The role of $\gamma$ in Offline RL

This is a pytorch implementation of discount factor in Offline RL on Datasets for Deep Data-Driven Reinforcement Learning (D4RL), the corresponding paper is On the Role of Discount Factor in Offline Reinforcement Learning.

For experiments on D4RL, our code is implemented based on TD3+BC, please click on the folder TD3+BC_gammma and then:

$ python3 main.py

For experiments on Toy Example, please click on the folder Toy and then:

$ python3 bcq.py

$ python3 dqn.py

If you find this open source release useful, please reference in your paper (it is our honor):

@inproceedings{hu2022role,
  title={On the role of discount factor in offline reinforcement learning},
  author={Hu, Hao and Yang, Yiqin and Zhao, Qianchuan and Zhang, Chongjie},
  booktitle={International Conference on Machine Learning},
  pages={9072--9098},
  year={2022},
  organization={PMLR}
}

• If you have any questions, please contact me: yangyiqi19@mails.tsinghua.edu.cn.