BriefGPT.xyz
Feb, 2022
同质化马尔可夫博弈的高效通信演员-评论方法
Communication-Efficient Actor-Critic Methods for Homogeneous Markov Games
HTML
PDF
Dingyang Chen, Yile Li, Qi Zhang
TL;DR
该论文研究了协作多智能体强化学习中的集中式训练和策略共享,提出了一种基于一致性的去中心化演员-评论家方法,以减少通信成本并保证收敛,从而有效地降低了训练时的通信成本。
Abstract
Recent success in
cooperative multi-agent reinforcement learning
(MARL) relies on
centralized training
and
policy sharing
.
→