Value-Guidance MeanFlow for Offline Multi-Agent Reinforcement Learning

TL;DR

VGM$^2$P introduces a flow-based, coefficient-insensitive policy learning framework for offline multi-agent reinforcement learning, improving efficiency and performance by leveraging advantage-guided behavior cloning and classifier-free guidance.

Contribution

It proposes a novel flow-based policy method that enhances efficiency and expressiveness in offline multi-agent RL without sensitivity to behavior regularization coefficients.

Findings

Achieves state-of-the-art performance with only behavior cloning.

Demonstrates efficiency in both discrete and continuous action tasks.

Maintains performance without iterative sampling or distillation.

Abstract

Offline multi-agent reinforcement learning (MARL) aims to learn the optimal joint policy from pre-collected datasets, requiring a trade-off between maximizing global returns and mitigating distribution shift from offline data. Recent studies use diffusion or flow generative models to capture complex joint policy behaviors among agents; however, they typically rely on multi-step iterative sampling, thereby reducing training and inference efficiency. Although further research improves sampling efficiency through methods like distillation, it remains sensitive to the behavior regularization coefficient. To address the above-mentioned issues, we propose Value Guidance Multi-agent MeanFlow Policy (VGM$^2$P), a simple yet effective flow-based policy learning framework that enables efficient action generation with coefficient-insensitive conditional behavior cloning. Specifically, VGM$^2$P uses global advantage values to guide agent collaboration, treating optimal policy learning as conditional behavior cloning. Additionally, to improve policy expressiveness and inference efficiency in multi-agent scenarios, it leverages classifier-free guidance MeanFlow for both policy training and execution. Experiments on tasks with both discrete and continuous action spaces demonstrate that, even when trained solely via conditional behavior cloning, VGM$^2$P efficiently achieves performance comparable to state-of-the-art methods.