MARL-LNS: Cooperative Multi-agent Reinforcement Learning via Large Neighborhoods Search

TL;DR

This paper introduces MARL-LNS, a novel framework for cooperative multi-agent reinforcement learning that improves training efficiency by alternating subsets of agents without additional parameters, achieving faster convergence.

Contribution

The paper proposes a general framework MARL-LNS and three variants that enhance training efficiency in multi-agent RL without extra parameters.

Findings

Reduces training time by at least 10%

Achieves same skill level as baseline algorithms

Effective on StarCraft and Google Football environments

Abstract

Cooperative multi-agent reinforcement learning (MARL) has been an increasingly important research topic in the last half-decade because of its great potential for real-world applications. Because of the curse of dimensionality, the popular "centralized training decentralized execution" framework requires a long time in training, yet still cannot converge efficiently. In this paper, we propose a general training framework, MARL-LNS, to algorithmically address these issues by training on alternating subsets of agents using existing deep MARL algorithms as low-level trainers, while not involving any additional parameters to be trained. Based on this framework, we provide three algorithm variants based on the framework: random large neighborhood search (RLNS), batch large neighborhood search (BLNS), and adaptive large neighborhood search (ALNS), which alternate the subsets of agents differently. We test our algorithms on both the StarCraft Multi-Agent Challenge and Google Research Football, showing that our algorithms can automatically reduce at least 10% of training time while reaching the same final skill level as the original algorithm.