MARLlib: A Scalable and Efficient Multi-Agent Reinforcement Learning Library
Abstract
A significant challenge facing researchers in the area of multi-agent reinforcement learning (MARL) pertains to the identification of a library that can offer fast and compatible development for multi-agent tasks and algorithm combinations, while obviating the need to consider compatibility issues. In this paper, we present MARLlib, a library designed to address the aforementioned challenge by leveraging three key mechanisms: 1) a standardized multi-agent environment wrapper, 2) an agent-level algorithm implementation, and 3) a flexible policy mapping strategy. By utilizing these mechanisms, MARLlib can effectively disentangle the intertwined nature of the multi-agent task and the learning process of the algorithm, with the ability to automatically alter the training strategy based on the current task's attributes. The MARLlib library's source code is publicly accessible on GitHub: https://github.com/Replicable-MARL/MARLlib.
Cite
Text
Hu et al. "MARLlib: A Scalable and Efficient Multi-Agent Reinforcement Learning Library." Machine Learning Open Source Software, 2023.Markdown
[Hu et al. "MARLlib: A Scalable and Efficient Multi-Agent Reinforcement Learning Library." Machine Learning Open Source Software, 2023.](https://mlanthology.org/mloss/2023/hu2023jmlr-marllib/)BibTeX
@article{hu2023jmlr-marllib,
title = {{MARLlib: A Scalable and Efficient Multi-Agent Reinforcement Learning Library}},
author = {Hu, Siyi and Zhong, Yifan and Gao, Minquan and Wang, Weixun and Dong, Hao and Liang, Xiaodan and Li, Zhihui and Chang, Xiaojun and Yang, Yaodong},
journal = {Machine Learning Open Source Software},
year = {2023},
pages = {1-23},
volume = {24},
url = {https://mlanthology.org/mloss/2023/hu2023jmlr-marllib/}
}