GitHub - ZhihanLee/Constrained-SAC-PPO: A pytorch implementation of Constrained Reinforcement Learning Algorithm, including Constrained Soft Actor Critic (Soft Actor Critic Lagrangian) and Proximal Policy Optimization Lagrangian

This is a prototype of Constrained Soft Actor Crirtic or Soft Actor Critic Lagrangian (CSAC or SAC-Lagrangian)

The basic SAC algorithm comes form ElegantRL :

I established a Constrained SAC algorithm to deal with CMDP problem.

See 'Class AgentConstrainedSAC' in "AgentSAC.py" for details.

log 22.7.8: Detach update lambda from update_net method

log 22.7.30: A pytorch implementation of Proximal Policy Optimization with Lagranian (PPO-L) will be released soon

log 23.1.3: Add a pytorch implementation for PPO-Lagrangian with LSTM, see details in LSTM-PPO-L.py

Name		Name	Last commit message	Last commit date
Latest commit History 22 Commits
AgentBase.py		AgentBase.py
AgentSAC.py		AgentSAC.py
LSTM-PPO-L.py		LSTM-PPO-L.py
LSTM-PPO-L.svg		LSTM-PPO-L.svg
README.md		README.md
agent.py		agent.py

Provide feedback