Skip to content

The model of the world objects of the Deep Reinforcement Learning showcase "learning agents".

License

Notifications You must be signed in to change notification settings

mvrahden/learning-agents-model

Repository files navigation

Learning Agents Model

This repository holds the model for the learning agents-showcase found here.

World Composition

The World consists of several derived subclasses of WorldObject. The following UML-chart visualizes the relations between the classes.

Agent Composition

The Agent itself is configured with a collection of sensors, the Sensory, and the DQN-Solver. The Sensory helps perceiving it's environment, whereas the Solver helps to derive the right actions from its perception. The following UML-chart shows the relations between the individual classes and gives an overview of the properties and behavior of the agent.