CoMIX: A Multi-agent Reinforcement Learning Training Architecture for Efficient Decentralized Coordination and Independent Decision-Making

TL;DR

CoMIX is a novel multi-agent reinforcement learning framework that enhances decentralized coordination and independent decision-making, enabling agents to adapt dynamically between selfish and collaborative behaviors for improved performance.

Contribution

This paper introduces CoMIX, a new training architecture that models incremental decision-making to improve coordination and independence in multi-agent systems.

Findings

CoMIX outperforms baseline methods on collaborative tasks.

The incremental decision process enhances agent coordination.

Agents adapt behavior dynamically between selfish and collaborative modes.

Abstract

Robust coordination skills enable agents to operate cohesively in shared environments, together towards a common goal and, ideally, individually without hindering each other's progress. To this end, this paper presents Coordinated QMIX (CoMIX), a novel training framework for decentralized agents that enables emergent coordination through flexible policies, allowing at the same time independent decision-making at individual level. CoMIX models selfish and collaborative behavior as incremental steps in each agent's decision process. This allows agents to dynamically adapt their behavior to different situations balancing independence and collaboration. Experiments using a variety of simulation environments demonstrate that CoMIX outperforms baselines on collaborative tasks. The results validate our incremental approach as effective technique for improving coordination in multi-agent systems.