-
Notifications
You must be signed in to change notification settings - Fork 1
/
bayesian_optimization.py
85 lines (72 loc) · 2.34 KB
/
bayesian_optimization.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
import numpy as np
from ax.service.ax_client import AxClient
from ax.service.utils.instantiation import ObjectiveProperties
import statistics
from tqdm import tqdm
from game_logic import Game
from AI import MDP2
ax_client = AxClient(random_seed=42)
ax_client.create_experiment(
name="moo_experiment1",
parameters=[
{
"name": f"m1-3",
"type": "range",
"bounds": [250, 750],
"value_type": "int"},
{
"name": f"m4-6",
"type": "range",
"bounds": [25, 100],
"value_type": "int"},
{
"name": f"m7-9",
"type": "range",
"bounds": [8, 50],
"value_type": "int"},
{
"name": f"m10-15",
"type": "range",
"bounds": [5, 20],
"value_type": "int"},
{
"name": f"depth_scale",
"type": "range",
"bounds": [2, 9],
"value_type": "float"},
],
objectives={
"score": ObjectiveProperties(minimize=False),
"time_per_move": ObjectiveProperties(minimize=True)
},
parameter_constraints=["m1-3 >= m4-6", "m4-6 >= m7-9", "m7-9 >= m10-15"],
overwrite_existing_experiment=False,
is_test=False,
)
def evaluate(parameters):
param_array = np.array([
parameters["m1-3"],
parameters["m4-6"],
parameters["m7-9"],
parameters["m10-15"],
parameters["depth_scale"],
])
scores = []
time_per_moves = []
num_runs = 3
for game_run in range(num_runs):
g = Game(use_gui=False, no_display=True)
g.setup_board()
m = MDP2(g, game_obj=Game, verbose=False, best_proportion=1, core_params=param_array)
print(f"starting run {game_run} with parameter array {param_array}")
current_score, current_time_per_move = m.run()
scores.append(current_score)
time_per_moves.append(current_time_per_move)
to_return = {"score": statistics.mean(scores),
"time_per_move": statistics.mean(time_per_moves)}
return to_return
if __name__ == '__main__':
for i in tqdm(range(15)):
print(f"starting parameter iteration {i}")
parameters, trial_index = ax_client.get_next_trial()
ax_client.complete_trial(trial_index=trial_index, raw_data=evaluate(parameters))