releso.agent

Parsing of the agents available in ReLeSO.

Out of the box the ReLeSO package uses agents implemented in the Python package stable-baselines3. Currently the agents Deep Q-Network (DQN), Proximal Policy Optimization (PPO), Soft Actor-Critic (SAC) Advantage Actor Critic (A2C) and Deep Deterministic Policy Gradient (DDPG) can be used directly but the others can be added easily.

The following table shows which agent can be used for which shape optimization approach:

Agent	Direct shape optimization	Incremental shape optimization
PPO	YES	YES
DQN	NO	YES
SAC	YES	NO
DDPG	YES	NO
A2C	Yes	YES

Author:

Clemens Fricke (clemens.david.fricke@tuwien.ac.at)

Classes

A2CAgent(*, save_location[, logger_name, ...])	A2c agent definition.
BaseAgent(*, save_location[, logger_name, ...])	Base agent definition.
BaseTrainingAgent(*, save_location[, ...])	BaseTraining agent definition.
DDPGAgent(*, save_location[, logger_name, ...])	DDPG agent definition.
DQNAgent(*, save_location[, logger_name, ...])	DQN Agent definition.
PPOAgent(*, save_location[, logger_name, ...])	PPO agent definition.
PretrainedAgent(*, save_location[, ...])	Pretrained agent definition.
SACAgent(*, save_location[, logger_name, ...])	SAC Agent definition.