Recursive Reasoning Graph for Multi-Agent Reinforcement Learning

TL;DR

This paper introduces the Recursive Reasoning Graph (R2G), a novel multi-agent reinforcement learning algorithm that enhances agents' ability to anticipate others' responses, leading to improved cooperation and competition in complex multi-agent environments.

Contribution

The paper proposes R2G, a recursive reasoning model within a centralized-training-decentralized-execution framework, achieving state-of-the-art results in multi-agent tasks.

Findings

R2G outperforms existing algorithms in multi-agent particle games.

R2G demonstrates superior performance in robotics simulation environments.

Recursive reasoning improves strategic interactions among agents.

Abstract

Multi-agent reinforcement learning (MARL) provides an efficient way for simultaneously learning policies for multiple agents interacting with each other. However, in scenarios requiring complex interactions, existing algorithms can suffer from an inability to accurately anticipate the influence of self-actions on other agents. Incorporating an ability to reason about other agents' potential responses can allow an agent to formulate more effective strategies. This paper adopts a recursive reasoning model in a centralized-training-decentralized-execution framework to help learning agents better cooperate with or compete against others. The proposed algorithm, referred to as the Recursive Reasoning Graph (R2G), shows state-of-the-art performance on multiple multi-agent particle and robotics games.