Coordination Graphs for Constrained Multi-Agent Reinforcement Learning
约束多智能体强化学习的协调图
机构 * Department of Electrical and Computer Engineering, Linköping University(1 链çe普大学电气与计算机工程系)
专题命中 模仿学习与强化学习 :navigation(abstract);分类 cs.AI
AI总结 提出CG-CMARL框架,利用协调图和拉格朗日对偶分解联合动作空间与约束耦合问题,实现独立于智能体数量的模型学习,并通过Max-Sum消息传递和拉格朗日乘子控制目标-约束权衡,生成帕累托前沿。
Comments Accepted at the Reinforcement Learning Conference (RLC) 2026. 40 pages (12 main + 28 appendix), 5 figures, 16 tables, 7 theorems