Symmetries-enhanced Multi-Agent Reinforcement Learning

TL;DR

This paper introduces a framework that embeds extrinsic symmetries into multi-agent reinforcement learning to improve generalization, scalability, and efficiency, especially in systems lacking intrinsic symmetries.

Contribution

It proposes the Group Equivariant Graphormer architecture for distributed swarming, enabling symmetry-enhanced learning in systems with few or no intrinsic symmetries.

Findings

Reduced collision rates in swarm tasks

Improved task success across scenarios

Enhanced zero-shot scalability

Abstract

Multi-agent reinforcement learning has emerged as a powerful framework for enabling agents to learn complex, coordinated behaviors but faces persistent challenges regarding its generalization, scalability and sample efficiency. Recent advancements have sought to alleviate those issues by embedding intrinsic symmetries of the systems in the policy. Yet, most dynamical systems exhibit little to no symmetries to exploit. This paper presents a novel framework for embedding extrinsic symmetries in multi-agent system dynamics that enables the use of symmetry-enhanced methods to address systems with insufficient intrinsic symmetries, expanding the scope of equivariant learning to a wide variety of MARL problems. Central to our framework is the Group Equivariant Graphormer, a group-modular architecture specifically designed for distributed swarming tasks. Extensive experiments on a swarm of symmetry-breaking quadrotors validate the effectiveness of our approach, showcasing its potential for improved generalization and zero-shot scalability. Our method achieves significant reductions in collision rates and enhances task success rates across a diverse range of scenarios and varying swarm sizes.