随时约束的多智能体强化学习

Oct, 2024

Anytime-Constrained Multi-Agent Reinforcement Learning

Jeremy McMahan, Xiaojin Zhu

TL;DR本研究针对多智能体环境中的随时约束问题，提出了随时约束均衡（ACE）的解决方案概念。研究提供了一个全面的随时约束马尔可夫游戏理论，包括可行策略的计算特征、ACE的固定参数可计算算法，以及近似计算可行ACE的多项式时间算法，显示出在最坏情况分析下，计算可行策略的近似保证是可能的最优解。

Abstract

We introduce Anytime Constraints to the Multi-Agent setting with the corresponding solution concept being anytime-constrained equilibrium (ACE). Then, we present a comprehensive theory of anytime-constrained