FoX: Formation-aware exploration in multi-agent reinforcement learning

TL;DR

This paper introduces FoX, a formation-aware exploration framework for multi-agent reinforcement learning that reduces the exploration space by focusing on meaningful formations, leading to improved performance in complex tasks.

Contribution

The paper proposes a novel formation-aware exploration method that guides agents to explore diverse formations, addressing scalability and exploration challenges in MARL.

Findings

FoX significantly outperforms state-of-the-art MARL algorithms.

The framework effectively reduces exploration space by formation-based equivalence.

Demonstrated success on Google Research Football and Starcraft II tasks.

Abstract

Recently, deep multi-agent reinforcement learning (MARL) has gained significant popularity due to its success in various cooperative multi-agent tasks. However, exploration still remains a challenging problem in MARL due to the partial observability of the agents and the exploration space that can grow exponentially as the number of agents increases. Firstly, in order to address the scalability issue of the exploration space, we define a formation-based equivalence relation on the exploration space and aim to reduce the search space by exploring only meaningful states in different formations. Then, we propose a novel formation-aware exploration (FoX) framework that encourages partially observable agents to visit the states in diverse formations by guiding them to be well aware of their current formation solely based on their own observations. Numerical results show that the proposed FoX framework significantly outperforms the state-of-the-art MARL algorithms on Google Research Football (GRF) and sparse Starcraft II multi-agent challenge (SMAC) tasks.