arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.11883cs.LGcs.MA
收藏

正则化Softmax深度多智能体$Q$ -学习

Regularized Softmax Deep Multi-Agent $Q$-Learning

Tsinghua University(清华大学);University of Oxford(牛津大学)

Ling Pan, Tabish Rashid, Bei Peng, Longbo Huang, Shimon Whiteson

摘要:

解决$Q$学习中的高估问题是单智能体强化学习中广泛研究的一个重要问题,但在多智能体环境中受到的关注相对较少。在这项工作中,我们的经验表明,QMIX,一个流行的$Q$ -学习算法的合作多智能体强化学习(MARL),遭受更严重的高估在实践中比以前承认的,并没有减轻现有的方法。我们纠正这一点与一种新的正则化为基础的更新计划,惩罚大的联合行动值偏离基线,并证明其有效性,在稳定学习。此外,我们建议采用softmax算子,我们在多智能体环境中以一种新的方式有效地近似该算子,以进一步减少潜在的高估偏差。我们的方法,正则化Softmax(RES)深度多智能体$Q$ -学习,是通用的,可以应用于任何$Q$ -学习为基础的MARL算法。我们证明,当应用于QMIX,RES避免严重高估,并显着提高性能,产生各种合作的多智能体任务,包括具有挑战性的星际争霸II微观管理基准的最先进的结果。

英文摘要:

Tackling overestimation in $Q$-learning is an important problem that has been extensively studied in single-agent reinforcement learning, but has received comparatively little attention in the multi-agent setting. In this work, we empirically demonstrate that QMIX, a popular $Q$-learning algorithm for cooperative multi-agent reinforcement learning (MARL), suffers from a more severe overestimation in practice than previously acknowledged, and is not mitigated by existing approaches. We rectify this with a novel regularization-based update scheme that penalizes large joint action-values that deviate from a baseline and demonstrate its effectiveness in stabilizing learning. Furthermore, we propose to employ a softmax operator, which we efficiently approximate in a novel way in the multi-agent setting, to further reduce the potential overestimation bias. Our approach, Regularized Softmax (RES) Deep Multi-Agent $Q$-Learning, is general and can be applied to any $Q$-learning based MARL algorithm. We demonstrate that, when applied to QMIX, RES avoids severe overestimation and significantly improves performance, yielding state-of-the-art results in a variety of cooperative multi-agent tasks, including the challenging StarCraft II micromanagement benchmarks.

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。

↑