Dual Self-Awareness Value Decomposition Framework without Individual Global Max for Cooperative Multi-Agent Reinforcement Learning

TL;DR

This paper introduces a novel value decomposition framework for cooperative multi-agent reinforcement learning that does not rely on the traditional IGM assumption, enhancing problem-solving capabilities and avoiding local optima.

Contribution

It presents the first fully IGM-free value decomposition method with a dual self-awareness approach, including an explicit search procedure and anti-ego exploration mechanism.

Findings

Achieves competitive performance in cooperative tasks

Successfully avoids local optima in experiments

Introduces a new IGM-free value decomposition framework

Abstract

Value decomposition methods have gained popularity in the field of cooperative multi-agent reinforcement learning. However, almost all existing methods follow the principle of Individual Global Max (IGM) or its variants, which limits their problem-solving capabilities. To address this, we propose a dual self-awareness value decomposition framework, inspired by the notion of dual self-awareness in psychology, that entirely rejects the IGM premise. Each agent consists of an ego policy for action selection and an alter ego value function to solve the credit assignment problem. The value function factorization can ignore the IGM assumption by utilizing an explicit search procedure. On the basis of the above, we also suggest a novel anti-ego exploration mechanism to avoid the algorithm becoming stuck in a local optimum. As the first fully IGM-free value decomposition method, our proposed framework achieves desirable performance in various cooperative tasks.