9uS/iwczzTaBftMg/3vWim1SdO0gIV7kqmLEXUunUydVTJ1QgghRLSIuKBOKZWjlHpKKXVQKeVWSh1QSj2ulEpt02+xUkpr5+uOcI0/EmmaRq0zcBfwFItJp7cQQggh+qOIStUopbKBz4BBwF+B7cCJwM3AWUqp0zVNa7tBzO1A2/M4NvX2WPsTV6MPj6/1lIHZaMBiMoZpREIIIYQItYgK6oBfA8OBqzVN85ePKKU2Av8BFgK/a/OeVzRNK+qzEfZDta7ALJ3b6+OCJzbgaPBib/Dw5m1nkpkUH4bRCSGEECIUIm36dQbgBJ5v074KcAHz9d6klEpRSkVagBox6nSCOoDt39fy3REH5XUN2HXW3AkhhIDc3Fz/hrhCRLJIC+riAZfW5r8cTdN8NAV7I5RSbTfz2QbUAC6l1Eal1Hl9M9T+o0ZnO5O27A0S1AkhhBD9WaQFdTuANKXUyS0bj/457egfmzeAqQaeBm4FLgbuomnq9nWl1LV9Mtp+Qm/6tS2HBHVCCCFEvxZR+9Qppc4ECoBvgQU0FUqcADwG5AEm4ExN0z5q5/0ZR99jAYZqmmbv4F43AjcC5OTkTHr++bYzvqFlt9vDdsDvp4c9PPVVx+cKLjglnpOz+9cMdjifabSSZxp6oX6mNpuNkSNHhux6/Y3X6/WflypCQ55p6AX7TL/55htqamq63H/GjBm6+9RF1E9xTdM2KKXmAX8CXj/a7AX+RlMW7xKgtoP3VyilngIWA1OBdzro+zRNmT7y8/O13t5wNZybuh76dD98tb3DPiPGjGP6hEF9NKLQkI1yQ0+eaej1xubDsbxRrGyUG3ryTEMv2GdqsViYOHFij+8bUUEdgKZpLyql1gDjgWRgj6ZpZUqpzwEP8E0nlyg6+mvPD9KLEnobD7cl069CCCFE/xZxQR2ApmleYGvzn5VSA4CJwHqdferaGnX019JeGl6/U17X8dQrSFAnhBBC9HeRVigRQClloGk61gj8/mhbnFLKptN3KE0bFVcAG/tynJHsUFVncbBUvwohhBD9XURl6pRSScDnwMvAd4AN+CEwCbhb07QPjnZNAr5TSr0C7AKqgDHAfx997Yeapjn7ePgR62BV549CMnVCCCFE/xZRQR3gpmnfuauBgUA98AUwR9O0t1v0cwIvAacCc2kK5I4A7wJLNU37vC8HHem6lqnz9sFIhBBCCNFbIiqo0zTNDczrQr8GmrJyohOapvHcDVM4VOXkUFU9Byvr+ecn+wP6SaZOCCGE6N8iKqgToaeU4sTBNk4cfGwJ4tSRmfzPyk2t+klQJ4QQQvRvEV8oIUIvKT4wlpdCCSGEiAwFBQUopcjNzQ147dprr0UpxeLFi/t8XCLySVAXgxJ1gjqHW4I6IYQQ0WHv3r1cf/31DBs2jPj4eAYNGsS8efPYtGlT529ux+LFi1FKdfh14oknhvBTBE+mX2NQUnzg0SUOKZQQQoiIkJCQwJgxYxg8eHC4h9IvrVu3jrlz51Jf31QkaLPZKCkpYdWqVbz00kv84x//4Mc//nG3r2+xWLDZAnZVAyAzM7znHkimLgbpZupk+lUIISLC5MmT2b17N++99164h9LvlJSUcPnll1NfX8/MmTMpKiqiurqakpISfvSjH+HxeLj++uvZsWNHt+9x1VVXUVJSovtVUFAQug/TDRLUxSAJ6oQQQkSjBx98kNraWoYOHcqaNWsYPnw4ANnZ2fzzn/9k0qRJuN1u7r333jCPtHdIUBeDEs1xnDw0lTNGZjL7hBwuPWUwV+QPRdO0cA9NCCF63fbt27nuuuvIy8vDYrGQmprK6aefzlNPPUVjY2NA/6KiIv+aKYCPP/6YCy64gKysLBISEjj55JNZvnw5Pp9P935ut5vHH3+cqVOnkpqaislkIicnhwkTJnDLLbfwySeftOrfUaFEZxoaGnjkkUc49dRTsdlsWK1WxowZw8KFCykpKdF9z7///W+UUkyfPh2AtWvXMmPGDFJTU0lKSmLKlCk899xzQY+lr/l8Pp5//nkAbr75ZpKSklq9bjQaWbhwIQCvvfYatbW1fT7G3iZr6qLYEXsDFy//mKHpVoamJTAkLYHjshO54KRBvHLL6eEenhBC9Lnly5dz2223+QOwxMRE7HY7GzduZOPGjaxatYrXX3+dhIQE3fe/9NJLzJs3D4/HQ2pqKo2NjXz11VfceuutvPfee7z44ovExR370erxeJg1axbr168HmraZstlsVFRUUFZWxrZt26ioqOC0007r8WcrLy9n9uzZbNmyBYD4+HjMZjOFhYUUFhbyzDPP8MYbbzBlypR2r7FkyRLuvfdeDAYDycnJOBwOPvvsM66++mpKS0tZsGBBj8fZW3bu3ElpadOx77Nnz9btM2vWLKAp0P7oo4/4wQ9+0Gfj6wsxnalTSl2olHq6pqYm3EPpFdu/r+H7aief7qvkxU2HePTdQv703t5wD0sIIcLi1Vdf5dZbb8VqtfKHP/yB0tJS7HY7TqeTd955hzFjxlBQUMDtt9/e7jWuv/56zj33XPbt20dVVRXV1dUsXboUg8HAK6+8wtKlS1v1/89//sP69etJSEhg5cqV1NfXU1VVRUNDA/v372f58uVMmDAhJJ/vmmuuYcuWLaSlpfHCCy/gcDiora3liy++YPz48VRVVTF37lyOHDmi+/6vvvqK++67jyVLllBRUeFfi3b55ZcDcNddd1FZWRmSsfaGnTt3+n8/btw43T6ZmZlkZ2cH9A/Ge++9x6hRo4iPj8dmszFp0iTuuecef0AZTjGdqdM0bS2wNj8//4Zwj6U3fLC7LKDthEH6FTtCiOjy2J4dPF7YvR9abX134RUdvp639sWQ3Oe20eNYMOaEkFyrLa/Xy2233QbAypUrueSSS/yvmUwmZs6cyZtvvsn48eP5+9//zuLFixk4cGDAdYYMGcIrr7xCfHw80JTpu/POO3E4HNx333089NBD3HbbbSQmJgLw6aefAk0BV8uKS6PRyLBhw7jllltC8vk2bNjAW2+9BTQFknPmzPG/lp+fz7p16xg7diylpaX86U9/4v777w+4RnV1Nb/73e+4++67/W05OTmsXLmS9evXU15ezmuvvcY111zT5XEVFBQwY8aMbn2madOmBVV4cPjwYQDS09OxWCzt9hs0aBBlZWX+/sE6dOgQRqORlJQUampq2Lx5M5s3b+bJJ5/khRde4JxzzunWdUMhpjN10eZwjZNfrv6K+f/4nP/7qpj39wQGddPHZIVhZEIIEV4FBQXs37+f3NzcVgFdS3l5eUyZMgWPx9NuMPGLX/zCH9C1tHDhQiwWC7W1taxbt87fnpKSAtDtAKKrVq9eDTQFcC0DumY5OTncdNNNALzwwgu617BYLLrTqxaLxT+duX379qDGZTabycnJ6dZXenp6UPdyOBwAWK3WDvs1T63b7fagrj9q1Cgefvhh9u7di8vlorKyktraWp5//nkGDx5MZWUlc+fOpbCwMKjrhlJMZ+qiSaPXx+V/+YTvq50AfLCnPKCPQcG00RLUCSFiz8aNGwEoLi5mwIAB7fZrXo5z8OBB3debiwnaSklJYeLEiXzyySds3ryZuXPnAnDeeefx0EMP8eqrr3LRRRdx7bXXMm3aNDIyMnrwaQJt3rwZoMOs2Nlnn80DDzxAYWEhDofDn01sNm7cuIC2Zs175lVVVQU1rqlTp7ZboBFqzcV+zQUtofajH/0ooC0xMZGrrrqK0047jVNOOYWKigoWL17MX//6114ZQ2ckUxclPvm2wh/QtWfS8DRSE8x9NCIhhIgczZkyt9tNaWlpu18ulwvAv3FtWx1tCNz8Wnn5sX9UT5s2jfvvv5+4uDjWrl3LZZddRmZmJmPHjuWOO+5g797QrHNuvmdH4xsyZAjQFPzoratLTk5u973N05l61cGRornatb2/u2bNr7etju2JllPpr7/+eruV0L1NgroosaekrtM+M47P7oORCCFE5Gn+IXvJJZegaVqnX905W7W9baHuueceCgsLeeCBB5g9ezYpKSns3r2bZcuWMW7cOJ599tmefLRWGhoaQnat/mbQoEEAVFZW+oNzPcXFxQC6ayZ74tRTTwWgtrY2bAUlMv0aJQ5UdvwvE4CzWwR1Byvr2V9Rj73BQ73bg6PBw8RhaZw4WAophIgGC8ac0GtFB211VkgRCXJycoDuVzw2Ky4ubnf/uOZsYFZW4DKXvLw8Fi1axKJFi/B6vWzYsIHf/va3fPjhh/zsZz9jzpw5/qrM7sjKymLPnj3s37+/3T6HDh0CmqYn++o4q40bN3LppZd2671Tp05lzZo1Xe7fsuJ1586dnHLKKQF9jhw5QllZWUD/UGgZ1PfWFHBnJKiLEt8dcXT4+vjBNsbkHEut/3NjEX/76LtWfe4673gJ6oQQUal5H7g9e/awY8cOTjihewHv+vXrdYO6uro6/7o2vWCiJaPRyPTp03nttdfIyMjA4XDw5Zdf9mjPtFNOOYWPPvqI9evXo2mablDx/vvvAzB69Oh2186FWvN0d3cEm+0aO3YsOTk5lJaWsm7dOt2/h+YiFrPZzBlnnNGtcbXn888/B5qmsdPS0kJ67a6S6dcocdHJg/jh5KG6r5njDDx02Umt/iOXo8KEELHknHPOYdiwYQDcfvvteL3edvt2VAywbNky3G53QPtjjz2Gy+UiJSXFv8EtoNu3mdlsxmg0Aj2fNm3eS27Hjh28+uqrAa+Xlpby1FNPAXDllVf26F7BmD59epemu/W+gj1H1WAwMG/ePACefPJJfzVsM5/Px6OPPgrAhRde6K9M7orOTlw6dOgQf/7zn4Gm4hiDITzhlQR1UeLK/KE8cOlJ/Ob8sQGv3XvBOMYNav3Nm6QT1Nkb2v+fnBBC9Gcmk4knnngCpRTr1q1j1qxZfPbZZ/4f1h6Ph02bNrFo0SJGjBjR7nUOHDjAJZdcQlFREdC06P6RRx7hvvvuA+BXv/pVq9MorrnmGubPn8/bb79NXd2xtc9FRUX89Kc/xeVyYbVaOfPMM3v0+c4880z/VibXXXcdq1ev9geumzZtYtasWVRVVZGTk+Pfry8aLVq0iJSUFA4cOMCll17KgQMHgKZCkmuvvZYvvvgCs9ns//tqqfl4NqVUQED54YcfMnv2bJ5//vlW1bz19fW88MILTJ06lYqKChISErq1HjNUZPo1ysw/PY8dxbW8svV7AG4/dzQ/OnVYQD/J1AkhYs1FF13EihUruOmmm3u6pREAACAASURBVHj//feZMmUKFouFxMREqqurO8zeNVuxYgXz5s0jLy+P1NRU7HY7Hk/T/zsvvvhifvnLX7bq73K5WLVqFc8884z/iDC32+2vwDQajfz1r38NyRq3Z599llmzZrF161auuOIKLBYLJpPJH0ympaXx8ssvh3w7lUgyYMAAVq9ezdy5c3nnnXcYPnw4NpuN2tpaNE0jLi6OFStWBD39rmka77zzDu+88w7QtNed1Wpt9X2Tnp7Oc889x9ixY1sF8H1JMnVRxmhQPHrVyWz6zUx23T+Hn58zSndtRWK8MaDN7pagTggR3ebPn8+ePXtYsGABJ5xwAnFxcdTU1JCRkcGMGTN4+OGH/Vk4PZdddhkffPAB559/Pkajkbi4OCZMmMATTzzBmjVrWp37CvDggw+ydOlS5syZw4gRI3C73Xi9Xo477jjmz5/P5s2b+clPfhKSz5aVlcUnn3zCsmXLyM/Px2Qy4Xa7GTVqFAsWLGDHjh0hOWM20s2cOZOtW7cyf/58hgwZgtPpJCcnhyuvvJJPP/201ckeXTV+/HiWLl3KxRdfzMiRIzGbzdTU1GCz2Zg6dSpLlixh9+7drabew0F1Nk8cC/Lz87Uvv/yyV+9RUFDQ7qaV4fDerlKu/2frzzx9TBbPzJ8cphEFL9KeaTSQZxp6oX6mu3btYuzYwGUWsaKurq7D/dR6Q1FREXl5eUDna6v6o3A802gX7DMN9r9rpdQmTdPy27ZLpi5GyfSrEEIIEV0kqItRUighhBBCRBcJ6vo5TdNo9AZ/HEkoMnW1rkZufPZLqhztl+wLIYQQom9IUNfP7SmtY9KSdSxctZW3tpfgdHct26ZXKBFsULfq84O8s7OUy57ayKGqzk+0EEIIIUTviemgTil1oVLq6ZqamnAPpdve2l5CrcvDmi3fc9O/NjFxyTs8sq6w0/fpT78GF9S9sb3pSJx95Q4u+8tGdpfU4ulG1lAIISJZbm6uf0NcISJZTAd1mqat1TTtRput/x6N9db2klZ/djX6SLWaOn2f1WSk7U4nDR5fl4OywzVOthyo9v+5tLaBK/7yCaV1sXuYtBBCCBFOMR3U9Xf7KxzsLgnc4HD2iQM6fa9SikSz3rq6rk3fvt0mmATIsVkYnGrt0vuFEEIIEVpyokQ/9vaOwMBq/GBblwOrxHhjwJSr3e3BltB5pu9NnaBuUKqV17YVc6jKyeFqJ4svOkF342MhhBBChJ4Edf3Yp/sqA9rmdCFL16ypArb1dGl9F9bVVTncfFEUeO8PC8v5sLDc/+efnzOKjKT4Lo9HCNE1mqbJP5iEiBKhXKsp06/92N6ywKnX00d2/fzA7hZLfFFUia8L34MHq5xdHosQomuMRmOXzigVQvQPXq8XozFwR4rukKCun6p3ezikEzSNyk7q8jW6u6bus+8Cs3R6ZJsTIUIvISEBu90e7mEIIULEbreTkJAQkmtJUNdPfVvmoG3GdkiaVXdT4fbo9e1Kpu5znaDOplNxqxd0CiF6JiUlhcrKSsnWCREFvF4vlZWVpKSkhOR6sqaun9Kbeg0mSwfwyzlj+N+zR5IUbyTBHEdifJzulGyz6no3a7cd5uvvA/f1u2TiYJ7ZWNSqTTJ1QoRecnIyTqeT/fv3k56eTlJSEkajUdbYCdFPaJqG1+vFbrdTWVlJYmIiycnJIbm2BHX91N6ywOmXUTnBfVOMDqJ/hb2BK//6Cd+WOwJeG5puZXJeuk5QJ5k6IUJNKUV2djZ1dXXU1tZSVlYWU1k7l8uFxWIJ9zCiijzT0OvsmRqNRhISEsjMzCQ5OTlk/yiToK6f2luqE9QFmakLxoNv7tYN6AAm52YwJC1wGxUJ6oToHUopUlJSQjZl058UFBQwceLEcA8jqsgzDb1wPVNZU9dP6U6/Bpmp66odxTWs3nyo3ddPzUtnSFrgIs/iagnqhBBCiL4iQV0/5Gr0cqAycL3ayF7K1D3wxu6AooyWpozIIC3BRJyhdfq43u3F6Y6daSEhhBAinCSo64e+LbcHBFmDU60dFjl0V3ldA3tKA7OCzc4fP5BhGQkopUhPNAe8XuGQs2CFEEKIviBr6vqhb3SKJHorS5eVHM8Xd5+L2+OjtNZFcbWTPxd8y7B0K6Nzkvnh5GH+vumJZsrqWgdxlQ637tSsEEIIIUJLgrp+qOhI4NTrcVm9VyQBYI4zMDQ9gaHpCeTnpmM0BFbqZCbFA62zehUOd4/vfajewbNF3zBrwGDy07t+YoYQQggRS2I6qFNKXQhcOHLkyHAPJSj7KwKrUPMyg8+G7Smp41cvbcPR4MHR4MHe4OH4ASm8cNNpHb5PL6AD9Kdf7d0L6jRN47OKcp757hvWlXyPDzjgcEhQJ4QQQrQjpoM6TdPWAmvz8/NvCPdYgnHuuBxSrCb2VzjYX1HPgcp6hmckBn0dj8/H1oPVrdpqnI3dHldGUmBQV9nNNXXvlx3mvz//uFXbupLvKXbWM8gq07lCCCFEWzEd1PVXPxg/kB+MH+j/s8fr69Z19AorunJMWHsyepCpc3g8LNj8GYvGncRxScmclTWAARYrJa5j26L4gPdKi/lJbv/KrAohhBB9Qapfo0Cc0UCcMfi/Sr2zXx3uHgR1SfEBbV1dU/fYnh28W1rM/M82UN7gwmQwcMXQ3IB+H5Qe7vb4hBBCiGgmQV0M08vUOXqQqdNfU9e16df1ZSUAHKx3cP1nH+HweJiRMzCg38YjZbhi6EgkIYQQoqskqIth8XGGgKKHRq9Gg6cpaHI1etlbWtflKdm2068mY9fOsmv0+fjOcaxq9uuaKs5f/w65iUlkmFtn/xp8Pj49Utal6wohhBCxRNbUxTClFIlmI7Wu1kGbo8FLfJyRnYdrufTJjQAkW+IYZLMyZUQ69118ou71jh+Ywks3n0Z6YjwZSWaS4+O6dEhxkcOOp81uypXuBtLM8UzLHsCaQ/tbvfZ+2WGm62TxhBBCiFgmmboY19EUbMuzW+tcHvaU1nGwqv3zXJPi45g0PJ28zERSLKYuBXQAe+tqA9pOtKUBMCM7MHhrnqoVQgghxDES1MU4vWKJ5unWw9WugNcG2iwhH8PeupqAtpHJKQCcmZUT8E16oN5BbWP3t14RQgghopFMv/Yzv3nlazKT4snNSCQ3M5HcjARSEwILFLpKtwK2OVNXE5iVG5Rq7fa92rPXHpipG51sA8BmNjMsMYkiR+uj0Yoc7Z9HK4QQQsQiCer6EUeDh399eqBVm0HB7iXnYY7rXtI12RL4LVB5dBsSvUzdoNTQZ+q+0Zl+HZmU4v99nm5QZyel7ZuEEEKIGCbTr/3I99X6mbPuBnQAQ9ICM297y5oCqMM6mbqBttBm6hp9PvbZA7Nuo5NbBnXJAa9/p/MeIYQQIpZJUNePHKqqD2jTC8qCMTonMGAqLK1D0zQO6RRFDApxUHfAYaexTeVrhjme9PhjW5nkJiUFvO+7Npk7IYQQItZJUNeP6AVZg1N7dg7qmAGBQd2ekjoOVjoDToMwGRUDgiiU8Po0KuwNVNe3f6pEoc7U66jk1hOrkqkTQgghOidr6vqR73WCup5m6sboZOq+Lbfz6b6KgPbxg22dTvW+vOUQT37wLZUON1X1bnwa3HjWCH79g7G6/ffpFDyM7EpQ57CjWWwdjkUIIYSIJZKp60d0M3U9DOoykuLJTGpdPdvo1Vi96VBA3/zc9E6v52r0sbfMToWjKaADOFLX/lFhxc7AKeXchNbTrQOtVuINrb9V6zyN1LWZthVCCCFimQR1/cghnUKJnmbqQH9d3edFlQFtk4andXqtnJT4gDa9Ao9mh3WCuoHW1lPKBqXITQxcV1fikzNghRBCiGYxHdQppS5USj1dUxO4+W0k+l6vUKKHa+pAP6jT05Wgbkha4Hg6CuqKnToVttbAQLXtFGxWvAUnvk7HI4QQQsSKmF5Tp2naWmBtfn7+DeEeS2dcjV6O2FsXHBgUQRUutOekITYmDE1lTE4SG/Ye4XBN4P50uRkJZCYFZuHaGqyzOXFJjQuP10ecMfDfECWuwEB1kDUwMPxx7nGcN2gIeYnJ5CYmkWwyUVBQ0Ol4hBBCiFgR00Fdf6K3ni4nxdKjPeqaXXrKEC49ZQgAZbUuJv/hvYA+Xc3mJcbHkZpgorr+2DFeHp9GaV1DQMBX7/FQ0+a4rzilyIwPDFRPz8rp0v2FEEKIWBXT06/9id4UZijW07WVnWLhrNFZAe3njM3u8jVaBW9GDUY1MPfTd7n5y41UNhwrmtBbT5dtsWJUKrhBCyGEEEKCuv5Cb+NhvanOULhl+nG0jKvSE82cf9KgLr//2Lg0GNOAyvZypNHFW4e/57rPP8J7tGq1UdOYlJbBIGuCP5AbZOmdzySEEEJEO5l+7Sf096jreZGEnlNHZLDip/n857ODJFviuPXskSTFd/1bxT+udC8qrXUxw1fVlbx0sIgrh+VxfIqN1WecDYBX0yh3uXBJRasQQgjRLRLU9RO9sUddR84+Poezj+/eOrbBaVZQGuQ26r7+0K5tnJ0zsNXaOaNSDNCpehVCCCFE18j0az+ht6aut6Zfe2pwqhUGelBW/c2BK91uLv7wXf6zfx/lDYGVtkIIIYQIngR1/USJzjYjvZmp64mMFBMM1s/SNSt2Obl72ybOePd1XjpYFPQ9Grxeir0eCuv6xx6DQgghRG+T6dd+wOvTKK0NDOoGpPR8j7resMF+GGXuvB+A2+fj19s2MT41jdHJHZ/lur26iod3b+c7Rx2H6h34gLN3fs2KU8/o+aCFEEKIfk4ydf2A3eVhZHYSNqvJ35ZiiSMxiOKFvmL3NLLywLcB7VojWA1G3fe4fT7u3PoFHl/HJ0T40FhfXsKBowEdwD5HXU+HLIQQQkSFiIsKlFI5wH3A+UAOUAK8DPxW07TqNn3HAA8B0wAzsPlov/f7dNC9zJZg4q0FZwFQ7/ZQUuOi2tnx9Ga4vF58iOrG1idfaBqwzcJDV0/kz8U72FNXG/C+Bq+PIw0NHRZL5CYGboB8sN5Bo8+HySD/PhFCCBHbIuonoVIqG/gMuA54BbgVeBW4GfhAKZXQou9xwEbgNGApcCeQBLytlDq3j4feZxLMcYzISuKUYZ2fwxoOZ2Tm8Mvjx5PoOZZVpMwILgPGBiOvT5vFU/lT/S8ZleLWUWP5v7PO7bT6NcVkIsPc+qgyr6ZxqN4R0s8ghBBC9EeRlqn7NTAcuFrTtOeaG5VSG4H/AAuB3x1tfgBIBSZpmrb1aL9ngR3An5VSx2uapl9+KXrN4IQEbh51PPu2NvDitgOQ6YXKpmnX6vpGjEoxe+BgXjr9bO75ejMPTcjnxNSuB6h5SclUVDa0avvOYScvqWvHmAkhhBDRKqIydcAMwAk836Z9FeAC5gMopRKBi4CC5oAOQNM0O/A3YDTwX30xYKEvNcEETgMcNIGj6dus2nlsWvaU9AzWnnVuUAEdQF5iUkDbQcnUCSGEEBGXqYsHXG0zbJqm+ZRSTmCEUioTGHW07yc61/j06K//BXzem4MV7Zt6XCYGgyLVasZmNWGzmhg3KKVVH0M3zngdaA08RaPUFbiHnxBCCBFrIi2o2wGMUUqd3DIDp5Q6GWhO6QwDmg8i/V7nGs1tg3ttlKJTM47PZsbx2SG/bo7O2bAlEtQJIYQQERfUPQbMBV5QSi0AtgMnHG1vBExAwtEvgAadazRv6NbhwahKqRuBGwFycnIoKCjo6dg7ZLfbe/0e4VDkbWSYIa5bWbfuKPcE/pXvPlxMQU1Bn9w/2kXr92k4yTMNLXmeoSfPNPTC9UwjKqjTNG2DUmoe8Cfg9aPNXprWye0ALgFqgfqjr8UHXASad+St13mt5b2eBp4GyM/P16ZPn96jsXemoKCA7txD0zQ+/66SgTYrObZ44uP093rra40+H38q3MnyvbvIT8/k9ydN7HTz4FDIrqnmsQ/XtWprsFi69WxFoO5+n4r2yTMNLXmeoSfPNPTC9UwjKqgD0DTtRaXUGmA8kAzs0TStTCn1OeABvgESj3bXm2JtbtObmu13quobuerpT/1/zkg0MyIrkRdvmtrBu3rfr776kpcP7Qfgy8ojnL9+HTccN4ZbR43FGtd731Z606+ypk4IIYSIwKAOQNM0L9ByTd0AYCKwXtO0eqXU1zRNvZ6m8/YpR3/9stcH2gcO17QOWCocbmwJpnZ6940Sp9Mf0DXzaBp/+WY3cUqx8PgTe+3e6WYzZoMBd4vTJxxeD3WNjSSbwvtchBBCiHCKtC1NAiilDDRNxxqB34N/65K1wHSl1IQWfZOA/wb2EiWVryU1gWe+DrSF98zXTVVHdNsnpWXwPyOP79V7K6XIjg/8/JKtE0IIEesiKlN3NCj7nKZjwb4DbMAPgUnA3ZqmfdCi+13AOcA7SqlHaVprdwNN06/nR8vGw4d1groBKR2fvNDbtlRVBrQlGuP492nTiDf2/pq/ARYrh5ytl0yWuJyMTE5p5x1CCCFE9IuooA5wA9uAq4GBNBU7fAHM0TTt7ZYdNU37Ril1OvAgsIhjZ7/O0TTt3T4ddS8qrdUJ6mx69SF9Z2tVRUDb70+aFBDQrfykiJJaFzXORqrrG6lxNvLEDyeSmmDu0f1zrFaoat0mmTohhBCxLqKCOk3T3MC8IPrvAi7uvRGFX1lt4BYeOSnhmX59bv8+lhfupFgngJqYlh7Q9vSGfRysbN230uHucVA3QPaqE0IIIQJEVFAnApXVBWbqspP7PlO3paqCX2/bpPtautnM0ITEgPZUq5mDtA62apyNPR6LbEAshBBCBIr4QolYV1YXmKnLSu77TN0rhw60+9rJaRkonc2HbdbAatTqEAR1epm6UqcEdUIIIWKbZOoinF5QF45M3e7a6nZfm5iaoduut/VKbQiCutHJNq4aloerpJRTjx/LAIuV4YlJPb6uEEII0Z9JUBfBvD6NCrtepq5vgzpN09hTV9Pu6yfrrKeDdjJ19T0P6sak2HhwQj4FVQVMHz6ix9cTQgghooFMv0awCnsDvjYbs9isJiymvj0qrMTlpKZRPxhLN5s5JU0/U5eqE9SFYk2dEEIIIQJJUBfBImXqtbrRzbiUVMyG1t8uKSYTS8ZPIqGdY8F6K1MnhBBCiEAy/RrBdCtfU/o+qBubksrr02bi8fkoctj5uqaKKncDlw/NI6WDo7lSddbUSaZOCCGE6B0S1EWwUp096rLDUPnaLM5gYGRySpdPbtDL1NU43aEelhBCCCGI8elXpdSFSqmna2raLwIIJ72Nh8ORqesumzVwk2HJ1AkhhBC9I6aDOk3T1mqadqPNZgv3UHTpbzzcvUxdtdvNiwe+48UD31Hj7ptsmaypE0IIIfqOTL9GsFAVStQ1NnLFx+/zjb0OgKe/3cOrZ57bboFDqPTmmroih503GupZv30LJS4nZS4XE9My+M0JE0JyfSGEEKK/kaAugoUqqPt/3+7xB3QA39jrWLGvkFtHj+vR+DrTWydKAHxnr+NFtwO++8bfZjH27VYvQgghRCSRoC6CLbtiAsXVTsrqGiirc1FW20BuZuAZqx3xahr/+G5vQPvz+/fxv6PG6h7v1aza7cZiNHY7WEowGzEZFY3eY5vtuT0+XI3eHu+1p3v+qxwVJoQQIoZJUBfBRmYnMTK7Z8dfbSgrwe7xBLQXu5xsq6liQqr+aRAAjxXuYO33B7gmdyQ/yR1JenxwWUKlFDariSP21mv4qusbGWDrWVCne/6rS4I6IYQQsSumCyViwaoD37X72twN77HfYdd9zatpvHroAJVuN48V7uT0d1/nnm2bOdIQWLzREf1tTXo+BZtmNgf8i8Th9VDXzskXQgghRLSToC6KlTe4eLe0uMM+537wFuvLSgLat1ZVUN14LMPm8nl56VARJhXct4x+BWzPq2+VUqTpjKVEsnVCCCFilAR1UezlQ/vxaFqHfTyaxsO7vw5o/0An0DsjMwebOXDvuY6kJpixmowMSLEwJieZyXnpmONC822XapCgTgghhGgma+qilKZprNrf/tRrS9trqqlxu1sFbB+UHg7oNz1nYNDj+H/X5GM0tF+M0RNpygi0Xi8o6+qEEELEKsnURalNVRXsc9QFtGeYWxc7mA0GJqVlcMR9bPuUEqeTnbXVAe+dnj0g6HH0VkAH6E+/SgWsEEKIGCWZuiilVyBxakYWz502jXu/3sIRt4sbRozhBFsq8W22LFlfHjj1enyKjUHWhF4bb3ek6Uy/SqZOCCFErJKgLkI9+OZuPvuugsykeDKT4slKMnPBhEGMzknu0vt/mjsSm8nEa8UHKXU1VaxeNSwPpRRLTjpF9z0H6x18XV3F377dE/DajOzgp157W6oUSgghhBB+EtRFqN0ltWw50HoK9ITBti4HdSempnFiahp3jZvAFxXlvHH4EOcNHNJu///7/gC3b/4MXzuvR2JQ17SmrrViZ30YRiKEEEKEnwR1EeqIPfCIsMyk4I8IMyrFlMxspmRmt9vn+/p6frX1y3YDuoEWK5PSM4K+d2/L0pl+3e+wo2lahydlCCGEENFICiUi1JG6wL3csroR1HXF73d+hcvnbff1CwYPxRCBQVKqMmBtsx7Q4fVwpCEwIBZCCCGinWTqIpCmaVQ4AgOTjKTg9ojritpGN7t0Kl1bumjQsB7dw+3xUeNspMbppsbZSEOjj6kjM3t0TWjagHh4YhK7a2tatRc57GRZLD2+vhBCCNGfxHRQp5S6ELhw5MiR4R5KK7VOD43e1psGW01GEuND/9eVYjLz1rRZPF64k798szvg9bzEJE6wpXb7+kfsDeT/7t1WbakJJrbeO6vb12wpN0EvqKvjvzJ6HjQKIYQQ/UlMT79qmrZW07QbbTZbuIfSSrneerrk0GfpmsUbjfxy7HgenTg54LUfDT+uR+vT2jv71efr+KSLrspNSgpoK2rnPFshhBAimsV0UBepulskUeZy0uBtf21cZ+YOGc6Tk04jwxyPSSnmDcvj2hGjun09AJPRQKK59bo3TYO6Bk877whObmJgNbAEdUIIIWJRTE+/Riq9oC4jsfOgbsmOr/ig9DDTcwYwe8BgpmcPJNkUmCnryHmDhnDeoCE0+nyYdKpLu8NmNeFwtw42a+obdbN4wRqekIRRKYYmJJKbmERuYhKT0mTqVQghROyRoC4CVdh1Kl87mX5t8Hr5oPQwDq+H14sP8XrxIcwGA6tPP5vxqWlBjyFUAR2ALcFMcY2rVVuNszEk185Pz2DXDy4N6XiFEEKI/kiCugjUnenXjUfKcHhbT2lajEbGpIR/vaDNGvhtVu0MDFy7I06COSGEEAKQNXURqTtB3ftlhwPaZmQPxBwBQU+qNTDLGKpMnRBCCCGahP8nvghQrrPxcEd71GmaxgelgUHduTmDQjqu7tJbO1ddL0GdEEIIEUoS1EWgYDN1hXW1fN/mzFOjUpyVnRPysXVHaoL+tiZCCCGECB0J6iKQ3mkSmR1k6j7QmXrNT88kxdR7e9sFI6WdveqEEEIIEToS1EWgSp3q1462NHlfZ+p1RvbAkI6pJ3QzdTL9KoQQQoSUVL9GGFejN2BPN6NBtbunW21jI5urKgLaz86JnKBOd01diKpfAbyaRrGznsPOekpcTg47nbh9Xm4dPS5k9xBCCCEiXVBBnVKq+WT3Q5qm+XphPDGvqj4w2ElLMGEw6B/V9emRMrxa6yO3BlsTGJkUeNJCuPR29WuN281Z773Rqi3BaOR/R43t0RFnQgghRH8SbKauCCgFBod+KALAoBSXThxMhcNN5dGvjtbTbSgvDWg7K2tARAUzvV39mmY2E28w0OA79u+Meq+XOk9jxKwrFEIIIXpbsEFdDbBfsnS9JyfFwiNXndzl/hvKSwLazsyKjKrXZnpr6mpDmKlTSjHQmhBw5mux0ylBnRBCiJgRbKHETiRLFzH2O+zsr3e0ajMAUzOzwzOgduhVv1aHuPp1gMUa0FbSZpsXIYQQIpoFG9Q9DQxWSl3SG4Ppa0qpC5VST9fU1IR7KN3ykc7U64S0dGzmyMpOJcfH0XZJYL3bi9sTuoTvQGtCQNthlzNk1xdCCCEiXVBBnaZp/wSWAyuVUrcrpTJ6Z1h9Q9O0tZqm3Wizhf981O746EhgUHdm1oAwjKRjBoPq9b3qBupk6g5Lpk4IIUQMCbb6dd/R38YDDwMPK6WOAI523qJpmnZcD8Yn2qFpGl9UHAloPz3Cpl6b/WrO8RgU2KxmUhNM2Kwm3bV23TXAqjP9Kpk6IYQQMSTYQolcnbaso196tHbaRQ99Y6+jwt365Il4g4EJqelhGlHHfjh5WOedemCgRWf61SlBnRBCiNgRbFA3o1dGIYL2fb0Dm8lETeOxKcxT0jKINxrDOKrwGaiTqTvskulXIYQQsSOooE7TtPW9NRDRZM3mQ8QZDWQkmklPNJORaCYzKT5g8+HpOQPZPPtidtfW8FlFOZ9XlPNfGZlhGnX4DdDN1NWjaVpE7dknhBBC9BY5JizC3P/azoCNeT+/+xyyky0BfQ1KMc6WyjhbKvNHjOqrIUakdLMZs8GAO2ADYg8pptCt3RNCCCEiVbBbmvgppcxKqSlKqcuVUteEclCxyuP16Z60kJYQWVuURCKllG4F7PfO9mp4hBBCiOgSdFCnlIpTSt1P03FhHwOrgH+06fO0UmqfUmpkaIYZG6p0Ajqb1YTJ2O3YO6YMTUgKaPvObtfpKYQQQkSfoKIFpZQRWAvcDaTQdBas3k/NN2mqlI2KTYr7SqXDHdCWkShZuq7KS9IJ6hx1YRiJEEII0feCXVN3IzAb2A3M0zRtm1JqAzC1Tb+3AC8wB/hjj0cZIyocDQFtaVES1O0trePvHxdR43RT42ykur6REwalsPTyCSG7R15ickDbd3YJ6oQQQsSGYIO6nwI+4CpN075ur5OmaU6l1LfA8J4MDhvfsgAAIABJREFULtboZerSoySoq3S4ee7zA63a4uNCO608IikwqNsnmTohhBAxItifqmOB/R0FdC1UAYOCH1Ls6sr065aqCt46fIjKhsCsXiRL1Sn2COUxYaAf1MmaOiGEELEi2ExdHNDV1EcSEBiliHZV2DvP1P2r6FvWHNoPwOjkFCZnZHH18BGMTUntkzF2l62Xz34FGGRNaLWtSY7FQl5iMi6vF0uMbsoshBAidgQb1B0ERiqlTJqmtfsTWSmVCowBtvVkcLGmqr7zoO6zinL/7wvraimsq+XcnEGMTen14fWI3jmvNc7GkG4ObFSKJ/NPIyfeSm5SEklxsj+dEEKI2BHs9Os7gBW4rZN+vwGMwBvdGVSs0tvSpOUedYed9XzvbH30lVEpJqVn9PrYespiMmJus4au0atR7/aG9D7n5AzixNQ0CeiEEELEnGCDuoeBeuABpdT9SqmhLV9USh2nlHoSWAhUA0+EZpixoUpnTV1a4rHgZGtVZcDr41JS+00Ak9oHU7BCCCFErAoqqNM07QDwQ6CBpr3qioDTAJRS9UAhcBPgAK7UNK1c/0qRQSl1oVLq6ZqamnAPBdCffm2ZqdtSXRHw+sS0yM/SNdNbV6d3goYQQgghghf0nhKapr0G5AOrAdfRayjAAjQCLwOTNU17L4Tj7BWapq3VNO1Gm80W7qEA+gFOq6BOJ1M3MS29V8cUSu2tqxNCCCFEzwVbKAGApmm7gauUUiZgFJBK08kSezVNc4ZwfDGlo0xdo8/H19V6QV3/ztTVOKVAWgghhAiFbgV1zY5WwO4M0VhimqvRG1A0YFCQbGn6K9pdW0PD0a06mqWbzQxLSOyzMfaUzRq4V51MvwohhBCh0aOgriWllAU4m6YNhzdpmrYlVNeOBe1NvRoMTdt9bKkKXE93clpGyLYD6Qt9sVddSw1eL6UuJ4ddTsam2EgxRcfpHEIIIYSeoII6pdQFwM+BlZqmrWzRPgx4Gxjdou1RTdPuCNVAo53e1GvLNWh6Qd3E1P4z9Qp9F9TdufUL3i8tptJ97JmunHIWZ2TlhPxeQgghRKQItlDih8A5wPY27Y/StNlwLU0bDvuA25VS5/d4hDGis8rXrbrr6fpPkQSAzRr4b4jeCOpcXm+rgA6gxCVLPYUQQkS3YKdf84GallOrSql04EKgAjhR07RSpdQPgX8DNwCvh2qw0WxKXgZb751JVX0jlQ431fVurKamo62q3A0UOVqfYaqAk1L7WVDXR9WvAyzWgLbDbTZtFkIIIaJNsEFdFnCoTdu0o9dZpWlaKYCmac8ppR4FJvd8iLHBYFCkJphJTTCTl9m6+EFv0+FRySkkm/rHpsPN+mr6VS+oK5VMnRBCiCgX7PRrEtB2nvB0QAM+aNN+AOhfi74iVHtFEv2NXlBX2xtBnVUnUydBnRBCiCgXbFBXAQxXrUsuzzn660dt+ppo2rtO9JDupsP9bOoV+i5TN9CSENBW4pSgTgghRHQLNqj7EkgH/gdAKTUbmADsaJ56PdqugOOAUr2LdEQplaSU+rVS6mulVJ1S6ohSaqNS6tqWwaRSarFSSmvnK2qqbn2axlf9/HiwZqkJZganWhk3MIXTRmQw54QBnDM29BWpOTrTryUuWVMnhBAiugW7pm45cD7wZ6XU7wEbTVOvy9v0m0LTVO3mYC6ulDIAbwJTgX8CTwAJNFXd/gMYC/yqzdtuB460adsUzH0j2T57HXUeT6u2pLg4RianhGlE3ZeZFM/Hi87u9ftkWywomr4xm1W63TR4vcQbjb1+fyGEECIcggrqNE17Wyn1M+B3NGXsXMAjmqY93abrdUd/Dfb811OBM4DHNE27vblRKfUksJumDGHboO4VTdOKgrxPv1HnaeSk1DR21VTTqDWFKRNS0zH2o02H+5rJYCAr3kJZg6tVe6nLybDEpDCNSgghhOhdQZ8ooWnaU0qpp2mqhC3XNM2n0+0RmrJ3hUFevjn9VNzmnm6l1P9n777D47yqxI9/72g0M+rVktx7L3FiJ3G60wMhBEKAsEACG8hvk1DCLoSlLks2QMLCsiG0bIAUIKGkOqQXpTo2bnEvcrdsq7cZTdPM/f3xjmyN3ldlpOk6n+fRo/jOO/NeT2Tp6Nx7zm0GnFZPUkoVA91a6x6rxzOB1trydIhTyyp46rxL8IdCbOtoZ2Nbi+XyoohWk5dnCuqOSVAnhBAii43omLBIIDfgfjmt9Y4Rzmct0A7crpQ6AKwB8oDPAMuAf7F4zmagCAgppdYCd2itnxvh/VPm9DtfBox9Z2X5uZTlO/ifjy+lwGn8L3Lm5HBaeQWnlWfeXrpUqHHlsZm2qDFpayKEECKbKa310FclkVLqPOB++hw5BnQB12utn+xz3W0Ye+zeAdowTrS4DRgP/LPW+oEh7nMTcBNAdXX1skcffTSOfwszt9tNYaF1liisNTe+0E3//xO/vSyfHJsssw5ksPf0D74uXglGZ+o+5izgfQ5zZaw4abD3VIyMvKfxJe9n/Ml7Gn+Jfk8vvPDC9Vrr5f3HBw3qlFK/G+brhzAybDuBF7XWh0c0S+OepwLfBvZhBGzlwK3APOBqrfVLgzy3AuMIMxcwWWs9rJYqy5cv1+vWrRvplIeltraWlStXWj7W6glw2h3Rf60il50t37s8oXPKdIO9p7+u28ldO7ZEjX1m+iz+Y9GpSZhZ5hrsPRUjI+9pfMn7GX/ynsZfot9TpZRlUDfU8utnMIoIh0oX9Y0MQ0qpB4Ava61jWu9SSi3GCOS+orX+dZ/xRzCCtf9TSs3UWocsJ6F1i1Lq18D3MCpoX4zl/qky1LmvInaWbU2kV50QQogsNlRQ9xCYVgUHep1yYAkwEbgRmAJcEeN8voKRZftr30GtdbdS6u/AF4BpwN5BXuNA5HNljPdOmTaPVVCXWUeADVdHd5CjHV46vMETH/NqilgyqTSu9xlv2atOgjohhBDZa9CgTmv9mVhfUCn1PoyecpcqpT6mtf5LDE+fGPls1UzM3u/zQGZHPsfc+DhV2rvNpyqUZGmm7vfv7OdnL++JGrv1wplxD+pq8ixOlZCgTgghRBaL9USJIUUqT6/HWLL9ZIxP3x75/Jm+g0qpUuBqjIKIvUopu1KqpP+TlVKTgZsxjjN7J8Z7p0y7xVFZpfl2Hj5Qx4bWFrw9GdupxaQ0SUeF1Vhk6hp9XnrCVh14hBBCiMw3opYmQ9Fav6iUagRMm/iG8DOMgPBHkf11b2Ms634eo6r1Vq11TyTI26+UehLYwcnq189hnGTxiVj386VSu8Weupw8zXe3bASMyHtmYTFnVo7jjsWnJXl28VVisazc4Y1/0OrKyaE010F78OR7Gwaa/D7GW2TxhBBCiEyXkKAu4gCwNJYnaK0PKqXOAL4LXAxcB3iBTcC/aa0fj1zqBR7DOIHiQxiBXDPwMnC31nptPP4CydJpkanyOnsgMhwG9rg7KXda9l7OKCVJytQB3DhjDkoZWbvxefnUuPIY53Ql5F5CCCFEqiUyqHMC5hTUELTWe4EbhrjGj5GVywpWy6+dNr9pbFFJfPedpUIyg7ovzJmfkNcVQggh0lHc99QBKKXyMfrKHUnE62cbq0KJZnymsYUlZcmYTkJZBXVWmUohhBBCxCYhQR3w7xiZutoEvX5WMWfqNPVBj+m6RVkQ1BUnMVMnhBBCjCVxWX5Vxkn0ZRh76D4L/BPQA9wbj9fPdh39CyWcmu5wdPGAy5bDjMKiJM4qMQZaftVaY3wZCSGEEGIkBg3qlFKWJzcMQWHs7f+y1nr7UBcLi0xdgbntxoKSUnKyIOhx2nNw5drwBU/+HUNhjScQotCZyC2eQgghRHYbavlVxfgRAl4BLtZa/ypBc846pj11heagbmEWFEn0SmaxhBBCCDFWDJUauXCYrxMCOoC6TOoPlw5CYU2nb+hMXTbsp+tV7MqloTO6utfty54Gy0IIIUQqDHVM2OvJmshY1eULovudrmsr1KYDd7MpU1dgsczq9icmU3f/3t0c8HRx3OfluNfLcZ+XF1ZeTkUW9PwTQggh+pJNTClmWna0a7QjOqSzK8XsItOpaBmryGX+sutKUKbu0UP72Ovuiho77uuWoE4IIUTWGdNBnVLqKuCqWbNmpWwO+Q47X7p4Nh3dAdq9QQ4Eu9hM9Ar2zMIiHLZEdZ9JPquCCLc/MUFdjSvPFNQ1+XyQPTGyEEIIAYzxoE5rvQpYtXz58s+nag7jipz866VzTvz59/v2sHlbQ9Q1c4uzKwKxDOoSlKkrc5gzcm3BmA86EUIIIdJe9qR/ssTOrg7T2Lyi7NlPB1CYxOXXUofDNNYekKBOCCFE9pGgLs3s6jQHddmWqSuyyNR1JWj5tSzXHNS1Bczn6gohhBCZToK6NBLWmt1WmbosC+qsMnWy/CqEEEKMzpjeU5duDnV78IaiD/Eosucy3pWXohklxuULa5hTXUSRy06hM5dCl51Si4bE8SDLr0IIIcaKmII6pdT1kf/8s9Za1rDirMXvY2JePvXe7hNjc4tLsu5M1KkVBUytKEjKvUpl+VUIIcQYEWum7vfAfq31Q4mYzFi3rLySty65ks5gkN1dHezq7KA4NzEZrLFCMnVCCCHGiliDumagJRETGasefOcAbn8PJXm5lObnUpKXy6lTylheXsny8spUTy/jyZ46IYQQY0WsQd0a4DyllF1rLYd1xsHD7x6krtEdNfbcl89j/vjiFM0ou1hVv7bL8qsQQogsFGv1611AIfDdBMxlTGrvNp95WpovS67xUpSba/oi7w6F8PcrSBFCCCEyXayZusPAN4E7lVKnAA8COwDPQE/QWh8a+fSym9aazv5nvwKleebskhgZm1KUOhy09ttH1x4MUJ2TXVXFQgghxrZYg7r9ff77A5GPwegR3GPM8AXDBELhqDFHjg1XrrQPjKfSXKcpqGsLBKjOslYxQgghxrZYA65Ye2tkVy+OOOuwyNIV59mzroWJlX95eD0tHj9dvh7cfuPj7a9fRIHFaROjVeZwmHLJsq9OCCFEtonpJ6jWWlJIcdTp6xPUlYQgqCjKGxuJzXUH22h2RwdWbn9PQoI6q7YmbdLWRAghRJaRIC2FTmbqNMz1o071cWB2M1fUvsCXN7yb1f3UiiyOCuuSo8KEEEKIERsbaaE01dFb+ZoLqrfgVcGurk72e9z8ZOkZKZtbohVaZOTc/sQEdRPz8plWUEhproMyh4NSh5Mp+ck50UIIIYRIlhEHdUqpFcBKYBKQp7W+sc9j44HcdK98VUpdBVw1a9aslNz/xPJrftj02KzCYuy27E2kWgZ1CcrU3TZ3IbfNXZiQ1xZCCCHSRcxRg1JqilLqDeBt4E7gFuAz/S77PrBfKXXWqGeYQFrrVVrrm0pKSlJy/xPLrxZB3Zyi7G4+XGix/Or2mwtHhBBCCDE8MQV1SqkyoBY4F6gHHgCOWFz6IEbl64dGN73sdjKo06bHZmd5UFdkkalL1J46IYQQYiyINVN3OzANeAqYF1lyPWhx3TuAFzh/VLPLcp3eSBCTZ738ms2sM3US1AkhhBAjFWtQ9yEgAHxOa9090EVa6zBQh7HfTgxgsOXXWVmeqUvmnjohhBBiLIg1qJsK7NFatwzjWg8wLvYpjR2dviDY9cnK1wiHzZb11ZlWmbouydQJIYQQIxZrUBcAzE2/rFUBXTG+/pjS4Q1aZummFxRmdeUryJ46IYQQIt5ibWmyGzhFKTVOa9000EVKqVnADOC10Uwu23V6g5b76WZm+X46SP6euma/j0afj2a/j2a/n+aAj09MmUFRbu7QTxZCCCEyQKxB3RPAcuAnwPVWFyil7MC9gAb+OqrZZblObxDKx17lK0CR0xxMuX2Ja2nysbdfY7/HHTV2/rga5uWmpp2NEEIIEW+xrvHdA+wDPqmUqlVKfRooAlBKna6UuhH4B3AZsA34XTwnm23CmjFZ+QrJz9RVOl2msWa/L2H3E0IIIZItpkyd1tqjlLoCeBqjXcl5fR5+N/JZATuAq7TWcsDmIN795sWseOkZGnzeqPFsr3wF6+rXRO6pk6BOCCFEtot5N77Wug44FbgVY89cCxACOjBOmfgysExrbdW/TvTRGQyaAjobRqFEtit2mZdfO72JW36tdJrre5r9/oTdTwghhEi2EZ39qrX2A7+KfIgR2uvuNI1NKSjEmZOTgtkkV0meOajrSGhQJ5k6IYQQ2S27+2akubouc1A3ewzspwMoctlRKnrMEwgRDJn3GMaDBHVCCCGy3YgydQBKqXzgHGAuRrFEF7ALeHuw0ybESXVucxu/sbCfDsBmU5w1o4Icm6IkL/fERyisyU1AorLSIcuvQgghslvMQZ1SygF8B/gikcrXftxKqZ8Dd0SWacUAjnR7TGNjofK1158+vyJp95JMnRBCiGwXU1AXCeieA1ZiVLm2Y5zx2gBUAzOBMuAbwDlKqcu01onbKDVKSqmrgKtmzZqVkvvfu2wFzX4/e9yd1HV1ssfdySmlZSmZS7azLJQISFAnhBAie8SaqfsqcCHQBnwNeLhv0KaUygU+DdyF0fLka8AP4jPV+NNarwJWLV++/PPJvrcn0pOt0ulknKuKsyurkj2FMcUqU9fi9xPWGlv/zX1CCCFEBoo1qPs0xkkRH9Rav93/wUiA9zul1C7gTYxTJ9I2qEule17dw29e34fdpiiO7Ce7+YKZfOz0yameWlbKt9vJz8mhOxQ6MdajNR3BAGUW++2EEEKITBNr9es0YI9VQNdX5PHdwNQRzivrdXqNTF1PWNPqCbC/2YO/JzTEs8RoWO+rk22fQgghskOsQV0b4B7yKoMbaI3x9ceMTotzTostereJ+LFegpV9dUIIIbJDrEHdy8AipdS4wS5SSlUBiyLXCwtWpydYnbIg4kdOlRBCCJHNYt1T9x3g/cBflFLXaa0b+l+glKoGHsXoW/ed0U8xO1mdc1pkcch9Nmv1BNhxrJMOb/DEx5TyfN6/eHxC7meVqWuSTJ0QQogsEWsUcQHwS+DrwAGl1OPAdqARqAIWANdgtDu5C1ipLCoLtdYPjWLOGS+sNW1BH0bNycn3p2iMZere3dfCLX/cEDV2yfzqpAZ10qtOCCFEtog1qHuA6EjkOotreh/79iCvM6aDumNeL4dmthplJ14FXht4bGMuU2d1/qvVsnS8WPeqk+VXIYQQ2SHWKOINjKBOjMI+j3E8mMoBCjUUhtAuLUEd0JHAoG5cv0ydAnwh8zK4EEIIkYliiiK01isTNI8xpa6z0zSmvIoChwR1iQzqTiur4OfLVjDelUeNK58ql4tcW6y1QkIIIUR6GltRRJrYYxHUOYJ2bLaxdbJBSX5yg7oqVx4fmCDNnYUQQmQnSVOkQJ3bHNQVhMdWkQRAocNO/zjWGwxJE2YhhBBiBCSoS4FD3R7TWDGOFMwktWyRI9L6S2S2TgghhMhWEtQlmT8UojHgNY1X5JjbbYwFya6AFUIIIbKVBHVJdqjbYyof1n5FqXPsZeoASiVTJ4QQQsSFBHVJts/dZR70qjHXzqSXLL8KIYQQ8TGmgzql1FVKqfs6OjqSds/9HqugzjbmTpPoley2Jv11BYNyVJgQQoisMDbTQxFa61XAquXLl38+Wffcb5Wp843dTJ1lUNeduKBuY1sLP925jeO+bo77vLh7erisZgK/Of2chN1TCCGESIaxGUmk0D6P2zwomboo7QnM1AXDYd5qbogaO+41F64IIYQQmWZML7+mglWm7n0zx7NwQnEKZpN6Vnvq3L7EHd1V48ozjR3zSVAnhBAi8w2YqVNKvRqH19da64vj8DpZoSMQoKXfAfJ2pfj5Ncuwj9Hjqgqd5i9BTyBxQV21RVDX7PcRDIflyDAhhBAZbbDl15WDPNbblcPqXKu+j/Xv3jGm7bdYep2SXzhmAzrAci9hVwIzdc6cHCoczqjgWgONPh8T8/MTdl8hhBAi0QYL6j47wPhM4OuR/34C2A40ANXAAuDDGD8n7wL2xWea2cGq8nVGYVEKZpI+rDJ1bn/igjowsnX9M6bHfd0S1AkhhMhoAwZ1WusH+48ppaYCPwHWAp/QWh+xuGYS8AhwC7AsflPNfFY96qYXFqZgJumjwCKo6w4k9uzXmrw8tne2R43JvjohhBCZLtbq1+8DhcC1WusGqwu01keUUh8FDkau/8yoZphFrDJ10wvGdqZu0cQSnvniuRS57BQ47RQ67TjtiV2OnpBnzsgd9pjP4xVCCCEySaxB3aXA1oECul5a6+NKqa2R60XE7fMWc/XEqexzd7Hf08V+dxfziktSPa2UKnTaWTQxue/B1HxzdvRQt0WrGSGEECKDxBrUlQGdw7w2HyiP8fWz2pSCQqYUjO3l1nQwJb/ANHaoWzJ1QgghMlusQd1BYLZS6iyt9eqBLlJKnQXMBXbHOiGlVCHwJeATwDTAH3md+4AHtda6z7VzMQoyLgAcwAbgP7TW8WjHkhCdviB3PrODIpedIlcuRS47lUVOPnjKhFRPbcywCqwPWjWFFkIIITJIrEHdHzD2yT2tlLod+IPW+kT7f6WUHfgUcDdGBezDsby4UsoGPAecDTwI/Bwj4/cJ4PfAfCKVt0qpmcA7QE/kfh3A54EXlFLv01q/HOPfLSnaPAH+vO5w1NiksjwJ6pLIKlN3zNtNIBzGMYbbywghhMhssQZ1dwOXAOcD9wM/UUrtBRqBKox2JyUYPereAH4c4+ufCZwL/Exr/ZXeQaXUL4GdwP/jZDuVHwKlwDKt9abIdQ8B24BfKKXm9c3qpQurHmxj9YiwVMm326l0Omn2n2xrEgbquz1MH+MtZoQQQmSumNISWusAcDnwI8BDJKgC3hf5XBoZvxu4InJ9LHrPyjpqcd/myGujlCoAPgjU9gZ0kevcGMHmHOD0GO+dFJ0W55paNeAViWVdLCH76oQQQmSumKMJrbUf+KZS6k6MrNo8jDYnboxs2lta65H+dFwLtAO3K6UOAGuAPIy2KMuAf4lctwRwAlb7+t6NfD498npppdMiU1csQV2UnlAYpRQ5NqsDS+JjSkEh69taosakAlYIIUQmiymaUEp9F2Ov3N2RwO2FyEdcaK3blFIfxMi2/aXPQ13AR7TWT0b+3LsBrd7iZXrHJg52L6XUTcBNANXV1dTW1o502sPidrupra3lH0fMmTpPR0vC75/OfrfVz67WEN4e8IU0gRB84wwXc8tzBn1e73s6EmG/+feOt3fuZPIBUz/tMWU076mwJu9pfMn7GX/ynsZfqt7TWFNE3wV2aa3vSMRkItzAVuBpjEKIcuBW4E9Kqau11i9hFE+AURnbny/yedAzn7TW92FU1LJ8+XK9cuXK0c984Htx3rNPMNkB/nF2mOgBn4KWHEAxe+okVq5clLD7p7sH9q+l4UhT1Nis+YtYOb960OfV1tYy0v9vrYcP8tSm6ERuuLyMlWecM6LXyxajeU+FNXlP40vez/iT9zT+UvWexhrUNRLZ15YISqnFGIHcV7TWv+4z/ghGoPd/karX7shDTouXcUU+d1s8lhLtwQD14RD1LUbgoqaBDgOr8wDZU2d1VFiiz3+dWiC96oQQQmSXWKOJ14GrlFIFo9g3N5ivYARlf+07qLXuVkr9HfgCRu+63kIKqyXW3jGrpdmUOGh1BJVPYRQJQ/EYr34tSklQV8gF42qYUlDA1PxCJhcUMGOMH9kmhBAis8Ua1N0BXA38TCl1UwJahvQGZFabqex9Pm/BWHo9y+K6FZHP6+I7tZGzbGzrO1l4PNZbmhRaBXUWBSXxVOl08cCK8xJ6DyGEECKZYg3qKoH/Av4TWKaU+gOwg0GWZLXWb8Tw+tuByzCqXe/uHVRKlWIEk23AXq11j1JqFXCNUuoUrfV7kesKgc8Be0ijyteDVlWVvpOVnbL8av77exKcqRNCCCGyTazRRC1G9asCTol8DEbHeI+fAdcDP4rsr3sbo1Di88B44Fatde9P+28AFwMvKqX+B+NM2s9jZPuuTKfGw5aZOq8Edb2s/v5dEtQJIYQQMYk1mjiEEaglhNb6oFLqDIwq24uB6wAvsAn4N631432urVNKnYPRCPnfOXn26xXpdkTYIVl+HVQqll+FEEKIbBNTUKe1npagefS9x17ghmFeuwNjWTatHRhi+XWsNx+2XH4NSFAnhBBCxEJOL08wT09P1BmjAFoD/r7Lr2M8U2e1/CqZOiGEECImEtQlmOV+Or8CLXvqeqWipclAtNak0XZMIYQQYtjGdjSRBJbnifYpksixKfIdgx+Hle1SWf36TnMjm9tb2evuYq+7k73uLh5ecT5LSsuTcn8hhBAiXkYU1CmlVgCfBk4FKoCB1g+11nrmCOeWFQ4MWSRhR6nEHVyfCVJZKPHbvbt5tfFY1Nhed5cEdUIIITJOzEGdUuou4Kv0HocwuDG/jnXI4jSJvLCdgE0RCusxv/QK1svPyVp+nVFYZArq6ro6k3JvIYQQIp5iiiiUUh8CvoZxBuy3gC8DC4FLMDJ2KzAqV/OB2zHOax3TrBoP/+/Vp3FpzQS8wRDeQCgFs0ovA539qrVOeBZzVlGxaWyvuyuh9xRCCCESIdZCiX/ByL5dp7X+LdABoLV+TWv9N631V4F5wHsYJ08ciudkM5FVocTUgkKUUuQ77FQUOlMwq/SSm2PDaY/+Ugxr8AYTH/BOKyg0jTX4vAm/rxBCCBFvsQZ1pwENWuvagS7QWrdgNA0uxGgiPGb5QyGOertN41PyC1Iwm/Rmua8uCUuw5Q5zUN0eDCT8vkIIIUS8xRrUlQBH+vw5AKCUKup7kdb6ELAN41SItKWUukopdV9HR0dCXv+It9u0qbDa5SLPLvvo+rPqVZeMYolSh8M01hGQoE4IIUTmiTW6aMLYL9erOfJ5JsZRXn05gKoRzisptNargFXLly//fCJe32rpdUr0pU9UAAAgAElEQVS+eblPwGULqmnrDlLotBsfLjsleYlvylySaxHUBQOEtcY2xquShRBCZJZYg7qDGHvmem0CPgp8ij5BnVLqVGAO0DDaCWayyfkF3DJrHoe6PWw9fozWHGW5h0vAt65ckJL7Omw2CnLseEIns4JhoCsYpMQiiyeEEEKkq1iDuleBFUqp+ZFzVx8Bvgd8RSk1AXgbqAFuwWh58nQc55pxZhcV87X5iwGora1l5cqVBMLhFM9K9FficODxRi/1tgcDEtQJIYTIKLHuqXsM2ACcAqC1PgB8HSOA+zhwD/BNoAzYDXw7XhPNBsc6vDS0++jwBgmFx3wLv7RRarEEK8USQgghMk1MmTqt9Sbg9H5jP1NKrQGuB6YD3cAbwP9prc2dd8ew76/aznNbj5/4c6HTzs8+vpRLFlSncFaizCIj1y7FEkIIITJMXMowtdargdXxeK1s1ukLRv3Z7e8h1x5rslTE20DFEkIIIUQmkYgiibosWnTIMWGpZ9XWpE0ydUIIITJMrMeEbcMolngdqNVaNw/xFNGHVVBXLEFdylntqZNMnRBCiEwTa0QxH6OlyS1wIsirBV4DXtdat8Z1dlmmq9/yK0CRK/G92DKB29/DG7ubaPUEaO8O0NYdJC83h69ePjfh97bK1MmeOiGEEJkm1qDubODCyMfZwKLIx62AVkptxQjwajGCvPb4TTXzdcry64DaPAFu+eOGqLHxJa7kBHVS/SqEECILxFr9+i7wLvBDpVQucCZGgHcRsAJYEvn4EkaQ957Well8p5yZgmFNoCe6R12OTZGXm5OiGaWXsgJzYNXqSU5gZdWPrj3gT8q9hRBCiHgZcZpIax0E3op83KGUcmJk724BrsEowlgaj0lmA6/FMaaFTjtKjqICoMCRQ26OIhg62b/P3xPGGwiR50hs4FvhcDLO6aI010Gpw0FproMlpeUJvacQQggRb6Ne+1NKzeTkkuyFQDVGM2IwjhUTgK/H3Gy40ClLr72UUpTlO2jsis6QtXYHmOjIS+i9l5VXsvayqxJ6DyGEECLRYo4qlFJTiQ7iJnIyiDuKcXTYq8BrWuv9cZpnxvNaBHWyny6aVVDX5gkwsTSxQZ0QQgiRDWJtabIPmNr7R6AJ+BtGccSrWuvd8Z1e9rCokaBAMnVRygrMlcBt3VKwIIQQQgxHrFHFNEAD9cCdwG+11hbhiujPKlMny6/RyvItmgB3m9vACCGEEMIs1hMltkQ+TwJ+CbQppZ5VSn1NKbVcZdiuf6XUVUqp+zo6OhJ+L1/IPCZBXTSrCti2JFXACiGEEJkupqBOa30KMA64FiOoOwRcAdwFrAFalVJPK6VuU0qdEu/JxpvWepXW+qaSkpKE30sKJYZWli/Lr0IIIcRIxRxVRE6NeDzygVJqHEafuguBlcAHgCsjj7VqrcfFa7KZzKqlieypi2a5/CqZOiGEEGJYRh1VaK2bgD8rpZ7HCOo+hdGnTgHS7CvCMlMn1a9RUrmnrjMYpNnvoz0QoD0YoD0Q4JSycmYWFiXl/kIIIcRojTiqUEoVAudzsrXJKZxczlWAH1g92glmC+vlVzlNoq9yqz11SVp+vWvHZv50cF/U2HcXLpWgTgghRMaItaXJZZwM4k4DcjjZoy6AEcT1nv36jtZazlqK8FoWSpj3kI1lpSncU1ftMvfC292V+AIaIYQQIl5izdQ9j9HSRAFBjCCuFiOQe1tr7Yvr7LLIhZPtfPichbh9Pbj9ITz+HhZPTHyBRiaxzNR5krP8uqC41DS2taMtKfcWQggh4iHWoG4txmkRtcBbWuvuuM8oS00vyWHlqZNSPY20VpJnztR1+pIT1C0uLTON7erswB8K4cyRZXIhhBDpL6agTmu9IlETEcKqxYvb30M4rLHZEtsCsdqVR5XTRaP/ZLI5qDW7ujpYUir1PkIIIdJfrM2HhUgYe46NvNzorJjW0B202JCYAIsssnVb2mUJVgghRGYYUVCnlKpQSv27Uup5pdRWpdTefo9fqZS6Xill3iQlxCB627woBUUuOxNKXHgDyQnqlpRYBHWyr04IIUSGiLmliVLqcuCPQBknK1/79+s4HfgO0Ao8M5oJirHl7186l3yHnfzcnIQvufYnmTohhBCZLKZMnVJqLsZJEuXAk8ANwDaLSx/FCPiuGe0ExdhSVeSi0GlPekAHsNgiU7e7yyiWEEIIIdJdrMuv3wDygO9prT+itX4YaO9/kdZ6J0aW7tTRT1GI5Khy5VHtckWN9WjNAY87RTMSQgghhi/W5deLgS7gzmFcewCYEuuEslGHN8g9G3w83biJQqedQqedcUVOPnvO9FRPTfQzraCIBl90u8Wj3m7mFktPQSGEEOkt1qCuCtiqtR7OelQPUBz7lLJPR3eQDY0hNjTWnxibWJonQV0ampiXbxqr90o7RiGEEOkv1uXXTozAbjimAU0xvn5SKaWuUkrd19GR2OOguvzmBrpWPdlE6k2wCOqOSlAnhBAiA8Qa1G0EJiilFg12kVJqJUbwt2aE80oKrfUqrfVNJSWJXVrz+M2Jzd7WHSK9WGfqPCmYiRBCCBGbWIO6BzGqWn+rlLLM2CmlpgL3Y7Q5+f3oppcd3BaZugLJ1A1Ka43b38PxDh9NXf6k3XdifoFp7Gi3ZOqEEEKkv1gjiz8BnwSuALYqpVYBkwCUUrcDi4EPA/nA37TWf4/jXDOW2yJTVyRBnaWnNtXz7Se34vH3EI50P/zoskn8+KOnJOX+VsuvsqdOCCFEJoj17FetlPoI8EvgeuCzfR7+ISebET8A3ByPCWYDj7/HNFbglEPireTYFF2+6Per/58TyWr5tcHnJRgOk2uTU/WEEEKkr5jTRVprL/BZpdTdwEeAJUAp4Aa2YmToNsd1lhnObRGUFDpzUzCT9FfkMr8vbougOFFcOTlUOJy0BIwl37JcBxPy8ukMBqlwOpM2DyGEECJWI14D1FrvAP5rsGuUUlO11gdHeo9sYRWUFEqmzpJVVXCXz7wnMZF+c/rZlESCuXy7LJMLIYTIDAlZT1JKTVdK3Q/sSsTrZxrLoE6qXy0VWbwvXUnM1AEsK69kVlGxBHRCCCEyyrB/aimlFFAJdGutLXs8KKVmAt/CKKbIxaiAHfOs99RJwGDFKqizWr4WQgghRLQhM3VKqQql1G+BNuA40KmU2qWU+nSfa4qVUr8AtgM3YAR0m4CPJmbamcUq0yTNh61ZL79KUCeEEEIMZdDIQimVB7wJzOVkZSvAbOABpZQdeAt4EeOcVwWsB+7QWj+dkBlnIKtMnQR11gocdpQC3SfH6w2G6AmFsedI9akQQggxkKEii1uBeYAf+BnwLkYPuquBjwH/CbQCU4EdwO3Sm87MuvpVgjorNpui0GE3ZTfd/h5K8x0pmpUQQgiR/oaKLD6EsS/uU1rrx/qMP6KUasII+iYCf4tcE0jMNDObVaGE7KkbWKHLHNR1+SSoE0IIIQYz1HrWfKCpX0DX697IZz/wLxLQDcwqqLMqCBAGywrYFO2r84dCHPd62d7RTlhL3Y8QQoj0NVRkUYyxR87KvsjnOq11a/ymlH2k+jU2VkvTyWxADHDl6y9xqNuNu+fkfddf9kHKpQGxEEKINDVUpi4HsMzAaa17O8K2x3VGWci6+bAEdQMptDxVIrkNiN09waiADqAtcsqEEEIIkY6knDDBekJh5o8vpqZAUVXkpMCRQ26OwmmXt34g6bD8Wu4wZ+RaJKgTQgiRxoaTLpqilPruSB/XWn8/9mklh1LqKuCqWbNmJewe9hwbT3/hXGpra1m5ciUA4bDG6OUsrBSlQa86q2XWtoBsGxVCCJG+hhPUTQb+Y5DHpwzxeNoGdVrrVcCq5cuXfz6Z97XZJKAbjGTqhBBCiNgNFdS9gRz1JZKs0Jn6PXVWQZ3sqRNCCJHOBg3qtNYrkzQPIU4oTINMXZnD3BOv1S9BnRBCiPQlJZgi7SycUMwNZ02lrMBBRYGDsgIHc6qLkjqHCll+FUIIkWEkqBNpZ8WMClbMqEjpHMpk+VUIIUSGkb4aQliosKh+bZWgTgghRBqTTF2CHW330uIOcNwTprHLR6HTTl5ujrQ0SXNWmbpWaWkihBAijUlQl2CP/uMw97yyx/jDm68A8KWLZ/Ovl85J4azEUKyqX6VQQgghRDqT5dcEszr3tdCZk4KZiFgU2e3k9sum+sIhvD3JrcIVQgghhkuCugRzW7TisOrDJtKLUspyCVYqYIUQQqSrtAvqlFLfU0rpQT6Cw7z2q6n8e/RyB8xBXYFk6jKCHBUmhBAik6TjnrrHgTqL8SXA14BVFo99BWjuN7Y+zvMaEatMndUxWCLaazsb2Xa0gxZPgDZPgBZPgC9dPJvTp5UnbQ6W++okUyeEECJNpV10obXeDGzuP66U+k3kP39r8bQntdYHEjmvkbLaU1fgSLu3Pe08sbGep987GjV29dKJKQ/qWqRYQgghRJpKu+VXK0qpfOA6oB54foBripVSaRctua0KJSRTN6TyAotjujzJDaisjgqTBsRCCCHSVaZEFx8DioF7tNYhi8c3A0VASCm1FrhDa/1cMic4EMugzpkpb3vqVFgGdUGLKxPnvHE15OfYKXM4qXA6KXc4mVtUktQ5CCGEEMOltNapnsOQlFJvAucAM7XW+/uM3wbMB94B2oC5wG3AeOCftdYPDPKaNwE3AVRXVy979NFHEzL3W1/x0D8WueeifIod0nx4MK8eCvLQ9uiihPMm2rlx8cklUbfbTWFhYbKnltXkPY0/eU/jS97P+JP3NP4S/Z5eeOGF67XWy/uPp31Qp5SaC+wEXtFaXzKM6yuArYALmKy1dg/1nOXLl+t169aNeq79aa2Z8+3nCIai3+Odd1yBK1cqYAfz3JZj3PzHDVFjl8yv4v4bTj/x59raWlauXJnkmWU3eU/jT97T+JL3M/7kPY2/RL+nSinLoC4T9tTdGPl8/3Au1lq3AL8GSoGzEzWp4fD3hE0BXW6OwmnPhLc9tcosl1+lnYgQQggxkLSOLiKFD9cDrcATMTz1QORzZbznFAvLylenXc59HQbrPXUS1AkhhBADSeugDrgKqAYe1lrHUnY4O/K5If5TGj6rIglpZzI81tWvEtQJIYQQA0n3oK536dXUm04pZVdKmUoRlVKTgZuBFowCipSxCuqk8fDwlOY76J/Q7PT1EAyFUzMhIYQQIs2lbYShlJoAXAGs1VpvsbikENivlHoS2MHJ6tfPRR77hNbam6z5WrE6TaJA2pkMS45NUZqXS1t3dOlwmydAVbErRbMSQggh0lc6RxifAXIYuEDCCzwGnAl8CCOQawZeBu7WWq9NwhwH5bE491V61A1fWYHDFNS1dic3qNve0c4bTcep7+7miNdDfXc3H5w4hS/MmZ+0OQghhBDDkbYRhtb6B8APBnncj5GVS1tuv7lPsgR1w1dR4GBfkydqrNWd3H11G9tauGtHdKJ4r7szqXMQQgghhiPd99RlNKvlVwnqhs+yWKI7uUHdpPwC01i9tzupcxBCCCGGQyKMBLp4fhUPlZ2B29/Duve2MmnaLGZVSdfu4UqHCtiJefmmsfpuj8WVQgghRGpJUJdA1cUuqiP7v/JbdrHy3OkpnlFmsQrqWpK8/GoV1B33eQmGw+TaJNEthBAifchPJZG2yvLNQV1bkpdf8+x2KhzOqLEwRmAnhBBCpBMJ6kTaqii0yNSloAGx1b66A54hjxQWWWZrfQcf+/VqfvriLt6pa8YXNBdCCSFEKsnyq0hb5QVOlILSvFzKCxyUFziYOS75exKnFRTyXntr1Ng+dxfnjatO+lxEYrV5Ajy+sZ66xi46vEGqi10snljCpQuqeWdvM2sPtLL2QCv3vFqHI8fGP587nX9/37xUT1sIIQAJ6kQaO2dmBXV3vp8cW2rPyp1ZWGQak7Ym2afZ7eey/3nDshinLD/X1Dg8EApbnlEshBCpIkGdSFv2nPTYHTCzsNg0ttfdlYKZiES699W6Aaur27qDpkbYAGfOKE/0tIQQYtjS46dmiiilrlJK3dfR0ZHqqYg0NqtIMnXZTmvNyzsaYnpOkdPOgvHmgF8IIVJlTAd1WutVWuubSkpKUj0Vkcam5hea/qE0+Hx0Bc2ZG5GZ9jZ5ONIWW0Xz6dPL0yabLIQQMMaDOiGGw5mTw5QCc4FGpi7BhrRmbUsTv9u3m3t372BbR1uqp5RyjhwbN5w1lcnlecN+zgpZehVCpBnZUyfEMMwsLDK1Mdnr7mRpWWb9YD/kcXPDmjej/i4/3bWV351xLiurx6dwZqk1pSKf/7x6Ed/Tmn3NHl7b2Uh1sYuXdzTw1Kajls85c3oFWmuOdfgocNopyctN8qyFECKaBHVCDMPMwmJeaTgWNbYvAzN1t7+3zhScauDHO7eO6aCul1KKmeMKT7TOOXNGOesOtFHfbl6a/cGzO9h53Gh98uNrl/DR5ZOTPV0hhIgiQZ1Ia1pr3P4eWj0BWjwB2jwBZowrZHqluSFwIs2yaGtSlwHFEu6eIHvdXZTkOmj1+1nT0mR53fbOdg5KQ2WTqiIXq754Lo9vOMKmw+0c6/Cx/qCxXL1m/8nehduPpf/XghAi+0lQJ9LaD57dwf+9uT9q7OtXzOPmlTOTOg+rtiZ1Xemdqfv70cN8e/MG2oPDO4XjuWNHkDa6ZuUFDj533owTf1713lG++MjGqGu21ksFvRAi9aRQQqS10jQ4/xVghkWm7qi3G6110ucyXP5QeNgBHcCjB/clcDbZY8kkc7X81vpOQuH0/VoQQowNkqkTaa3comN/izv5QV1Jbi6zC4spyc2l0umKfDgJAzlJn83wXDN5KhvbW/jDgb3Duv5gt4e7cgIs8Hmpcg2/CnSsmVKeT7HLTqev58SYNxhib5ObOdXm4F8IIZJFgjqR1qyCulRk6pRSvHjh5Um/72h9Z+FStnW0sbGtdeiLgZ2hIPl2+bYwGKUUSyaV8lZdc9T45iMdEtQJIVJKll9FWrPM1A1wlJMwc9hsfG3eYtP49IJCbpwxxzR+Ya4LheKd5kbeamogGA4nY5op89aeZr6/ajtPbarnQLNn2Mvpiy2WYLccaY/39IQQIibyK7lIa1ZBXavHn4KZZK6zKqvY+4Fr2dDWwpqWJhw2G9dOnoYNxbNHD3PMZ7TrGOd0cYU9n1vWvcMbTcaRWXOLivnLORdRnJudPdhe3dnI794+WYhTmp/LVy+by6dWTB30eUsmmoO6zVIsIYRIMQnqRFqrsArqUrCnLtPZlGJ5eSXLyyujxv9yzoX89fABdnd1cNjj4a6udlo9J7Nzu7o6efhAHbfOnp/sKSfFe/2ya+3dQYqH0UTYKlO3/WgnwVCYXDk6TAiRIvLdR6S1YlcuuTkqaswTCNHlk3NX42FSfgFfmbuQ2+YsZFtnO63avNy6urkxBTNLvHBYs+2oObu2dFLpkM+dWJpnyiL7e8LsOp7ebW6EENlNMnUirdlsigmleRxs6Y4aP9wa2+HrmcIXDPHw6oO8urORAmcON6+ciS8YZn+zh8sWVlNV5ErIfWcXFTO3qJhdXeYmuls72tBao5SyeGbmOtbpwxeMDmKLXPZhnf9qFEuUULsrupnz67ubWGSxNCuEiK+w1jywfw8vHKtncn4BN86cw/zioX8hy3aSqRNpb3JZvmnscFu3xZWZb9vRDu58dger97Xw8o5GPvKr1Xzy/jV8+8mtnHvXa+xpsM4E1TYc47YNa3jk4D72ubti7p9nU4q7l55OjTI3aOkIBjnc7RnR3yedHWwx/52mVRQMO3g9d1alaeyVHQ2jnpcQqRIKaw63duMLhlI9lSE9fGAvd2x7j7WtzTx25CBXvfEyP9q+mZ4sL+4aypgO6pRSVyml7uvokA3O6cwqc3K4NflBnT8UYnN7K680HOXPh/bziz07+O3e3XG9x7Kp5Vwyv8rysUBPmDuf3WH52CuNx3iq/hDf3Lyei197nhUvPcNfD+23vHYgS0rL+UFBGaeWlZse29qRfZWdh1rMX0NTKsy/QAzk4vnVprGNh9tpcUshj8g8B1s8XPrT1znv7tc464ev8MZu6yMF08Vf+n1/C2nNb/bu4sc7t6ZoRulhTAd1WutVWuubSkpkuSSdTbLI1B1pS/7y6+FuD1e/+QqfW/s2//7eOv5751YeOlAX13scaevmhrOnMVCy6O26Zjz+HtP42n5nujb6fRTYY69YVUqxuMQqqGuL+bXS3QGLoG5aDEHd9MoCZvQ7g1hrTEuyQmSC2/+2mX3NRva6rTvINx7fQk8oPbNevlCIXV3WyZhHDu7FH0r/TGOijOmgTmSGyeUWy68pyNRVOs372Zr9vrjew2G3ceOD6xho9TQY0qze2xI11hkMsttiL9wZFeblweFYXFpmGtvSnn1B3aFW8/Lr1PICiysHdtE8c1b11Z3ZWVgisle9O8ya/dENyuvbvewaYLtHr3BY09TlT/pxids72gkNcM+unp4TLZnGIgnqRNqbXGax/JqCPXUlubnk9kuhdYdCdPeYM2cjVVXk4mKLQKGv13ZFBw27LX5jnVZQaBmEDseiEnNQty1SLJFN+hffQGzLrwAXWSyVv723OeveK5Hd3jhi3U1gy5GBtyZtre/g4p++zul3vsyF/13Lgebk7bvd3DH4CTl/P3o4STNJPxLUibRnnanzJv0Hp1KKiiRk6/7pzCmDPv7qzsaov/uOTvN+t3nFI99SMKuwCKct+ltDWzBAvTd7ilO01pZB3bSK2DJ1p08rx2mPfq/au4M0yb46kSECPWHeqbf+xXSghtpaa7761/fYHwnkDrR0888P/oNgkpZrS3IdnFpWjsNmHcI8VX+IO7Zu4o3G4/jG2FKsBHUi7VUUOMjLja7K9AZDdKagB3Gl02kaa/bH9wf4OTMrmV45cHBxrMPH7X/bfKJCbWen+RvvaEr77Tab5fPXtTZbXJ2ZWj0B3P32JjrtNqqKzP9/B5ObY2PGuMKoMVeujaPt8Q30RXr5dd1OHtxfR1sg84P3V3Y00DVA28+BMnVb6zvZ2a8n474mD39bfyTe07P04UlTefzci9n6vg/z+LkXWV7zu/17uGHNm5z50iru3rGFJt/Y+DcpQZ1Ie0opywrYJm/yN/EmY1+dzab4n48vNWWA+np+6/ETJxdYBXWjydQBnFZWYRqrbTw+qtdMJwct9mROKc/HZou9F9+nV0zl21fO5/efPZ03b7+Q7f95BUsnS7+sWLV5AvzlH4d5+r2j+HvSN7vi7enhF3t28L2tGznzxVXcvO4dXmk4mnGtNA63dvPY+iPc8+rAxV47j3da/r94cbv194L/fXkP3YH4bUcZSq7NxqllFVw7edqA13QGg/yqbicXvfYcT9cfStrcUkWaD4uMMLksn90N7hN/Vgo6/Mnft2QV1DWNIKgLhY3li4UTirnmtEmm0wmWTi7lH9++hPUH21i7v5Vf1e6NevzLl8wmx6YIa82uOGfqAM6vquF3+/dEjb3ZdJyw1tiyoAmxVY+6qTEuvfYaarlcDK2x08cHfv4WjV1G5mv51DIeuWlFWh659sLxetyRfbRBrXn+WD1vNTWw5tKrsEeWAwPhMDs726l0upiQF9s+zWT4++ZjfPGRDYSH+BYaDGk+/9B6vvX++cytKTox/sI266DueKePr/71Pe657lTsSfx/95FJ0/jb4QODXuPu6aFqhPuMM4kEdSIjXHfGFC6aX8Xksnwml+czodTF6rfeTPo8xsUpU/fr1/fyxMZ6nthYz93P7+LyRTV84ozJnDWj4kTz22JXLhfOreKcmZW8sbuJbUeNCtdzZ1XyT2dOYW+Tm3XHm/GEon8zLrTbmTjKHyQrKsbhsuXgC5/8Lb01EGBzextLLfrYZRqr/XRTYyySyFQNnT7yHTkUuWJveZMoD60+eCKgA1h3sI3HNxzh46enX8BsFTy8f8Jk8u12/KEQ9+3dxYP762iJLM1eVDWef5u3iAUl6ZG9DYc1339m24mAbsH4Yi4d7+f8M0/jI79abbr+jd1NlOTl8vNPnMrqvS08tPpA1C/Y/T275Til+dv4wYcXJ+hvYLaichw3z5rHfXt3DVgV+7Ep01lROXgRWjaQoE5khEsXmBu9pkI89tQdbu3mf1462bQ4EAqz6r2jbDjYxiv/dgGufvsHHXYbT3/hXN7c08RX/ryJPY1dLPqPF4xvyuU9qPnRrz+3qGTU2TRnTg5nVY7jtT5LrjaMc2CzNaiLpUddJgqGwtzyxw28tL0Bm4J/u2wut144K9XTAoyK4f7+65kdaRPUPVN/mAf372G/x30iWOvr2knT6O7p4bNr3mRtv72nrzYe49XGY5xRXslX5i5MeWCx43gnDZ1+Fk4o5oazp7FwQjFvrF7HsqnlfOni2dzzyh7Tc65cPB6AR/9xiOe2Dr0N409rDnH9WVOZV1Mc9/kP5Pb5i/ns9Nls72znmLeb1S2NPH+snkA4TKXTyTfnLxn2a3n8PWyp76DVEyDQE2ZiWR5za4ooTqNfhAYiQZ0QMbDK1DX6Y2uE/MTGenr6rXsoBf/90VNMAV2vHJti5dwqakry2HGsT0+6AvNvpaPdT9frgqrxUUHdLbPnc9649AiuR6v/Jm8Y+fJrprj/zf28tN3o3xXW8OMXdnH5whpmVRUO8czECoc1Gw+ZK7irS1x0B3rId9jRWvN2XQu/rK3jo5OTu3ftraYGvrjh3QEfn5pfwPLyCv5751ZTQNfX2tZmPrH6dX6wZBmfmDojEVMdlt4+l3sa3dz+t80ALKrM4Wbg1AH2gq6cO46eUDimxtqv72qKW1B33Otle2c7e7o6OaW0bMDAeJzLxQWuGgCumzqDJr+Pn+7cyrnjqilxOCyf00trTe3uJn7/9gHe3dtCoF8lr1JGEdunVkzh0gU15Ixg/20ySFAnRAyqXOaCjeO+2BPVuJAAACAASURBVIK63h+sfX3u3OmcNdNcnNDftIr86KAu3/wDbl6cDrVeWVVz4r8XFpfyxTkLBmwhkEl8wZDlGbrzxycvq5BsWmvuen6nafydvc0pD+oG6jn54m3no5RxisrPXt7NPw4YDbALeux8OInz+9/d2wd9/IbpszngcXP/vuEdGfitzetx2Gx8ZJDN/YnUG9QFek5+75hfbvy7Pn16OaX5ubR3nyyH/aczp+DKzWHt/lY6vOYy2RybYk51UfT3JTixXWQ0jnu9fHfLBl5qOArAdVOmc9PMOcN+/jinix+esnzI6+oa3Xzz8S2sPTBw/zut4a26Zt6qa+aUSSU8/Lkz0zJzl/nfoYVIohqLoK4xhlL5Yx1etlj0frr+rGnDev60/q1OCq2Cuvhk6qYWFLKopBSHzcZPTj0jKwI6gE5vkPNmV1JZeHIpvabYxbgY25lkkt6AqL/NgzSXTZat9eYf/qdPK8NmU/xhzSE+ef+aqPm/driHNk9y+hltbm8dtJXPtIJCPjltJpvb25heMLzgWGMEdke6k9est1dPKMza/ebAZX65sUJQ6LTz42tPoTTfCFbOmlHB1y+fB8ArO82/jObYFLdfPpfvfODkHpDcHMXCCcWj3s5Q23CMy19/4URAB/Doof189O3XePbokbhVGz+1qZ4P3vsWaw+0DH1xRGWhkyJneubE0nNWQqSpaougrsnvI6Q1OcPYx/ayRZZuXk2RZYNlK1HfKHM1yhW9/GoDFsQpUwdw1ymns6mthblxChTTQVWxi99/9gzAqLrcUt9h6lk3Glprmtx+/MHwsP+/JtpjA/QP68287Gtyk++wU13sPFGoE2/+nhD7mjzsaXRT19DFlUsmMLemiK1HzYHlwgnG19sHFo/nR8/uwBM4WbDTE4Znthzj0yumJmSeff1un3l/WV/fmL+Ew90e/np4P7v6HdX36WkzmV1UzK/rdnG0X+NufzjM3w4f4La5C+M+58FsO9pJV7+v9UKnnanFJ39hu3RBNeu+dQkef4iS/JOZqGtPm0SBw85bdc0EesJcMGccHz51ItMqC+jyBfnxtUtYMKGY2VVFOAZpxzQcLX4/t65fTbdF4+D1bS2sX7+aSXn5fGfRUi6rmTji+/yqdi93vbgDpgfAZ4P6oTNvhU47//XhRQn7dzJaEtQJEQNXTg4lubl0BE8uQ4S0ptnvswz4+nvRIqiLpQgkat+XRZZudlEJ+fb4/bNeUFKaNlV7iVBV7OLi4tG3Odh5vJPfvrmfvU1u6hrddPp6eP/iGn75yWVxmOXobD/ayZ/XWR+bdLDFw/97eB0vbDO+Lj++fDI/vGbxiPr19fXUpnomleWxeGIpDruN76/azgPv7I9qoTGuyMncmiLLZbqFE4yl8LICB9edMYXfvrU/6vEnNhxJeFDX5PMNetzUFeMncmnNBFoCflM2b0JePt+Yv4Q8u53rpszgM2ve5J3m6OP9nq4/xJfnLEhqcLB6nzkbdcb0cnJs0VlDe46NkvzowGx2dRGzq4v40sWzTa9R5Mrlo8snj3p+HYEALx6v56n6Q5YBXV9hRvcL7C9r67j7zR2w1I/K0+hwCDps4DayluOKnJwyqZQcm5FNrm83ttl8/X3zGF8y9Pf6VJGgTogY1bjyooI6gAafd8igbtfxLt7Za/6mGktQF3WMVaH5m94ppeZzW0XitboD/LVfNux4R+o72O841snH7zO3qei1u8Ed1Z7iz+sOs2hSyagCpkBPmK8/thlfMIwr18bSyaW8u8+85Len0Y3Wmm0W2xF6M3UAHzltkimo23Conf3NnkFPXhmt544doceiPcZVEyazqLSMz06fjVKKSqeLO5cs42ub/oEGiuy53LtsBXmRX65ybTbuOmU5573ybNTr7Pe42dbRzqIk/pvdZFGQcvbMCgglfym4v6Pebj7w+ku0BQdeWr9m0lTmF5cyp6iYsyqryB3hlpDH1h/h7ve2wpIgKvISygZ6ToAph8r55hULuGxB9YlfbsJhzZt1zfx981E+eUZ6VGQPRII6kTFu/dMGmrv8dPl66PIHaevysuasHgqSvLehypVnWmo57vMyWMG81ppvP7mFUL+q1+piJ4smDH9ps6rIiSvXhi8YtszULSlNbrsRrTXHfN60bLCaTDUl5mxfQ2dqj5DSWvONx7fQ5Yttafl/X97NNadOHPG/q/eOtBtfn4AvGLYM6AD2NLhp6PTT0m9/nCPHxuzqk/vTFkwoZl5Nkali+YmN9fzrpcPfNB+r54+Zl6xvnjWP2+eb+699ZPI0zh1Xza7ODpaUllPar9JyUn4BZ5RXmqpjVx09nNSgrjfb1NfiiSV40+Cghbt3bBk0oPvL2RdyekXlqO+z6XA733hiC0zUJwK6XipPc/kVFVyxqCZq3GZTXDBnHBfMGTfq+ydadux8HiGl1FVKqfs6OlK/WVgMbc2+Vtbsb2X7sU4Ot3pxB4n5B1Y8jKRY4qlNRy03q19/1rSYlrpsNsXU8kh2oj0H3ZqD7vN9MFlBXYPPy2/qdvG+11/k0teexz/GDs3uzzqo85mC+GQ61NrNpsPmzMxQmt0BftcvMzaUYCjM4xuO8MF73+KjvzZnBmdYZNT2NHZZzm9uTZHpJIlrTjPvm3pm81HTWLy0+P2saTG377hywqQBn1PtyuP8qhpTQNfrgxNPZnguqhrPl+YsID/HuoVRohzrMAd1E0pTv5TYHgjwnEUQ3WtBcWlcAroOb5Cb/7DeqPw9lIvuModADxzYw47O2P/dpIsxHdRprVdprW8qKcmeTeDZrNhlzhy4/QOcRJ1AfZdZFVDldBEeoIt5rz+uOWgam15ZwOfOmx7z/U+cfHAsF3Y44R95fKf6NH6z/OykFDR8Yf1qzn7pGX60YzO7ujrpDoXYnsHfBOMh32E3fX32hDUt7uRl6wI9YX5ZW8dtj27kiY1HLJf6e505vRzHIMc4/fr1vbyw7TgvbW8Y8hxWbyDEdfe9y7/+5b0Bq2mvXT6J3JzoX16a3QFetDhuaskk89fw1Usn0n/r2b4mD0ctMk/xsKergwJ79Kb5yfkFo9rDdcX4SSeKqV5tPMbW9jauSWJbE39PiGZ3dCZMKetfSJLtyfqDBAapZr16UnyWPDu6gyePZNQK9jjQ/W7bozXf3bwBPcT3dCvPbTnGF/40sufGiyy/ioxRaBHUdaYgU/exKdO5uHoC1S4X45yuE+c9DsQXDPHeYfMPu+9fvRCnPfbf1E1tTVC0t/fwz2eMvAosFv5QmP7ffje1tXJq2dB99lLFFwyx8VA7c2uKTOfsxktNiYtOX/TxScc7k7ev7it/2cTfNx8D4MlN1lmsm86fwTffb7SfuPWPG/j7lmOW13kCIf7fw+sBmFNdyGM3nz3gsWKPrD3E+oPWLVN6nTOzkic31puOl3p8Y73p2jOmm7PN1cUuFk0oMbUDequumY/FYYN+fysqq1h3+QdZ3dzIc8eO8NLxeq4YP3FURQ0VTicfmjiFx44Yv+C92niMhn94WXX+JUkplrDa4zmu0JnQ83Xd/h5qdzXygSUTBrxGa82fDw2cGVbAByfEJ6ibUpHPE7ecw72v7uEXtXsJeW1UdxbSWBr9dbmurYVnjx3hygnD+9ryBUP84NkdPLTa+H970bwqUnXuzpjO1InMUmQR1KVi+XVyfgFLy8oZn5c/ZEAHsO1oh6k7+bgiJ+fOGtlygtUZpQcsjr1KFKtjwt5rH7hpZzrYebyLT/zfu5x2x0ss/6+X+eT97/KL1+rieo9qi4zHsSQVS7S4/TxrEaBV9Atgz59t7AkKhTX2nOEFErsb3Nw7yHtl1b+sr3xHDgsjrS6G48zp1r8cnGPx7+WtPQP3kBsth83GBVU1/OiU5ay59Cq+OHvBqF/zu4tO5ZpJU6lyurhgXA2/PfNc3D09bGxrYf0g/fDi4Wi7+WtxfIKydOGw5q/rDnPhf9fyhT9tZOOhgYP+7Z3t7OwceAvUx6ZMpyYvfkvEDruNf71sLo/dfDZLJ5fyyPvPZUaB+WvzR9s3D2tbyb4mN9f88p0TAR3AHc9spyuQmmydBHUiYxRabNx2pyCoi5VVFmPZlLIR/3Y+zeI4q4Mtyates8rIbWobfuPOVNjdZ5N9s9vP23Utg/6gicV+dxfXvvUq71QehoU+cJwM4BuSlKnb3eDGasXnoRvPoParK/nhNYu5eukElk01NuXbFPzs40v5wYcXD6un2G9e30eXz7zVwRcMsW6Axsa9zptdiT3HNqyTK6ZW5A+4HHjebHNQ93ZdM+Ek7Fu022wU5Y7+9IDi3Fx+cuoZPLjifEJa86E3X2HJ809yzVuvcveOLXGY6cCs9tMlojXHxkNtfPiXb/O1v22mqcvYfvCr2r0DXj8pr4B7l63gxhmzObWs/MQS9aXVE/javEV8d+HSuM8RYOnkUp645WxmVBTxnYWnmB4/4u3m4QMDz7vX/mYP2/udpjGxLA9PMDVBnSy/ioxhtfxj9YMm3VgGdVNjq3jTWp8IAlOdqVtcUobC6Izf62C3h6eOHOLKCZOGlb1Mtl0Wx4LNrRle5mgwYa35wvp3jT2FClSpRi/ww3su0IpjHT6mJGHL0v5mc1B/2YLqE61BplX+f/bOO7yt8mzj96ttybK8994rHomzp0kghIS9CoVCSwkFPkb5WgqlBUoL9CtllLaUUUpbKLNAIKyEDCdkb8dO4pE43nvJsiVZ63x/yEs6r2TZlmTZeX/XlQt8dCQdHR2dc59n3I8CN42xYhg+lm5eGI+CuEBc9vJ3477HuwfrcdfKFJtlx+t7MWji10IJiHW+bGSADD+7JAMAbDpaHbEg0XHSal5CEKQigc37dQ0YcKi2G4uSfTf1T0MsEGBPp22E82z/1MdqOYMWNY4KdP/BebJRjVK72sqSyg6odUao/PjncJVEgvXRcSOpzkGzGX1GI8Jknv/hDP8OVkVEYUVYBHZ32H4nr56twE0JyVA48f5cnRWB9XlR+PJkC/zEQvzvJem4fUki9ny326Pb7gjfO/syGA6gpV/dOQnAE3Ach6N1/CaCuRMUdQe7OlC842s8fOIw9va1QSwHxsoqtc6IXq13RicpxWKkKflzUh88fhAXl2xBWa97ImDupIoi6tIjpi7qvuto4zWJEAUHJFpvNtq8lH6t6ejnLXNVtGZHB+Cze5eOu94L31ah0s5WZP85fsrw8vxo7Pp5MT76yWJ8+9AKpA3tZ1fSrwudiDOZWEitt7vvveMea5jwFAlyBcR2kfpugwFdg55rrKHto2gPROq+vzCeN3LPYLZgSzm/KYaGVCj0iqCz57GcfJ4g6jIM4t/nxy/TeOLybKzPi8LWn67Aj5cnQ+TBOsXxYKKOMWOgzdqbjkaJidA/aEJauD/8xKMNERKhALkxExsef6i7E7UD/fiooRaPnDwCU6EWSLGNUnq1rs6BdUrtQD+u3bMdm5t8wPhqDPZiBHBPpO7tWvoJn0SbAKXZazV1tEhdcpjrxrw50QGQjpOGHTRZcMNr+20Gt9M6bJekhCAuWI75icE20fXEUDmEFPue9XOiRixPFlJE21hodagdmkHc/c7RSdvH6I1mvHyoAo9tK8WReu+UEYgEAiT7848/T0braI0SnojUiYQCrJ8TxVu+2c6CpqZfg79Wn8HHDbUwT2O36DDpSpWN7cwwr52rQL/JeUYoXCnDX2+e6xNjAZmoY8wYaOlXX6+pU8rEeG/jIpQ9eQm+uG8ZfnNFDn6yMnnCXa+HuyhF1AO2F0hv1tXlO+l0NXIcHi09ij6jb6TGewYMaNfYRkBEAoLkUNcGsDuiUTuAHW307lEAQIRpwt2vZo7jXUBcsUeooYm6CXw+kZ3hryPUOiMefP8EzBYOA4Mmqs/ckhT6sSEVCalD3n+0LBE7frYKhx9bg9gg55GjG4rioKSUtpU2qrGrqp3/gBMsFg5v7K7Bome344WKU3hXV4Xrju/Ayi1f43enTqB+gB/9dCeplGj3WzXVHrPDaKaJOg+Nu7o8n9/t+l11J6qHIubHe7pwSckW/LGiHD87cRi3HdjtE16XD2Tk8GZ4q41GvFdXM01bNHGYqGPMGGiWJtNdU8dxHPqMRpzV9KFj0PEFXCQUIDdGhduWJOKhoRojVzFZLDjWQxF1faPCMC9WhUg3zDB1lfnBzjt3B8wmfNfhWrrF09BSr0mhiikPHe82DOLvC5Zh9+rL6CsoLGhV68e9SHcPDmJTYx0ePHYQ87d8jpcqTwMAdCYTnjtThkXffoE+J077RrMF9d38KG3SBCJ1AJAZ6Vr0uLJNg89Lm3CioRcmu+hYTKAf4p1EK2gp2Oohm5MwpXTc5qEghQT3FMggokT8tp2ZmKj7z6F6PP3VGWvZgtJap0cEQL2hH2/WVEPnYZGRq+KXYGxpbcI/aqo98n4qPxEC5baKONoDkToAmBsfSBXod/77CNr79Phz1Wmb6NzeznY8UX7cY4K2s3/QpYaaRIU/ro3lj8h7s6bKJ0SnK7BGCcaMwVcsTQDrHfXbtWfRpteNDJ5+NDsPV4Ul4KOjjZCJhbhxfhy1Y3einOrr5Q239iNCqMR+uHpVLK6ZG4PUMRfLVrUeAwYTkkMVHvO/SlMGYE1ENLa1OXb1L2lvddnnyZPQmiTS3ZB6zQsMhtFiwe0HHTQZyDnoTCZoTaP1RY3aAcTKR8XWjrYW/PjQHpumk5L2FvwqJx+/PVWK9+prhpa1UlNDgHVyhH3qMdRfigC7yHa7Xod+kwkJCn9eNAIAMu32iUQowLo5kTh0vpuXRv7pB6XUbZmf6LyrOy3CH9+csl1W3T6xiFhWiBC/vCwLT31hFb8iAcHC5GDMjZ9Ynep7B4dKBOQciN3P1F8kotaNupNr4xLx1+oz6DfZnsP+VHUKNyYkwV809W7bsby/cTEAQGswoVWtR4taj3ClZ0QdIQRX5EfjFbuu19ouLe57/zjKY/nTOgwWq/+lJ2ZsXPrSbvQPmpAYokBymALJof64e1UKdRTeT1Iz8VFDrc1vsk2vx6ametwYP3GzeG/DRB1jxkC1NJmmRgmd2YTzdumZ8+p+rP73LmiGtukfe87jg7sWITZoanUWOykpvkWhYfj7L5bxRow9v7USf915FhbOWp/01g/nQy7xzM/8laLF+LD+PBq0Axi0mPFPu4LikvYWWDgOAi8YqzqDVk+XadckwXEc9nS2o3twEMURUQhw0b7iqfIT2NdJjxARAcDJOPTorZeHDr0eq3Z8jeyAQFwfn4grY+KRR5n7ea5fg4o+NT5qGDVk3dra5FDU1XSMX0/3ctVpvDwUHSkKCsEbC5bxxlktSg7BvcUpyIi0zlpNClVALBTgQE0Xvvf6AYf74LdX5eJcez/OtPRhnpPuVQBUWxNaJHU8ri+KxcnGXqzOisDKjDBwHFDa0Is/bavG8YYeaPQmfHz3EofP7xkwjNpQKPkRmHS5yuPHbZhUhv/Ln497j9qOVdOYTDjS1YlVEfy6NHcgl4iQHOaP5LCplR+Mx4+XJ+ODww282b4H2ztAovj7/I8F8z2yz9U648gkjYpWDSpaNRAKCB5Yk0ZdP8lfiXVRsfjKbmzZ62crcX1c4rSfz8aDiTrGjMGXLE0iKPNfDzV3jgg6wDo8+5a/H8THdy9BiL+Ut74raE0mavfVorBwnqAra1TjzztG1z14vhtv7a3FvcWpk3rv8RALBPh+YsrIdr5bV2Mz6qdzcBCn1b0TGlje2KPFnupOJIf5UzsdJ4O9hxTAj9Q9cOwgNjc3AACCJRJ8vvxixMhHxfgXTQ1YHREFvzHWBgaLBWaOg1IkRoRMhrP9FHGisKBbb90nnzTWwcxxKFP3oKysB0+fKsUtiamYExiEk3Ydw8+dKYNpTCpqV3srTqt7ka3ij6k638mPdI2dtXq0uxMvVo6Gx470dOHt2rO4L93WTDc3RoXcGP6IrkXJIViZHoZdVfzoCgAoJEI8eUUO9TF7aOnXsxOM1AHWc8FL3ysEYD0HFDy1FfbZNbXWCJWcLs4P1IxpiAjgW7JEEu8UvF8WHYsNLXH4YujYG+Zwt+dEnbcIVkjwrx8twHWv7oPeOGYfq/j7e01EtMfEEq2JKD5Y7nSSxk9SM3miTm00oFE7gHiFZ8XwVGE1dYwZA232q2aaInU0Udeq49fUqeQSBFC8mVzl/foa9NjVU4kIwWVR/MHiHx/jD8TeXNo87uxOdyAXibAoJIy3fGe7k0aCMZjMFry0rQornyvBI5+U4YbX9uP5rZVT3i6N3kidR5odNZpaO9zVOSLoAKu1xNOnS7G1tQl/qjyFldu/wn3HDuCy3d/a2LVIBAI8kz8PJ9ddhW+LL8WPkih3/nILzvVaL2JfNNt2BA9aLJAJhSgO51+8d9jtt36TCY+XHeOtpzOY8cxXFbzlYyN1vzvFT5UeoAyrd8bP12ZQO1cBawG8qySHKWD/Mi1q/ZRuzpQyMZJ4o/OA/Ke24pnhmjk7bLp2lXyRITd4ZpQcjeVhEbxlRzw8XcJb5Mao8NDF6bYLA/jno4WUc4e7oNn9JFOOl7HMCQzC4qFtmqMKwvMFC7B3zXqfF3QAi9QxZhD0RonpEXWRFFGnhRH2P6m/3FQ46dmKtQP9eKnyFG/5FTHxNnVZw/z3KF/UVbRqkPGrb5Afq8ILNxYgxYMpl1XhUdjd0QaJQIAFwWEojojEmgjHMx+H0RvNuPPfR3ji4NVd53DbkkSETjLKCQAHa7p59WYxgX42Rdz2URIA+LqlEV/b3anXDvTjmj3b8dOMHNyVmsmrS8sI4Ee5EG7GDnM/Fm7djHZKI811cYnoMQziT1Wnx/0stybaRlw1eiM2vFsC5OkBMQdYAAwIgFrxSOdr/UA/TlBGuDlKGTsiN0aF56/Px4MfnOA99l21daqDfeSYhkwsREKIYiR6EqKQIDXcH71ao8PZsq5QGB+Ec5Q09Ou7a9Depx+J6vUPmvCvfbV4+8DQSCchB+Jne3xwHGDiN/V6jCJK01FpbzcGzWZIhZ6oMPMuVxRE49lTJ8EFmQEDAQnii+gFIZMbmegKrpQn0Hg4Kw8WjkNhULBXZvO6CybqGDMGX7I0CaeIOk7CwWoIbD0BRARIx7VocESTVoub9pVAY1dETQDcnZrJW7+hW+u0vrC0UY373j2OL+9f5rET1LqoWMTKFVgSGu7Ugd2eTcebeIJOKhIgPy4Q3QOGKYm6vRRz3KWpISP7gOM4bGvlD5V3hInj8FxFOeIV/thg1wSSSRF1RMqhH2b0D/KjE4kKfyQq/BEnVyBK5ocWvXMD3SVh4TZ/3/LNHtSH9cLm25SbwSksyIiypjlj5Qq8Mm8x7rGr2wIAvdkM2QREw1WFMbg4OwI5T2wZWsIBUg6d/Xqcae0bmV4xHo9fng25WIjUcP9JlyXYUxgfSL2pAYBNJ5rx6w3ZCFZIcPc7R22PNT++wICeoKnbcybA9iQp/BEikaLLMPqegxYLTql7MTd4Zk3KoPH3hiog1mQ9ThX8DlR/kQjZAfyyAndBS++7Uk9Im3E9E2DpV8aMQS4Wwl6P6IxmGM2UE7OHUYpEkNtdEIkANrdJ8xImP9/16dOlaKVc5NdGxlD9rXZUOI+8qPzEKIgPhNbguVRspJ8fLo6MnpCgA/jbvjg5BJvuXYoP71o85akP+87yzWTHDoavGehH8zhiyp6VYZFYT0l/p/kHYCLf9nCERkgINqY6t7nJUKoQJrV2KnIch2cPl+GkkJ6iI3IOPcQaFRQQgnXRsYii3ITUaydey6aQiqwzWBUWYK4epEgPLNDhw4q68Z88RHFGOBYmh7hN0AFAYZzzus2TTWp8Xd7KTxXL+CIjO0iF31zpWo2gOyCEUKN1h2dBCrZuoB//Ou/comVeUKhHRwvSp8n4fhp1slzQoo4Qcjkh5HW1ml9zw/A9BAKCJSkhWJ4WivVzorAiVoQfL0uatJP8VCCEUOvqIBndlolaLAyjMRqxlRI9CpfK8Otc+nDr7eOIuuvmxWJlehi1hX86sVg4HKq1TQ/ur+nCy9urXfKVOljThVd3ncO+s/wLYLtGT7UzWZIyegFNVvjjyxUXY64TM+WxJCr88YeCIqpY9xOJkDiBmpuxXn83xSdThdcwy8ZE6X5RegSvt/Lr6MZySm2bP6RtV+0kzXVXpIUBSYaRtCURA+9pqp166dnjbj+y8S7SJxvUuCgzHD9fm2HbRU8RdQuiQpEV5Vk7E3toos4ddXVtfXrc+59jeGlbFb4qa0F1m8arN8EvVp6yafihsTI80mPvrzeaUUsxZU9zw4hAX8W3zvBehuO4zQA2FxUV3Tnd28Jwjf/8eNHI/5eUlGDVqmwna3uWCJkfz9YEEg4Y8oGdN8H5rsMc6u6gjs15d8kqRPvxu/Kae3XjTpN4c895rEj3XDHyZKls06BXa1skLyDAyzcVjluj9Z+DdXjs0/KRv395WSY2rhgdOL/1VBvvOVEpIvy9oRL3pmUiQCwBIQTZqkC8t2QVFmz9HGrKFIwb4pMQLpUhQuaHDdFxPCuQsWQEqPjHhAPGXsilQiHuTcvCryjNEACwPGz0wrcqPAofNdQ6fW37ebQJCn/st2uOmKyoK0gOAFHbCgOjwII/VZ7m3XQ067Q409eLdr0ewRIpytU9+KK5Af0mI25NTMV9aVluKQcQCQVYmBSMg+f59YMAcLKxFzKxEPcWp+KGojg88vFJ642QjC9wEuTej+LQRN3weYDmKegqFa0afFnWApSNLiuIC8QmF2b9TpUj3Z34fJxxgcl+StyckOyxbTjb3s/rio5WyXgejrOJC1rUMRhTwVmkTiISuFxjZM/eDn7U7bq4RKRQZkUCgFwipM51HEthfCBWpHmuGHmyHKyhp0dpzSVnWvpQ2apBflwgEoLleGmbbVrnpW3V+MHiRMjEQpgtHN74zm60j4BDX6QWr5+rLulNNwAAIABJREFUxEf15/FARg5uTkiGWCCARCDA2qhYfFh/3uYpG1My8Gh2nsufZ1FIOPZ3tmNNRDS2t7Wg10H0yl8kQpJd9OyG+CRsbW3C7g5bMTo3KMSmQ3JtVAyKgkOdRnJO8yJ1/GOntn9yok4rpneq/qv2LG5MSEK6UgUzx+HR0iNOxeeLlacQJpXhJjdd1C/JiXQo6kobe8FxHAghCPATYWVGGHp1RtSFdqELtiUJCdPQ4ZijCoRCKMKAebQuts9oRLm6B/kO5iy7QjUlUj2RmcCTRW82494j+0GL0XE9AkAnAPoE2Lgq26PNINTUqxuMx32ZCzr9ymBMBaqok1pPY3kxqkmPodrbyY8wraDYHgwTKJfgqXFqgB5em+mTHVy0izBtqPs7B+pw2cvf4cEPTmDti7uxvaIdNy+It0mTaQ1mHBp6PaGA4P2Ni3B1Yczoi0SboIX1otljNODJ8uO4tGTryIzaB9KzbUyH56iCcF961oQ+z43xSTh8yRX4Y+ECHL/0SpxedzV1veyAQN73IRYI8PcFy/CPBcvwvxm52JiSgafmFOI/i1faeHgJCcEb85fi/vRsXB2bgMey83mvX9HXaxPttReQAFA3iZo6AChV91CXmzkO/3faGhLa2dYybjQRsAo7nck9zU63LkpAcQY9Gt3ZbxiZfSoVCfGDxYn4+O4lEFLs6KZD1IkFAqqtx/4JdinbMzyCbSxTrVN1BZlQiOcK5kNpV1/r1yjDUkMMfhyTgRdWz/N49oBWfpExhc/PcRxOqa0ek/+oqZrKpnkMFqljMCYJzdZkOFJ3SY5jEeaMdr0OVRq+We7i0HDK2qPcOD8euTEqlDaosSApGE9/eRo7KzsgFBD88rIsLHYwZH064TiOLuqSbbd1YNCEZ786g2GNYjBb8OzXZzAnRoUzdsbCu6s6Ri4UUSo/vHhjAW5bkognN5fjVGwb7OVDir9yRMhF+8nx9cpL8HlTPQLEElweHTfhUU323aQfHmpEuEaMdqVtdOtSSqMFYL24F0dEoXgc49lAiQQ/zbAKeY7j8HLVaWhMo++hNZtRN9CP5KHoLk2o1NLMkl2glGKRMsyO9hYc7urEHsqNCY2OQT3+VXsWP6F0dE8UiUiAt364AA3dWvzon4dtx48JOByq68LVgaP7XWsy8WxmBABi/eQwmS34575aHK7tRn5cIO5cnjxpayJXWRIajj2dbZgfHIoloeFYEhqBORMw7qZR3c7/jtMoUz08wYrwSPx32UW44+AeNOq0yFMF4cPLir1q01JFmSYzGVHbYxjEu3U1+KyxHtX91nNOnFyBHyal+dzNMhN1DMYkoadfLQhRSPD9hfyh0K6wl3JnnhWgQqh0/BmNOdGqkZTvWz9cgOZeHYIVEsjE3ve6atfrUKbuQXlvD8rUPbgqNoFnAVLV1o9uuxFCMrEAebG2aev3DzdgwK5rt6ZjAFcVxMCe3dV8U92CuEA8eG0KfnjI1tBXRAh+YZdajfaTu0VgDLO5tBltbUKgwAgy9DX4C0S4Nm5yxweN4brAg3Y1c+XqHqeirlmvm7CtCcdxKO1xLOoA4PHyY/CbwGu+erYCtyamTrhr2hFxwXIsSQlBdZcGiDABYWYQfwserj+AelkWHhgSw7RIZbSfHFKhEE98Vo5/7bd29G451YbGHh2euXqOW7bPETcmJOH7iSkT+j6cwXHctEXqRt5LqcKny1fjibLjeDK30KOCbt/ZTvxhSyW0BhNuWhAPo9mCnZX880HGJNKvBosFz1eU26STG7QDONbThXmUesjphIk6BmOSaPooaSMJ53BQtCusj45DtJ8c+zrbsbejDSd6u7E0dHJRv+jAyXnkTZVXqs/guYpym2VqgwHro2Jt7mr3UjpW58YHQSqynvj1RjPe2F2D57+lpzlo6e2qtn60qHWIUtl+dvsJDYDVV89RnaK7SA5T4EhdD1AmAxdpAizA7ZnWJg13kh3AF3UPHDsIERHg0qgYyIRCRMv8ePYtZ/p6Uehi5y8ANOm0Nn5qNCr6JuYmoDYaUdLegvV2on8qxCZJoRKa0MeNRi9N4PBS1WnkBgZhdUQ06gf4zUUJCn9Utmrw7wO2Fi2fHGvEb67I8Wi0bqJR4fFo6xvkTdyRiQWI8fJ5IVQqw1+LFnv0Pfaf68Itbx4caYr4zWa6mTch9PnD4xEh88OS0HDeTfempnqfE3Wspo7BmCQfH2jmLRNKgVsWTT4KIxmqrflpRg7+u+wiHF97Je4ax8PMVbQGE47X9+Ddg/U4XOs82jIVaIXdR3q6eKOpaKJuaWooOI7D12UtWPPCLoeCDgDOtffbjPsa5rsq29flOA472vii7uLI8addTJURk9MBAXBOApyXoLPL/YbZtJmwAHDv0f0oH6qBo0Xr7j96ABpKx68jTvb2IIs2OWMc/EUinNtwHWo2XIdrYvm/j6nWjo3li+YGPFt3wkbQjeXlqtPgOI7a/avgxFj70m7YN5/rjRZ09nvPkNgd0JoEUsP9XZr8MZNo69PjvveO8bpcaSSGKCadubiKctx+2dQAo8X7PqnOYKKOMaPoHzThZGMvtp1uQ0mDEX/aVo2PHTjJe5J9ZztxqJIvjCxiDtUD/Jo4RwyazejQ66kWJgCgFItdSr0644uTzSj+YwlyntiCq1/Zh19+WobPTrg+RWGiLA4NR4aSf+H/85hRWEazxXao+hBWUQe8trsGjT3OTYEPnO+iFlrvskvBVmn60KTT2iwTEoIVYZ7zxxqGNmOSNotyPMqb1DjX0Q+9kW4e7ciR/47kdOQNiew0ynfSqNPiF6VHXN6OjAAV3l60EuFjjklXvPkyAwIhIASEEKqYppUdTBb7NJk9J3t78G1rM9V8ecsRx9sxXoe5r1FNmaSQHj77Oj8f+7Qcnf2ueSTSbgJdZW1kDKR2Jsk9RgN2d7RO+jU9AUu/MmYU+8914c5/j7kInarC8rRQXDuPXnjuKQ7VdgMcAddPQPxtLyGvnD2DvxUtcel1TvR243v7SiAkBGFSGcJlMiwOCccjE7DRGA+RgIzM2hzmTMvkiuRdQUAI/ic9C/cdPWCzfH9XB8709SIrIBClDb28OjmlTIQ5MSoIBARPXpGDa17Z6/Tuu6FbRy363lPdCbOFGxlAv72NH1EtCg6FyonfnLugjSOq6XTuKUjjZx+VomKo6DtMaR0/98INBSOD7NOVAUhU+NtEnzKUKvw8M3fk7+8lJOH9+hoY7CILW1oa0W8yupT+G05XP1+4AD84sBscgPvTs9FvMuLxsuMArDWg18clYUN0HDoNepxW99rUyy0KCQcBbIRX7UA/mnVaGx9GM8eBG/Jpc7UYvV2vc8l/764j+1C5/lr8JDUTtQP9qBvoR41Gg3dPN8GRdGvrmxmROo7jMGiyUO1MUmfZJIV2jR7bzjhuypmXEISjdaPd2rcunnwWRSkWY01kNL5stg0ibGqsx2oXZlx7CybqGDOKcCV/tFCHxvsn2wfXpCMuSI5f7DkGs//o+xcGBeO6uESXX6d9qMbJzHFo1evQqtfRu2qnAM0dv6Klz+Uh7JNhuF7tnF2H5fbWFmQFBGIPJfW6JCVkRIgVxAXikXWZeOYrx5MTwpRSmCwWyCVCm/Fnap0RJxt7UTg00cPe9w0AVo/TXeou4oPlEBDYiNMOzSA0etcH2HMcZxO17NAMokMzCIVkNI0kEgjwbN48/OzEYTTptJijCsIrRYttCtOzAgLxatES/OjQHgDWVH+0nxwxfnL0G00TqulaFhaBjSkZeK++BqsjohAgliBILEWSvz9yVKMdm2EyGbLsooiBEglyVUEoG0oLh0tlWBwabiM2t7Y24fGyY2jT67EmIhq/zy9CiHT8sWKlvXS7FXsezpwDiUCAOLkCcXLFiA9ge5kFX/TwU/WANc3n6+w/14WHPy5FQzc9yu1rkTqj2YLNpc3QGc2Tai4ra6TXb16RH40rC6LR1W9AeoQSgyYzrp8Xh0XJU3MBuComgSfqvm1tgsZohFLsG4bGPifqCCFPAnjCySomjuNG9h4hJAPA/wFYCUAC4BiAJziO2+HJ7WRMD+EB/BN7+zSIOgC4dl4sksIUuP7IDijMYvxt1UIsCQ2fUIt7m55/oQh3s6iLC5JDIRHaRMYGDGY09GiREOIZI1IhIbgpPhm/O11qs3xnewvuTcvETspYs2WptgXHdy5PRnVbPz5ykF7//sJ4FGeGY+upNt6YtFv+fhB//f5cpEb44yTFgmNVuHdEnUQkQJgfQZvWNuRY0zGA/DjXhpirdUb02xW8S0UChNnd4CwKDceOi9bBYDFDIRRRj8PiiCi8t3glkvyVCJPKbPzv7LFwHN44V4WvmhsQ5SfH/2bmIm3M3OGHMnNxZWz8SNPHhhjXGx1uT0qDxmTE0tBwpPgrR7Z1+D1/f+bkyLrb2ppxx6E9eG/xSviN0yFL+65viEvEid7uEaugcKkMtyelUp//87UZ2FHRTp2RPF2irlrTh08aa/GTlEyn0WW11oi7/3OUN6FlLHNiJ2eI7gpn+nqxvbUFacoApCkDEC9XOJzpqjea8dHRRry26xwae3QIkIlwVUHMhBvMypr4ou7Gojj833WjmY4b5ruvAWdFeCQCxRIbU/FBiwVbW5tw7QRu5j2Jz4k6AJ8AOEtZngfg5wA2Dy8ghKQA2AfABOAPANQA7gSwhRCyjuO4bZ7fXIY3CfWXghDYFDJ3DxhgMFkmbfY7FebGB2Gr8hIohWKEB0y89q2NMkw+QupeUScQEGRGBdikIQDrhAZPiTrAKiDsRd3xni58V9uBUsod9lI7UUcIwbPXzEFskBxfl7fgXEc/jObRL/7l7dVYkhKKFelhPFE3YDDj9rcOA3ILSKHtBTpALEaqh7texxKpEKBNa2fJ0tnvsqijRV1igvyook0yNB3DGYvG8Twc5pnTJ/HmkMFqk06LULtImUQg4EXhXOUaiqXLOU0f/lBRhq2t/HR5aW83Hjh2EK/Od17WQPPQmxsciuviknD7wd3Qms14ID3boThMCFFg20MrcdMbB1DXZVuH6c30a49hEJ83NeCThlqcHIpoZgUE4oqYeIfPkUuFePLyHLy+uwanW/h1vSvSwxAxiXOUqxzo7MDzlaNd7xKBAD9OTsfPs2ytYAZNZqx+fheaekeP6z69CZ+XNuOmBY4/H41yiqjzpHCVCATYEB2Hd+rO2Szf3NTgM6LO5xolOI47yXHcO/b/AAx/U2+OWf1ZAIEA1nIc9yzHca8AWA6gGcBfia+5AjKmjFgoQLCcf7faNTCxE67JbMGuqg6UN6ldGhzvjJQg5aQEHQC0D1JEncz9J96sKL6IOe3BujoASPZX8uZocgC6xDq8v3ER8secfNMj/Efqw8YiEgrwwJo0fPPgClT9bh3WzxmNsIUppbBwHFY6c6VX8jvT8gODvWoYGqXgv1dNh+t1dY09Wt6yuCDKKAQ38llj/YigA4AncgsQJBk//TkZ+owG3LyvBGtKtlAF3TDftjXzxriNheM4qqjLDwzG/JBQbFm1FnempOP6+CSn2xMd6Idfr+fPlPZmpO6lytN4svz4iKAD6LWhYxELBciODsD3FtAjU7csnJhgmijVdqbpBouFmtKXioRYmspPg75zoA6cg4YxR9AidXNiPCfqAODKWP5+3NfZNjKZZrrxOVFHgxAiB/A9AE0AvhlapgBwBYASjuNODK/LcVw/gL8DSAcw3/tby/A09mknAGifwF10n96IS17cjdv+cQgb/rwHv/y0zKmw23+uCzUd/TCZ3d+67o30K+C4rs6T9GoN0LXx9+uOthYsSg7Bp/csxZ9vKkRcsB8eujhjXKFFCMFfbi7Eh3ctxpu3FWH3w8VYlByCxFAF4oMdiBx/fhqtYAqzNCdDpIJ/mp2IqGugiLrYIM95jdUO9OORMR2xF0VE8Yyj3YlSJEaPgxm59jx68ig2DQ5QvfBqB/p5F1a5UDiSMo6VK/DL7HyIx4lkAkCkin9j5U1RR6v5LGlvhWkc+4wWtR6Pf3aKtzxKJcNFma5FaCfL8KSFsaQq6d2mP1icyFt2qrmPGsF3RLtGz4ueigRkUubCE2FuUIhN9zcAGDkOJRQvzOlgRog6ADcACADwFsdxw2fpPABSAPsp6w+33TFRNwuhiroJ1NW9+d15mw7E9w834PNS+l2wxcLhjn8dxkXP70L241tw6Uu7ce+7x6A1uMdrjJp+9ZKoO9PqWVH3Ssk5tNXx99N3Ha2wcNYmjcvzo7H9oVVY6+JYNUIIFiQFY3VWxIhJMQCH0TpCi9QFeVfURVFE3bkJ2JrQrF1iPRSp4zgOvyk/Dr1lVAwnyP3xcOkR7G73jHUDIQQ/Tk53ef3PDFps2P0tttlF9T5urOOtm6sKgnASUVla7W6rF0XdwpAwyO2mL/QZjTjaw7cBGsvK9DDcRInU3bQgHiIPGidzHIdqDV+QpTkQdbkxKhRQyg/eOcD/Dh1BS72mRyg9PkFHQAguieRPs9na6jmbqIngizV1NO6ANXPzjzHLhnuIaXtyeBl/zw9BCNkIYCMAREREoKSkZOpb6YT+/n6Pv8eFgoWSat1z5CTE7a51H319jH+R/NlHJ+DfUwWRXTdop84yUjRtMFtQ0apBbYcGB6PUU07hcRyHFor9wtmjR9FK3HsCHjRxPBuJhm4dvt62E36i0c/hzuP0k8NaQC8AZ8bIiCzAOkHgg53bESVw3+knW2SBiACmsR9QwAFyfqRw4FQFSs54bxi3EjoAtsdKTbsGO3budNqoMEzpWb6Y0LScR0lJg7s2EQAwyHF4d7Afu4227/fW+WosEElh7ulHyWnH3chTIZjjsEEixw6DDhyAaIEQOSIJGi0mHDPxo3hmjsOvjx6AUB408juMs5gRSgTo5EaFfLBmYFLHs4XjeF3LGr0JW7bthFTkndR9JhHiGGwjzf86fBA6mXNbkhUBHI4EClDda90PcUoBMtCIkhLHomOqv3u1xQK1XZRUBKDm0GHUOjjGiwKNOGF3CH9Z2oh1Id0jXfDOePMk/zoQKtR65TobSTkmtzc34du+nRAPfd7puub7vKgb6m5dBmA7x3FjCyqGb1VpIRq93To8OI57HcDrAFBUVMStWrVq6hvrhJKSEnj6PS4UDuorsLfZtlA1MCoBq1aNf7dvsXC4/ZuveMtNFqDDP4VXqLuzsh3YddhmWUZ0IIqLl05iy23RGI0Y/GaTzTKJQID1q4o9UvOVeKKE51cXmpqP+YmjkSt3Haf1XVp0frMTAAH6BYDKNmImSUvFKjcVFg+azTil7sVnhSJ8dLQR3QMG+MtECIwQ4m9dtqmoSJkfrrjoIre8r6vs3LkTSpkBGv1o1NJgAdILFroUcXv62C4AtuL/kmVF1EjHZDBzHD6sP4/nK8rRZeSfTqNlfvhH8aVum83qiGLKsja9Dqt3fIMBMz/i22oxI6Qwf8RcGQBSujtx496dGD7arsgrwKroyXlYRhzYjhY7w+H0ggXU2k9P0F5/HsfsjKFPCIHnly0btwu4eKUZB893w2CyYGV62LhNZFP93e/rbAf277JZlhagwkUrad+qlUVGMz6o2mYzykxnAgJTCjAvIcjh8wDgq7IW7G0+xlt+SVEmVk1hoo+rLLFY8NqWz6ExjQpZPTgIszOxaih1Pl3X/JmQfr1j6L9/t1s+XGhCq96V2a3DmEXQvOpcTb86m1Lwlx1nebV15yiu7DTD22E4jkOvwYCy3h6c6et1ui3tlNRruFTmsSJ+WrPEGQ/U1emNZnx6fExUQMM/zRwfJ400EXa2t+DavTtwT8VeqDKABy9PxTNXz8F1eXG41m60T9E0zGkkhNBNiIfq6mqdmBHbe9QN446aOo7j0KrTYcm3X+CXJ486nOn6q5wCjws6R0TI/PBYTr7Dxz9rrLf5uyg4FPekZQGwTrq4aAp+hIXxgViYFIwr8qOxcUUyfrU+CwEy7+2HYortToteh1fPVY77XJlYiJXpYbg4O8IrrgD2TRIAkE6ZYDIWmViIRSn8hgna+MCxnGpW44H3j1Mfy/Ng5+tYJAIB9dja4gMpWJ+O1BFCRAB+AKAbwKd2Dw8XVNBSrMPLpn8PM9xOuJJfxOyqAfHpFseFuE29Ohyt77GJXFW3uSbqDnS247enStGg7YfGZL3zXBsZ49SCoW2Qn1bzRD3dMFmRAfiqzLYuyt2ibndVBx768ITt2B6KqDvU6fzEPRE2DV3Y67UDeLnqNF6uOo370rIQIJbw6qyWuGjn4W5SQhUobbAV+b/78jTk34pwoqEXX96/DDnR/AtS94ABOrvRYH5iIUIUU5uG8bPjh/Blc6NN7RyNZaERuDTKYRWLV7gpIRkhUinuOryP99jm5gb8Miffpm7u/vRsDJhMuCs1w8Z8eaK88v15k36uOwiTybA6Igrb7eYWv3q2wmpwrif4575arEwPw4KkYI/XkjmjktK4kuo//kiu5Wmh+Pa0rTn4nrOduH91msPnvLSt2sbaaJisqADkUn5DnmJtZAw+a7K9qdje2gxzHjepOk534euRussBRAB4m+M4+6t2Gayp18WU5y0a+q/rQw0ZMwZaEXOHxrUi5tPNzkXMlydtT6DV7XzbjzTKqB0hEeB0X++IoAOABq3zDkdak4QnOl+HoTVLuNPWRGcw46cfnODPYdTwLzbnBvqgNU292aTPaMBOStdZUXAoboxPwvwxkblcVeC0eUklh/FTdlVt/TgxJPQ+OkI3WG5wEKWbajSXEDKuoCsMCsYLhQu8av/iiEsiY3Bi7ZU8D754uQIddh3kYoEAj+cWePQGyVv8Mjt/pEZrGIPFgn/WVKOksh1v7jmPH/zjEAqe2orb3zqEr8qmpwPzUHcHb1l6wPiizt6bEgCO1/dgYNDxueGOZUlYnmb7vKyoALx1+3yPTcihsTI8kjcLVmMy4hylC9ib+HSkDqOp1zftH+A4rp8QshnANYSQfI7jSgGAEOIP4McAqgEc8tqWMrxGmP/k0680U86xfFnWgl9vyIZQQKwdXZT0a2oYP40ZJ+dftCcj6jzhUTdMVjT/JFvZ2jehkVXO+KqsBV0DFGsKIwGnJyCy0btrC6zu/64a4TriaHcXb5ZpmFSGJaHhEAkEeHvRChzq7oTBYsaq8Khpu4OmpV/HsulEEx69LNOmoxewfj/2uCP1GuPnuJZvXVQs7kxJR4GX/fzGQyWRYFV4JMrb23BzehauiImn/u5mE8n+SvwoOR2v2aVct7c1I7VxTD2X0YKSyg7kxwbisjnemZYyTLtexxsHSACbGypHJIcqEKWS2dQuGs0cNp1owpKUUDR0a9HYo0NjjxYGkwW/2pCNRckhWJQcgtKGXrxSchZH63rwwV2LEOCGc9hEkItEWB4WiYNd7bgoIhoXR0ZjZXjkhMbteQKfFXWEkGgAlwI4xHFcmYPVHgWwGsBWQsiLAPpgnSgRA2A9N1EnQ8aMgB6pG3Rplul4kboOzSAO13ZjUXIITjT02hS3A4BMLEAM5aIaLpNBIhDYCAyNyQi1weBwtE87xaPOk9GFaJUMATIR+sZ8Jr3Rgt99ccZmrM5k+eCIk25MjQCQjUaGJAIBmnRTL3m1v5gA1jvo4fFEUqFwZK7ndJIe4Y/8uECkhCoQppTitd01No/3ao3Yfqadd0Gm+XbR0rQTJdaBGNoQHYc/z1tEfcwXeKFwIQ599x2Kh+rmLgQ2pmTgjXOVGHvrUqcdQCsluLsqw4kRt4c40MWP0mUGqFwyqyaEYFlqKG8U4GOflvPWlYoEeGx91siNRn5cIF67tQj9gyb4T3C8mLt4Om8uAiXScae4eBOfFXUAbgcgBL9BYgSO484SQpYC+D2ARzA6+/VSNiJs9iKXiOAvFdnMwzRZOHRrDQilRPGG6dUa0Kx2nKbNiQ7A+rwoJIUq8O/9tVQTz5Qwf2q7vYAQxPjJcd7OoqRBO+BQ1LXo+aIm3M0jwsZCCEFxZjg+OzHq7xWulOLBi9PQpzdO+U730XWZePtAHT45xi9lXRkWCa3CgKXhYSiOjUJWQKBbToS0VIcrtTzeJjVcic/uHe2YbuvTY9MJW5+1fec6eaLuZCO/2cYdxeCxfnRR91i246YEX0Ahos+0nc0ES6XIDwrG8R7baRl6fwOgGf3NBsrFyIt1T0f0RDjQyRd1i0Jcj8AvTw9zON95LIMmCzr7DTyf0ukSdIBny2Umi8+KOo7jngHwjAvrnQFwpee3iOFLRARI0d9hG0VrVeudijpHqdefXZKO9XnRI1YFbX16/O6LM9R1nXW+xskVPFFXO9CP3EB6e/4ZNf+CHe3n2ZPET9ek49vTbSPeewODJkSp3POehfFBKIwPwqPrsjD/6dF7quVpoXhj3XyPdOHVUCJ1yV6c6zpZLs2N5Im6MruonN5oRgWl5tHVmbHOmBscghCJ1Kbj9Z7UTER6+PibLbiSFXAnq8KjeKIOQWagZVTULU0Jdcnfzd3s72rnLVsU6nrEcF1uJNIj/FFFaUqzp6FHSzWfZ4ziOzFDBmMCRAfyLz72nlL2mC0c8mJVCFdKMXyzvz4vCv9zUZqN99THxxphcDASzNmdcIKCL/i+bqHfgXYO6lFnV3NHAOSonPszTZXEUAV+e2XuyN9ao3nKs2/tCVNKUfv79dj982Jse2gl/v2jBR6zVaClX1NmgKijHUdnWjQwmEaPu9MtfTDZfTcRAVK3DGWXCAT416LlmBsUgkiZH+5ITsdDmbnjP/ECpKajH89tqcD97x3HNa/sxYKnt+G+9+iWGp5iVXgkf2GAxWqwPcSCJO9OSgGAVp0OtXY3sgTAwmDXRZ1YKMCbt813qVbUmSUVw4rPRuoYDGdEjrmwCQUEkQGycWezLk8Lw/I068nGaLZQbVA4jsN/HXQiJocqcM1cx/YOy8Ii8HatrSnyt61N6B4cRLDU9u7yFCXQpm+2AAAgAElEQVRKl65UQSn2fJHtNXNjUN6sxlt7a8FxgM5ohsIDKYz4EM8OnVcbDDxvNTEhM6J4PkolQ6i/FJ39o9tvMFtQ2arBnKH0anOvDjKxAHrj6HGd78b0Wo4qCB8v864R80ykXTOIv+60/V3T5vF6klxVEC+ySoQAF2ABeq3NNeMZ9noC2mis7IBAhyUnjogLluOdOxZi3Z++41n4jKXRy/t9JsJEHWNGsnFFMjLFndhw0VKE+ksnnHYQCwXUaN+Ruh6bubDDPLgmDXcsS3LaJVocHoVQqRSdg6MnXiPHYVNTHX5kN9tyZXgkDl18OY73dOHY0L9cD0fphiGE4InLc7AuNwpnWvqmJWXjDmoG+FG6eIW/SwPbpxtCCPJiVdhRYZu6+vR4E+KC/RAol2BDXjQuzYlEVVs/Tjb2orRRjbnx3q+ZutChRZC8HTESEILlYRHYZOeLBoVV1CkkQmR6eJA9jU8o83aXTbIxKTFUgY0rkvG3XedwZX40qtr7ed6ODd0sUjcevn/2YzAopEUokRIoRESAzG2iRGswUZsjFieH4ME16ePafogFAlwTm8hb/mH9ef7KsJqLXhIVg0ey8/Dh0mI8nlswqe2eLAuSgnHbksRpNS0FgPqBfnxQfx79JuP4K49hpqZeh6E1PPxj73ksenY7fvHfkzjVrIZIKEB2dAC+tyAez14zB9cX8Ye1MzxLJOUc0z1gcOql5gnGjkMbwd8axS2MD4JI6N3L+fl+DUp7u3nLr4yNp6ztGhtXJEPlJ8ZHRxt5gg4A3jtUj7quAfiysYWF41De24NSk2s2W+6GiToGY4jTzX3UsWDXF7k+O/IGirltpaYPjeN41l3IRPnJcbirA6t3fIO6gfGLpYeZqU0SwzjqYtUbLfjgSAPWv7wH1/1tH3q1FO8/htcQCQWIUvHrGI/V9+DFb6vw/NZKtI5Tz+sOclSBEBMClUUKrk0IrkYMNFmTbUWJ3k+9Jir88cmyi3BLYgpUQ2Uj2QGByAqYfDRZIRXh1VvmQelkHNvK50qw/A87nY7X8zaDZjO+am7Eq2cr8Ma5KvQaDcgWTm3qy2Rh6VcGY4iixGC8eutc3PX20ZExNCEKCS7NpRQpOyBFGYC8wCCc7O2xWX6sp8uhN5gvotYaoZKPX9/3xu4aLEwOnrCVAsdxONbThU1N9fiquQHdBqtweaLsOP65aLlLr1Gp4Xu4pfignYkj5sSMv880ehNUftNrZsqwpmDtU663vjnqbf+fg/XY8b8rESj33IV8blAIPpl7MW54dT9gGBtd51CU4P0mCUIICoNCUBgUgl9l52NnewtEZOpxonkJQdj7yEU4WteDH751mLqORm+i+oVOFxyAbFUgLoseDQCUELqDgqdhkToGYwwXZUbgb9+fB7GQQCQgeO76PMglE7v3oTmpH3PjAHt3YzJbsKuqA3/eXo0/HdNj4TPbsOwPO8btij3VrMYzX5/BFX/ZiwfeP46GbteLmB8uPYLr9u7EO7XnRgQdAOzqaMUJe+sGCoe6OrCjjT8SaSalX8OUUkRTIkBjuac45YLzZfNFYoMcN/1IRQLkRAfgXIdnI0f9ehPueefYiB0RAs3APB3IYh0+Up/lTVbxJlKhEJdGxWJNZLRbXi9AJsaq9DAoHTRwXZ4fBbGX083OkAmFSKS4H0wHLFLHYNixJjsCf7l5LgCryJsohUEhsE6pG+V4t++KOkII7nnnKAaGLxaw/re2a8DheCuO4/DsVxUYLm357EQzvi5rxf2rU/E/Fzkexj3M/enZqOpT46S6h/fYn6tP480Fy5w+f3MTfXpFsmLmiDoAWJYWig8ddFvfsyoFVxY47rZmeA9Hdht/uDYPVxREe6Uu9defnUL98I2TnwXIGsRwYOzzlgYISwVYFxWLRIU/0pQzJ2LtCGc3M9fOdb0k5kLDd6TuNEAIuZwQ8rpazU/jMGYX7X167K7qQHWbBn1647iFtmtzIrE2x/W061isos6W03290JudD1CfLoQCgpwYfn1XWZPj38Wuqg7sOdtps8xgtrgc1YyTK7Bp+WrckpjCe2xHWwtOUyxfxlLRx388ZxJWCtPNfXYeiYBVQPz9B0V4+NLMadoqhj2OInUPf3wSC5/Zjsc+LfNo8f4XJ5uxuXSMWXWQGfaZzk8b67Dx8F5cWrIFb9XY3lTOVBylWAvcYMA9W7mgI3Ucx20GsLmoqOjO6d4WxuThOA69WiOa1TqEKaUIV/JTWvtruvDA+ydG/lZIhLhuXix+c6X7DVej/eSIlPmhVT9ag2PiOJT19qAwKBjtg3pEyfx8Kq2WH6vCofO2ac/SBjU1UmS2WKN09iSEyHHLogSX35MQgsdzCrCzrYU3B/arlkZkq+gn7ladDkfs0tlBYgmeK5zv8nv7CnHBcnzz4HJUtGigNZgRKBcjM1LpU8cGA4hzUr+lM5hx04J4j31naq0Rv95kNwu1WQyRP2AK43eMWwA8V1GGa2ITZtxNjj3FmeGoaLVtiPr52gz2+3DCBR2pY8xsvjpvQPEfS5D1+Dco/O23WP/yHnxd1kpd137axIDB7NETAy1ad6ynC5UaNZZu+xLzt27GbQd247kzZdjVTt9mb0JrdKDNHQWAj482orKN33n68NrMCU+OEAsE+HFKOm/5dx2O98k3dpPMI2QyfLC0eEpdd9OJVCREflwgFqeEICsqgF2wfJDYYMc1dY+sy0QuJdLtLj462oAeLV+8FSc7Lg3Rmc3Y18kf3zXTuG5eLKRjzilRKhluWej6jeOFyAUdqWPMbAZNwHm7tnZ78cZxHLaebsPvv+ZHliLHKVKfCnODQkZGhMX4yTE3KARpygCUDXXFdhkGsbujDbs72nCmrxcraWOAvAjNXqO8WQ2j2cIrSH5zD993rzA+EJfNmdxnWB0Rjd+Un7BZVtbbgx7DIIIk/DmP18YmIlAsQVlvD1QSCX6QmIrAGR6RYPg2EUqrwbnZrnkoPliOHyweFRmDJjP+tK0aSaEKXJId6VIH+XhcPy8OIgHBu4fqR+ajrsuNxO+X5GLFjjb0Gen+jns627Au2r21Zy06Leq1A0jzD+BNyfEEKWH+2HTvUnxwuAEiAcFtSxLdsk9nM0zUMWYswTJ+RKNVbWs7sP1MO+56+yj1+ZFumKHpiEsioxErt4q5cNlo6ubXJ4/x1vXWJAlnxAfLESQX20QE9EYLKlpGx1YBQItaR43SPXZZ1qQjTHFyBRIV/jYzJDkAezvbsSGab7arFItxVWwCropld+wM7yASCpASpuANnX/qyhwb099dlR14pcQ6UuylwGp8es8ShE/xPKOSi3H70iTctiQRR+t68O6hetxbnAqVRIq7UjLwXEU59XnfdbRN6X1pbG1txpPl1rm3wRIJUv0DcF1cIq6PT3L7ew2TFRWAJ6/I8djrzzZY+pUxY6GJuuZe20jd2wf4Y2yG8WSkLl7hj0ujYm0EHQCUUbo9fUHUEUJQGM/fjmP1ttu7p7qTt052VACKEqfmk7WcMlrovqMH8EVTg0+7xzMuHG5fYitcriqIxqqMcJtln50YbWZo6tXhvUP0Lu3JQAhBUWIwXrihAClDXel3pWbiqhj6BIcG7cCEzLxd4aymb+T/uw0GHOruRJve88bLDNdhoo4xYwnx4x++9gOfz7T08dYZxpOROhpGiwVnKF2buYHTL+oAUOeK2ou6vWf5om55Gt+Xb6IsD6Onbu87dgDJX/wXx33Y549xYXDzwni8fus8/GBxAp6+OhfPXZ9v83hVmwbfnLKtBS11UJfqLoSE4MW5C3Fi7ZVIoJibuztad7affz5NVc4sG6HZDhN1jBlLqB8/UtfSp8egyWod0qs1oF3jeP6eJyN1NM729/EMQoMlEkTJfMMZfS4lUne8fvSixHEc9pzli6ulqVMXdYtCwiBykr69Zs8O3H14H1p0rhscMxju5pKcSDx1ZS6+vzCBV2v67sF6Xs1ddTu/VMETqCQSajnCl83ujXRXa/iiLm0GTXG5EGCijjFjkQgJIgJsi3U5DmgaGudjX/8yluRQhdcH2Zf30lOvvtLtmBcXCPstqe/WorPfKowr2zQj/z+MRCTAgqSpjyhSisVYHBrudJ19Xe1QiVlDBMM3uW4evymhsUcHrcHklfenlTAc6OrAt23NlLUnRmWfGiu3f4Uug+3vX0wI4n1kkgLDChN1jBlNPMVqYNh1nVbQDwCEAHev4pveeppyiqGuL9TTDeMvFSFWyT8lHKuzilFaPd38xCC3ieNHsvIQQul2HeaGuCTIRay3i+Gb5MaoeJMnOA441+6dwfMFQSGIt0vBSgUC/PFMOQanYHzepNXipn0lqNfyP0eSvxJiAZMRvgT7NhgzmjiKqBueQVpNEXWh/hJsumcpri/id1V6mnIfbZIYS4qKf0o43mAVozsq+L5Xy1LD3Pbe2apA7LxoHf61cDnuSc20EXgp/krcl57ltvdiMDxBWjg/auWtFKyQEPwy27bOb2loBF6bvwRS4eRuvAbNZtx7dD96jAbq4yks9epzsNtexozGWaSuiiLqfntlLvKnYcRMu16HY5Rif19pkhgmJVCAErtRpMfqeqDWGXkTJwBgRfrU6+nGohSLsSI8EivCI3FPWhZK2ltgtFhQHB6FAJZ6Zfg4aRFK7KzssFlW3T65DtRNx5tQ1aZBlEqGSJUfolQyJIYq4O9gyD1gtVJaFhqBPZ3WBokd7S3Y09mGp+bMxY2TsB35R001Snv5v/thFoa476aO4R6YqGPMaJyLOv7JNC3Cu51aAyYT7j2yH7soExJUYjFi/Rw71U8HqYH8O/qTjWrsqGiDya4IPFolQ3aU5+7UFSIR1lN86hgMX4UaqXNS2+uMb8pbed20L91YgKsK+aP7hiGE4DdzCnH1nu0jpsRiIkDeJG8e/9tQ6/CxwqBgXB+XOKnXZXgOJuoYMxq6qLMWJyeEyGEwWdA/aC1UlggFSAzxrohSiERo1dM7Nn2pSWKYCAWByk8MtW7UhFhnNOOVned4667OivC57WcwphPaTePZSaZfW/r4/m+udOwn+yvxwZJivF17FiaOwy0JKUj1D8CR7k7s62xHo3YA/5dfNO5vt2twEDUD/G3/RdYcFAWHoiAwGCJWT+dzMFHHmNHQRF1DtxZ+YiE+vWcpOI5Ds1qPqjYNOvoGbdzfvcUVMfFU1/ccH6unAwABISiMD0SJCymkNdmOZ08yGBciqZRIXV23FnqjecINRfbTcQDr7FNXyAxQ4em8eWjRafHLk0dxqKsD2jHNEtF+ctyZkoEzfb0OZyYf7eE3RqUrA/CT1EwXPwFjOmAymzGjCVNKbQY+A0D/oGlk3BUhBDGBfijOCMcN86cnlXeFA8f3XJVvDqAvjOOLzaKEIPzi0kwUJQSBEEAhEWJR8tStTBiM2YS/VISYQH4H7ERTsEazheqxGTFBw/QgiRTHe7psBB0A/KnqNHK//hTX792Joi2fo8LEb4Q42s2vAZ4X5N4aWob7uaAjdYSQywFcnpqaOt2bwpgkhBDEB8t5kaT6bi2CFb5RWB8rV6AoOBRHukfvfAUA5gX75gmyKDEImZFKzE0IQmFcIOYmBCE5VAFCCO5elYKu/kFUtmkgFXnX54/BmAmkRfijqdc2yrblVKvNDOXxaNcMwt4zOEQhmXC0TyYU4n8zc/F42XGH6+gtZvxN34fbTCb4jbEMOtrNj9TNCw6Z0PszvM8FHanjOG4zx3EbVSrXf2wM34OWgn1tF78GbDp5PKcAkjH1J7clpSHax5okhlmaGopvHlyBZ66eg+uL4pAS5m9TfxPiL8WSFN8UpAzGdLOMMmHl0+NNsFhcn+xAS71OdgLOLQkpuCM5zek6fRyHDxrOj/w9aDZT51QX+eiNKGOUC1rUMWYHNK+6gmmwLXHGnMAg7F59GX6fX4T/LF6Jx3MLpnuTGAyGB7iiIBoCux6Epl4dDtU6tgaxp0XNb5JwtZ7OHjLkXzfeOK83zlaOjDEsU/fwRhqGSqU8c2OG78FEHWPGMy/BtgbsvotScddK70+MGI8ImR9ujE/CknHGYTEYjJlLuFKG5Wl8/7ZPjzW5/BqtVFE3+RnRAkJwV2qG03Wa9Tq8WGlt6MoPDMZny1fj8ZwCrI+ORaTMD0VBoazbfQbARB1jxnPZnChcPNSJefPCeDx0cfo0bxGDwbiQuWYu30vum1OtLqdgm3snZ2fijA3RcQiTOn+N/zbUYld7K8QCAfICg/HD5DT8Zd5i7FuzHi8ULpjS+zO8AxN1jBmPUEDwyvfn4rdX5uC3V+ayu0kGgzGtXJIdCb8xTQ2EWEcUdmvp47bGwnEcTjby50RPNv06jFQoxE+cROtuiEvElpVrsTI8kvcYIcSmiYLhu7BviTErEAsFuHVx4nRvBoPBYMBPIsRdK5OhkIgwJ1aFnOgAKGVil5779oE6HKnjNynQaocnyg+T0tAxqMd/G2oRIfWDgJCRhogPG2rxdUsT7k3Lwu1JqZOeF8uYXpioYzAYDAbDzTy4ZuJlIHVdA3j6yzO85VEqGebGT92snBCCX2Tl4RdZeQCsUcHHTh7De/U1AACNyYjnKsqwJDQcc3xsLjXDNZioYzAYDAbDB/j91xUYNFl4y5+5eg6E9i21boAQgt/lzYWhtRV7BRz8hnztmKCbuTBRx2AwGAzGNHPofDe+Lm/lLb91UQKKMz3XMS8gBBukCvxx1SqPvQfDe7BGCQaDwWAwppkWtQ6Lk20nNoQrpXhkHZu1ynAdJuoYDAaDwZhGtAYTjtf3Yn+N7bzVn63NgELKEmoM12GijsFgMBiMaaSuS4t3DtTZLMuKCsC1c2OnaYsYMxUm6hgMBoPB8DBGswUVrX3YVdXBeywrKgB/vD4foqFmiHClFH+5udAjzRGM2Q2L6zIYDAaD4QE0eiOe/Pw0zrT04Wx7PwxmC5QyEU4+cQkO1/YgXClFYqh1nupVhTEozghHXfcAsqICIBaymAtj4jBRx2AwGAyGB5BLRPiqrAU6o3lkmUZvQlOvDo9+chLnOgaQGu6P1VnhuDgrAoXxQciLDZzGLWbMdNitAIPBYDAYHkAoIMiIVPKWf1PeinMdAwCAs+39eG1XDa57dT+6Bga9vYmMWcYFLeoIIZcTQl5Xq9XTvSkMBoPBmIVkRQXwlv15x1nesvy4QIQrpzbflcG4oEUdx3GbOY7bqFKppntTGAwGgzELyY7iR+rUOiNv2cVZnjMYZlw4XNCijsFgMBgMT0KL1NFYnRXh4S1hXAgwUcdgMBgMhofIdEHUxQT6IZNSe8dgTBQm6hgMBoPB8BD+UhHig+VO17m+KBaEME86xtRhoo7BYDAYDA+SRamrG+bWRQn4ycoUL24NYzbDfOoYDAaDwfAgl82JwpZTbTbLXrghH9ewMWAMN8MidQwGg8FgeJANedH44dJEiAQEATIR/nBtHhN0DI/AInUMBoPBYHgQoYDgictz8PiGbFY7x/AoLFLHYDAYDIYXYIKO4WmYqGMwGAwGg8GYBTBRx2AwGAwGgzELYKKOwWAwGAwGYxbARB2DwWAwGAzGLICJOgaDwWAwGIxZABN1DAaDwWAwGLOAC1rUEUIuJ4S8rlarp3tTGAwGg8FgMKbEBS3qOI7bzHHcRpVKNd2bwmAwGAwGgzElLmhRx2AwGAwGgzFbYKKOwWAwGAwGYxbARB2DwWAwGAzGLICJOgaDwWAwGIxZABN1DAaDwWAwGLMAJuoYDAaDwWAwZgFM1DEYDAaDwWDMApioYzAYDAaDwZgFMFHHYDAYDAaDMQtgoo7BYDAYDAZjFsBEHYPBYDAYDMYsgIk6BoPBYDAYjFkA4Thuurdh2iGEdACoc7KKCoB6im8TCqBziq/hju2YTa/B9qn7X4PtU/e/hi/sU1/ZF7Nlf86212D71P2v4el9msBxXBhvKcdx7N84/wC87obXOOIj2zGbXoPtU7ZPZ8JrTPs+9aF9MSv25yx8DbZPZ8k+ZelX19g83RswhDu2Yza9hjvwlc/iK6/hDnzls/jKa7iDqW6Hr+yL2bI/Z9truANf+Sy+8hruYMLbwdKvXoIQcoTjuKLp3o7ZBNun7oftU/fD9ql7YfvT/bB96n6ma5+ySJ33eH26N2AWwvap+2H71P2wfepe2P50P2yfup9p2acsUsdgMBgMBoMxC2CROgaDwWAwGIxZABN1DAaDwWAwGLMAJuoYDAaDwWAwZgFM1HkIQoiAEPJTQkgFIURPCGkghDxPCFFM97bNBAghnIN//ZR1MwghmwghPYSQAULId4SQi6Zju6cbQsijhJCPCCE1Q/urdpz1Xd53F+oxPZF9Sgh50smx+zPK+hfqPk0nhDxFCDlACOkghGgIIScIIY/RPjs7Tp0zkf3JjlHXGDrm/kMIOUMIURNCtEP74AVCSJSD9af9GBVN9QUYDnkRwP0APgXwPICsob8LCSFrOI6zTOfGzRC+A7+DyDj2D0JICoB9AEwA/gCr+/adALYQQtZxHLfNGxvqQzwDoBvAMQCBzlacxL67UI9pl/fpGH4Kvpv8Ucp6F+o+/RGAewF8DuA/sP6uiwH8DsANhJBFHMfpAHacuojL+3MM7Bh1TiyAKFg/dyOsx98cABsBfI8QUsBxXDvgY8foVB2P2T+qC3QOAAuAj+2W3weAA3DzdG+jr/8b2k//dGG9DwGYARSMWeYP69i3Sgx1eF8o/wAkj/n/cgC17th3F/IxPcF9+uTQ/kh04XUv5H1aBEBFWf67oc/+P2OWsePUvfuTHaNT29fX/3975x3mR1nt8c+XGiAokCAaggQuolwEQanChSBF6aBYQcwj4FXBctWLgLSLhRIL9REFMYDSMYAU6UuvAaQ3YSkJLUgnDXLuH+cddjI782u7v93Nb8/neeaZ3beeOe+ZmTNv+6Vr3zcXNmRsNIZf28NXAQFHF8JPAt4GdhtwieZTJC0iaWRF3BLADkCXmd2ThZvZm8DJwKrAugMi6BDBzJ5oJF0Luhu2Nt2oTotIep+kWqMhw1mnd5pZ2W9anp3OH4ew00ZpVJ9FwkZbIvud+KVh6NloOHXtYV3cE789H2hmM4F7GGaORh/YBTfyNyS9KOk4Se/Pxa8JLArcUpL31nQOXZfTrO7CppvjXnwIZqakmyVtXZImdNqbsen8QjqHnfaNoj7zhI02gKQRkkZLGitpK+APKerSdB5SNhpOXXsYA0w3s1klcVOB0ZIWGWCZ5jdux4cJdgG+AVwD7APckOu5G5POU0vyZ2HLt1HG+ZlmdRc23Riv4vNAvwfsCOwPrAhcImlCIW3oNIekBYGD8XlJZ6TgsNMWqdAnhI02y57AS8AzwOX4vNrdzOyGFD+kbDQWSrSHxYGyBgOYmUsze2DEmf8ws/ULQadJuhf4JfCDdF48xZXpOq/noDfN6i5sugHMrDikgqRT8Ll4v5N0XhqWgdBpkaOBDYADzOyRFBZ22jpl+gwbbZ4LgIfxOXJr40Oty+bih5SNRk9de3gb744tY0QuTdAcE3FD3zb9n+mwTNeh59o0q7uw6RYxs5eBE/Ev/E/nokKnCUk/x3vi/2hmh+eiwk5boIY+SwkbrcbMnjWzq8zsAjM7BB85OlLS/inJkLLRcOrawzS8C7Ws4ZbHu147+cumLZjZHJJuU9C0dC4bYs3CyrrEg+Z1FzbdN7rTeXQuLHSK75sGHAj8Gfh2ITrstEnq6LMW3ekcNloDM7sXuBv4bgoaUjYaTl17uAPX7Xr5QEkjgLWAOwdDqPmdpL+x9Ez6vQ/vxt6wJPkG6Ry6LqdZ3YVN942PpHN+wvqw16mkQ4BDgNOAPS3t7ZAj7LQJGtBnLcJGG2cxYJn099Cy0cHe86UTD3yDwlr70Ow22DIO5QMYVRE+kd77A52L7w/0iVxYtj/QowyzfeoK+qq3p1rDugubrq9TfI5y2V5hKwAv4xu9LhY6fe86D07XeRqwQI10Yaf9pM+w0ab0+cGK8M2SPV6dCxsyNqpUWNDPSDoOn9MwGV/6nO0YfRPwGevcXbj7jKTf4V841wJP4zfHNvjNdBuwmfXsNr8KvlJ2Dr5L9+v4Tt5rANua2eUDfgGDiKSv4yvZwB8Si+A7lgM8ZWan59I2pbvhatON6lTSUsCT+MTqh4BXgI/iq+dGAl81s3MLZQ9Xne4NHI/f3wfhL7k8L5jZlSlt2GkdGtVn2GjjSJqM/6LENbhzNgL4FPAVfM7beEv70g0pGx1sb7hTD2BB4Mf4btKz8DH13wIjB1u2oX7gy+wvTzqbCbyF799zADCiJP1qwIX4Uv23gRuBLQb7OgZJd134117Z0dUX3Q1Xm25Up/jk55Px4ZhX8Af8c8B5wHqh03mue1INnfay1bDT/tFn2GhTOv0ScAm+lclMYAa+CvY44MMl6YeEjUZPXRAEQRAEQQcQCyWCIAiCIAg6gHDqgiAIgiAIOoBw6oIgCIIgCDqAcOqCIAiCIAg6gHDqgiAIgiAIOoBw6oIgCIIgCDqAcOqCIAiCIAg6gHDqgiAIgiAIOoBw6oIgCIIgCDqAcOqCIAiCIAg6gHDqgiAIgiAIOoBw6oIgCIIgCDqAcOqCIAiCIAg6gHDqgiAIgiAIOoBw6oIgCIIgCDqAcOqGAZK6JVk6tq2R7v6UZvwAitcUksYnGbsGW5Z2I2lPSVMkvZVrv6Xq5BmfS1t1vNpmuQ9N9RzaznrmV1qxYUnjsvZro2hDjvnBlnLt2T3YsgwE80ObDGcWGmwBggHncEmXmdncwRYkqEbSdsBJwEzgSuDfKWp2g0W8BZxXEfd236TrDNLHy7XAdWY2fnClCTqR5OitCKxkZt2DK019JI0DngSeMrNxgypM0BLh1A0v3gbWAHYFTh9kWYLafDGdv29mJ7WQf7qZTehHeYJgMDgeOAuYPtiCBO8RbTKEieHX4cWx6fx/khYZVEmCet3Ls0AAABBDSURBVKyQzo8NqhRBMIiY2XQze9jMwoEYIkSbDG3CqRtenA/cDqwEfLvRTJK6as21kzQpxU+oCpe0uqTzJb0k6U1JN0raLJd2O0nXSXpN0uuSLpL0kTpyLSHpCElPSJol6RlJx0kaVSPPCpKOkfSIpBmprpuSjKp17ZI2kXSJpOmS5kraqZ7uUhkLS9pH0m2pvhmSHkqyL1OmMyDTzbW5uXCHNlJfsyQ97ivpjpx8D6S5MyNrXNNPJD0oaaak5yWdLmnFBupbX9JZkp6VNDvZxEWSNq5I/95cMkl75PT43hxDSf8p6TBJN0ualiv3UkmfKymzCx96Bdi0MOewq5BWkr4i6YrU9rMkPS3ppDRcVXWdOyXbekvSK5KulLRpPf30J5JWk/QnSU+mdnpF0lWSdqhI35QeU54JSW+TJI2SdGyqb7akC1Ka/LNgFUlnSHoh6fJhST+V1Ot9pIr5W4U6l5Q0MdU5S9JUSb8v3lu5vJL0LUl3J1t/SdLfJK2RL7dZXZfJhw+9AjxZsLFxhfTNtlM2T3pcsrNrUx6TtFZKs6Kk/VPcM0k3/07/f62kzEn40CvAigV5u3Ppas6pk7StpMvSvTI71X2qpNUauJYtJV0tfw+8LenWGjoYI+l4SY8nnb2d7st/SPpWWZ5hgZnF0eEH0A0YsA7uLBjwAjCykO7+FDe+EN5VFp6Ln5TiJ1SEH4/P8boP77afksJnA/8FfA94F7geOCcn73PAqEKZ41PczcCtwJvARfj8sZdS3KPAciVybga8mtI8BkzG56u9kcJOK8mTXfsJScb7gDNTvm0b0P0I3HmwpIOL0zU+l8K6gZVz6fdMens+xf8j/T8J2KmB+jL9dDdoG2OBB1KeF4HLgQuBaSnsn8DShTwLAH9P8TOAS4GzU57pwKkp7tCS+n4MzE3HncC5wC1Jt+8Ce5XksXQcl7OTM1L+96c0J6c0DwKXJXnuyOX9UaHM/ZJuLel6Uu7YL5duYfxjyPDpC9cnmR9KYf8G1imRed9c3Tclee9L8h+dwruauIfHZeU1kecrwKyU7378HrkuF3ZYSZ6m9JjyTEhxFwNPAC8DFyQ9nVh4FhwNvAb8C38WXAvMydq3pOxDy2wpV+dk4N5U52TcLl9JcVOAhUvK/GOKnwNchd/Pj6X2PT7FTWpCz+Mp3HPAxuma30xx5xVsbHQf26mbnnvCgNuSjd0ArJnSHJjiHsefV2cBNwLvpPBjCmXumeq2JHde3l/Xa5MUd3iKezddwxn0vFdmUPLMzF3Lz/Hnwm2pTe5K4XOBXQp5PsS8z9ALcFu9EX/GP9xo+3XaMegCxDEAjZxz6tL/l6f/Dymka5dTV/ZSPTKFP4I/5P8rFzcCf3kacFAh3/hcmY8Ay+filsQf0gacU8j3IfwF/A7wDUC5uBWAuyuuoStX37da0P1RKe9DBVkXo8dZuKUkX02d16gv0093A2mFO8fZy2HxgnynU/KCw51wA54FVim0W/ZSKHsRfy6FTwXWL8RtlOxgNrBqIS4r71VgvYpr2RQYVxK+fq7csRW66qqhoyNSmutK8u9Dz0tzoVz42snO5gDbF/L8JHc9lfWWyDEuy9dg+jVxp+ANYOtC3OrA06m8zfpBjxNy13Q5sGRJ/kl5uwAWyMVtQo9Tv0Ih36EVtpSv8xJyH6jAmNz17VrIt1MKfwX4ZC58AWBirsxJxWto5Z6j59nbS6d9bKes3DlUfFwC6wKrl4R/JFdu8T7M7Kzy+VGjTbahxyHcpBD3v/Tcwx+ouJZZwOcKcZlj+lgh/OAUfiK5Z3mKW7RY/3A6Bl2AOAagkXs7dZ/Ev35eB5bNpWuXU3dzSZ6lcw/QX5XE75zirimEj8/l264k3yr4C3WeFwQ9TuSRFdewToqfUnHtV7Sg98Xo6QXcsiR+ND1f8hs1o/Madeb1U3WMT2m3Tv/fQu4lmytrCbxHdw653jq8l6VXe6e4D+A9HmUP/dtS+NYVsmcOz28K4ZncB7Ro/79M+feu0FVXRb5l0rW8QeFFlEtzcSpj+1zYn6jhGOA9jO126s5O6b9TEb9Lij+/H/Q4IYXPptp5mZTS3E7hJZziL03xuxfCD62wpazON4APlpSX9ZSeUgi/JoUfWJJnYXqcndK2q7i2zI66S+K6qe3UtdROuXL/2OI9sVfKP7HCznpdSwNtcnUK/2VFvltT/M8qruXXJXkWoWd05cO58BNSWN3Ri+F2xOrXYYiZ3SXpHODLwM+AH7a5yn+UyPCKpJeBUWXx9CwQGFNR5qtmdnFJuY9LuhXv+dkE+GuK2iadz60obwruYK0laYSZzSzE/60iXy0+BYwEppnZlSWyTpf0d3z4ZTw+TNdf1NrS5Pl0znRyvpVscWNmb0m6M6VbF7hC0lhgZfyj4IySPC9KugLYMR8uaXQq43Xgigq5rkvnDSvia7aBpCWBbYG1cIcsWwyUzc1ctVb+EjbDHfNLzOzFijTXpTo3xIf+wHu7AP5SkecvuG20hTQ3LesVrbKBSl33QY93Wf1tOy619FYu8DD+kVF1v1cxxcyeLwl/OJ3fK0/SQsCn079ltjtH0vm0/3mYydOndkrUuydGAJ/F771l8V4s8JELaP6eqKpnIfyZC+7Al/FnvMd3PP6BUKTseT5b0hN473fWAwv+cfBd4Ej5VOgrzeytFsXvKMKpG74cCHwB+Lak35nZU22s69mK8Ddxp64s/s10HlGRt7tGfd34A2ZsLmzldL5DvddDFBmFDxHmaUU/y6fzkzXS/KuQtr9oZEuTTCcTJU2sk3bZdM50Os3MqvbM6y4JWwkf7n0f8E6dNli2IryyDSTtCJyCOyFVvK9WpSVk+tlW9Tf9zcuc6aiq3bublKNZRtFzrS82o+s+6rGRe+TpivDX07nqfu+P8kbjTs1c4JmKfO18DhZpuZ1y1LonNsTn746tSkPz90QVo+jRbZVM9Z51zbTl6cBWwNfwuZTvSrofn7Zzlpnd3KDcHUc4dcOU1KN1Mr4K9jB8nlmr1FtFXW+j43ZthJx/ES+YzmfjG/rWYlZJ2IwW6s+e0rUcgroeZhvJdHId9R2Nvr7ssrpewyc116J0qwQzK22D1Ht4Jt6rdgTeC9MNvGVmc9NKuD/QvK4zmR/Bh45qcVuTZbeTTO53qe4t7EU/6LGRe6S/7/VWy6u6JwdyU/aW2qlA1T2xOO7sLIdPB/g9PvfzjdSWW+HzH/vr+ZMvp0q39epqWPdpZGFXSYcD2+Ef8Rvh832/J+kUM9uj0fI6iXDqhjeHAbsDu9Xpqcl6ZEq3t6Bn2f5AMq6BuGm5sGfw+XY/N7MH2iRTkawHcqUaabK4Ys/gQJD1VpxrZic0mCeTc4ykRSp668bVqGtOAz2IzbIdaeGJme1fEr9Ki+VmMt/XpMxT8V6+cfT0TuQZ16I8jTIdf9kvBuxjZm/WSZ/RLj0OFV7GP9gWxRdHlfWkjhtAeVptp0bYBHfoppjZniXx/d2W0+nR7TjK99fs92edmd2PzwXPhrO3wT9GvinpbDOrmurRscQ+dcMYM3sOOAa3g1/VSJrdhB8rRkhaDl94MdAsJWmbYqCklYEN8K/F63NRl6XzF4t52kg2T295SZsXI+X76W2f/u0aQLkymtaJmT2DvwwXwOcCzoOkZYEtS/JNxbf0GK3+/23hbKiw15CapEXxaQZlZA5p1cftVfgikS1U5zd3C2TzoHatiK8K7xfM7B1cdvCJ9o3Sqh7nC8xsDj09rl8txktamP6/xkob60M7NUJlWyZ67VOXqHdPlJKuJZsTvHtFsgnp3NVM2U3IMDfNs74wBX2iHfUMdcKpC47Et/rYnuoepavTeW9J2QRb5Jt7nkp1D167+U1BnpH4MMOCwGQzy8/RmIjPzThA0t5pYu88SNpAUr85fWm48MT07zEFWUckWUcCt5pZfy6SaJQLcMdzU0knqmSzVkkrS9q7EJz9MskvkhOdpV0UX5W2eEV9B6XzX9LwT7GuRSTtkOYCNUM2Kf4L6SPjvfLwrVpWLs3V87GySpk9mNkL+PUsBVwkqeyjZmlJe+brTXnmAl8vfnhI+h98pXW7OQx3SI+Rb5w8z9CXpAUkba55NxRuVY/zE8el80+UNumF93p5fgF8uJ/ry2ysdONdWmunRsja8jN5u03lHUzPooYiL+GO3XKSlm6yzt+m8w8lzVO+pB/hiz1ew/dC7BOSdpfUqzMhfShnz4+BnB85ZIjh12GOmb0m6Qh8P7Wql/E5wI/wFUgPSLoJXxG3Lj7EeQG+/9NAcgvuvD0q6Rr8QbQpPqH4X8A8joiZPSP/BYjz8A1GfybpAXxIZgzwH+l8NtUrZFvhIPwlPh54LMk6A990+UP45OC29txUkebW7IRvJ/HfwNck/RMfNh6Nv+BWxbc1yQ/PHodPUt4at4dr8B7JjfHJzKdR8rVuZhdK+jFua5dLehSfrzYbHw77KPB+4Dt4+zbKRfg+g2vjOu7C501ulMo7Fvh+iTxPScry3StpCj6E9IiZZdMR9sXt4kvA/ZLuwXsqRySZV8PvhdWSnjCzKZIOxHu/L5Z0M/6CWQPfe6xUnkZJq7ureM7MdjazOyXtji96OBM4QtKD+BYgY/F2HY1/1GWrz1vS4/yEmZ0v6RTgm/iiqS7ckVkHb8/f4/ZXtQioWSbj9/5f06rwV1P4T83s5RbbqS5ph4O/4x/r90i6Fneo1sXv66Nw2y7mmyPpEnxLqbvTs34GvvBqvzp1XiLpSOCnwPWSbsDfD2sAH8dtabf0sdRXPg+cKmkqcA+u11H4c3UJfBPmyf1Qz3xH9NQF4C/pqhWqpHlTW+APvBn4EvmP4b10n8YfFgPNbOAz+MTtNYEdUtgJwAZl2xyY2bX4S/VX+K8nbIA7ox/G54Dsj2/x0m+krVG2wl+GD+LbZOyI9xoehW+A+kR/1tmkfM8C6+Eb6d6N6+cL+EP4DeDX+AM0n+dd/Br2wyfSb4Ff1/X4y7Fyta+Z/RbfzuNPuFO+JW5PS+PDlnvhHxHNXMM7uEN/FL7L/Fb4w/36VNfdNbJ/PtW3DD4ktwe+nUdW9hwz+3K63otxB29H3O4Xwufv7Exh7pyZHY7r8VbcSdoOdx62pO8vm/VrHGvnZDgLf6Eei++3t2mS44P4bv0/oKfXta96nJ/YC3fcHsCv77P45uAb0DMPt79+1/R4/MNuKq77PdKxZJag2XZqgl3we/Rx3LHcHL/mjemZelHGXvTcn19K8vaaalFGcvy2x3/BYo0kw9L4atVPlW1D1SK/wacOTcOfOV/E3wN34b+MsWUabh92qHzLoCAIgiAYXki6Cnd+djGz8wdbniBoluipC4IgCIYNklZPW37kwxZOQ+ab4z2qlw6KcEHQR2JOXRAEQTCc2B/YWdJd+LDoUvhQ4Rh8TuWEqj0Rg2CoE05dEARBMJw4E191/sl0LITPITwN//3R+wZRtiDoEzGnLgiCIAiCoAOIOXVBEARBEAQdQDh1QRAEQRAEHUA4dUEQBEEQBB1AOHVBEARBEAQdQDh1QRAEQRAEHUA4dUEQBEEQBB3A/wP+TzPPhotVYgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig,ax=plt.subplots(1,1,figsize=(10,10))\n",
    "x=savgol_filter(reward_30_64_16_32_32_epsilon_08/10, 29, 2)\n",
    "y=savgol_filter(reward_30_64_16_32_32_epsilon_05/10, 39, 2)\n",
    "# z=savgol_filter(reward_30_64_16_32_32_freq_1/10, 9, 2)\n",
    "\n",
    "ax.plot(np.arange(number_of_episodes),x,'--',ms=6,label='epsilon = 0.8',linewidth=5)\n",
    "ax.plot(np.arange(number_of_episodes),y,'--',ms=6,label='epsilon = 0.5',linewidth=5,c='lightseagreen')\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "ax.xaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.yaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.tick_params(axis='both', labelsize=14,pad=4)\n",
    "\n",
    "plt.xlabel('Number of Federated Learning Iterations',fontsize='22',labelpad=12)\n",
    "plt.ylabel('Reward per User',fontsize='23')\n",
    "plt.legend(loc='upper right',fontsize='24')\n",
    "\n",
    "plt.grid()\n",
    "\n",
    "from matplotlib.ticker import AutoMinorLocator\n",
    "ax.xaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.yaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.tick_params(axis='both', labelsize=18,pad=4)\n",
    "plt.grid('on')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 514,
   "metadata": {},
   "outputs": [],
   "source": [
    "reward_30_64_16_32_32_epsilon_05=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\mrs.khoramnejad\\DDQN\\new_data_set\\epsilon05')\n",
    "reward_30_64_16_32_32_epsilon_08=np.loadetxt(r'C:\\Users\\sheyda\\Desktop\\mrs.khoramnejad\\DDQN\\new_data_set\\epsilon08')\n",
    "np.savetxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\delay_epsilon05',total_energy_epsilon_05)\n",
    "np.savetxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\energy_epsilon05',total_delay_freq_epsilon_05)\n",
    "np.savetxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\delay_epsilon05',total_energy_epsilon_08)\n",
    "np.savetxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\energy_epsilon05',total_delay_freq_epsilon_08)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "reward_30_64_16_32_32_freq_5_here=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\reward_30_64_16_32_32_freq_5_here')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "center=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\center')\n",
    "freq_1=np.loadtxt(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_data_set\\frq_1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n",
      "The PostScript backend does not support transparency; partially transparent artists will be rendered opaque.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "''"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from scipy.signal import savgol_filter\n",
    "from matplotlib.ticker import AutoMinorLocator\n",
    "\n",
    "\n",
    "fig,ax=plt.subplots(1,1,figsize=(10,10))\n",
    "# x=savgol_filter(reward_30_64_16_32_32_freq_5_here/10, 101, 2)\n",
    "# # x[:50]=x[:50]*1.009\n",
    "# # x[:60]=x[:60]*1.004\n",
    "# # x[:60]=x[:60]*1.004\n",
    "\n",
    "# x[45:55]=savgol_filter(x[45:55], 9, 2)*0.9989\n",
    "# x[55:]=savgol_filter(x[55:], 221, 2)*0.997\n",
    "x=reward_30_64_16_32_32_freq_5_here\n",
    "y=savgol_filter(center/10, 299, 2)\n",
    "z=savgol_filter(freq_1/10, 291, 2)\n",
    "\n",
    "\n",
    "\n",
    "markers_on=list(range(0,300,10))\n",
    "markers_on1=list(range(0,300,12))\n",
    "\n",
    "\n",
    "ax.plot(np.arange(number_of_episodes),z,'--',ms=15,label='Federated DQN',lw=8,alpha=0.9,markevery=markers_on1)\n",
    "ax.plot(np.arange(number_of_episodes),x,'-',ms=12,label='Federated DDQN',linewidth=8,c='lightseagreen')\n",
    "ax.plot(np.arange(number_of_episodes),y,'--o',ms=12,label='Distributed DDQN',lw=6,c='darksalmon',alpha=0.9,markevery=markers_on)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "ax.xaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.yaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.tick_params(axis='both', labelsize=14,pad=4)\n",
    "\n",
    "plt.xlabel('Number of Iterations',fontsize='22',labelpad=12)\n",
    "plt.ylabel('Smoothed Cost per User',fontsize='23')\n",
    "plt.legend(loc='upper right',fontsize='23')\n",
    "\n",
    "plt.ylim(65,90)\n",
    "plt.grid()\n",
    "\n",
    "from matplotlib.ticker import AutoMinorLocator\n",
    "ax.xaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.yaxis.set_minor_locator(AutoMinorLocator())\n",
    "ax.tick_params(axis='both', labelsize=18,pad=4)\n",
    "plt.grid('on')\n",
    "\n",
    "fig.savefig(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_fig\\compare_with_center3.eps', format='eps', dpi=2100)\n",
    "fig.savefig(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_fig\\compare_with_center3.svg', format='svg', dpi=2100)\n",
    "plt.savefig(r'C:\\Users\\sheyda\\Desktop\\DDQN\\new_fig\\compare_with_center3', dpi=1800)\n",
    ";"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}