arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.01955cs.LGcs.AIcs.MA
收藏

PPO在合作多智能体博弈中的惊人效果

The Surprising Effectiveness of PPO in Cooperative, Multi-Agent Games

Tsinghua University(清华大学);University of California, Berkeley(加州大学伯克利分校);Shanghai Qi Zhi Institute(上海期智研究院)

Chao Yu, Akash Velu, Eugene Vinitsky, Jiaxuan Gao, Yu Wang, Alexandre Bayen, Yi Wu

摘要:

近端策略优化(PPO)是一种普遍存在的基于策略的强化学习算法,但在多代理设置中的利用率明显低于离策略学习算法。这通常是由于人们认为PPO的样本效率明显低于多智能体系统中的非策略方法。在这项工作中,我们仔细研究了PPO在合作的多智能体设置的性能。我们表明,基于PPO的多智能体算法在四个流行的多智能体测试平台中实现了惊人的强大性能:粒子世界环境,星际争霸多智能体挑战,谷歌研究足球和Hanabi挑战,具有最小的超参数调整,没有任何特定领域的算法修改或架构。重要的是,与有竞争力的非保单方法相比,PPO通常在最终回报和样本效率方面都实现有竞争力或优越的结果。最后,通过消融研究,我们分析了实施和超参数的因素是至关重要的PPO的实证表现,并就这些因素给出具体的实用建议。我们的研究结果表明,当使用这些做法,简单的基于PPO的方法可以在合作多智能体强化学习强有力的基线。源代码在\url{https://github.com/marlbenchmark/on-policy}上发布。

英文摘要:

Proximal Policy Optimization (PPO) is a ubiquitous on-policy reinforcement learning algorithm but is significantly less utilized than off-policy learning algorithms in multi-agent settings. This is often due to the belief that PPO is significantly less sample efficient than off-policy methods in multi-agent systems. In this work, we carefully study the performance of PPO in cooperative multi-agent settings. We show that PPO-based multi-agent algorithms achieve surprisingly strong performance in four popular multi-agent testbeds: the particle-world environments, the StarCraft multi-agent challenge, Google Research Football, and the Hanabi challenge, with minimal hyperparameter tuning and without any domain-specific algorithmic modifications or architectures. Importantly, compared to competitive off-policy methods, PPO often achieves competitive or superior results in both final returns and sample efficiency. Finally, through ablation studies, we analyze implementation and hyperparameter factors that are critical to PPO's empirical performance, and give concrete practical suggestions regarding these factors. Our results show that when using these practices, simple PPO-based methods can be a strong baseline in cooperative multi-agent reinforcement learning. Source code is released at \url{https://github.com/marlbenchmark/on-policy}.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑