rlgt.agents

package documentation

The rlgt.agents package defines the GraphAgent abstract base class, which formalizes a reinforcement learning agent for graph theory applications. It also provides three concrete agent classes that inherit from GraphAgent and implement the Deep-Cross Entropy, REINFORCE and Proximal Policy Optimization (PPO) methods, all based on PyTorch. In addition, the package includes several auxiliary classes for defining random action mechanisms.

This package can only be used if the optional agents extra dependencies are installed.

Module	`deep_cross_entropy_agent`	This `Python` module contains the `DeepCrossEntropyAgent` class, which implements a reinforcement learning agent for graph theory applications using the Deep Cross-Entropy method with `PyTorch`.
Module	`graph_agent`	This `Python` module defines the `GraphAgent` abstract base class, which formalizes the concept of a reinforcement learning agent for graph theory applications.
Module	`ppo_agent`	This `Python` module contains the `PPOAgent` class, which implements a reinforcement learning agent for graph theory applications using the Proximal Policy Optimization (PPO) method with `PyTorch`.
Module	`random_action_mechanisms`	This `Python` module defines the `RandomActionMechanism` abstract base class, which formalizes the concept of a random action mechanism in the context of reinforcement learning agents for graph theory applications...
Module	`reinforce_agent`	This `Python` module contains the `ReinforceAgent` class, which implements a reinforcement learning agent for graph theory applications using the REINFORCE method with `PyTorch`.