DRL-Base-EMS for HEVs
The Vehicle Env is OpenAI-Gym like Env, so it needs to work with Gym Env files.
- You can simply install the base Gym library, use with the following command:
More information about Gym Env can find at https://github.com/openai/gym
$ pip install gym
The Algorithms library from Tianshou is currently hosted on PyPI and conda-forge. It requires Python >= 3.6.
- You can simply install Tianshou from PyPI with the following command:
More information about Tianshou library can find at https://github.com/thu-ml/tianshou
$ pip install tianshou
- Get ready with Gym Env and Tianshou library
- Download the
Vehicle_Env
file underDRL-Base-EMS
- Open the Gym Env file of your workspace with suffix like
\gym\envs\
- Add all three content of
\DRL-Base-EMS\Vehicle_Env\classic_control\
to\gym\envs\classic_control\
- Change the content of the
\gym\envs\classic_control\__init__.py
and\gym\envs\__init__.py
by the instruction in\DRL-Base-EMS\Vehicle_Env\__init__.py
- Now, you can try the different DRL Algorithms in your workspace.
We acknowledge the following repositories that greatly shaped our implementation:
- https://github.com/thu-ml/tianshou for providing popular DRL Algorithms in PyTorch and operating guide. Please cite their work if you also find their code useful to your project:
@article{tianshou,
title={Tianshou: A Highly Modularized Deep Reinforcement Learning Library},
author={Weng, Jiayi and Chen, Huayu and Yan, Dong and You, Kaichao and Duburcq, Alexis and Zhang, Minghao and Su, Hang and Zhu, Jun},
journal={arXiv preprint arXiv:2107.14171},
year={2021}
}
If you find our code useful to your work, please consider citing our paper:
@article{DRL-EMS,
title={A comparative study of 13 deep reinforcement learning based energy management methods for a hybrid electric vehicle},
author={Hanchen Wang, Bin Xu, Yiming Ye, Jiangfeng Zang},
journal={https://www.sciencedirect.com/science/article/pii/S0360544222033837},
year={2022}
}
If you have any questions, please create an issue in this repository or contact Hanchen Wang (haw110@ou.edu)