PMAT: Optimizing Action Generation Order in Multi-Agent Reinforcement Learning

Multi-agent reinforcement learning (MARL) faces challenges in coordinating agents due to complex interdependencies within multi-agent systems. Most MARL algorithms use the simultaneous decision-making paradigm but ignore the action-level dependencies among agents, which reduces coordination efficiency. In contrast, the sequential decision-making paradigm provides finer-grained supervision for agent decision order, presenting the potential for handling dependencies via better decision order management. However, determining the optimal decision order remains a challenge. In this paper, we introduce Action Generation with Plackett-Luce Sampling (AGPS), a novel mechanism for agent decision order optimization. We model the order determination task as a Plackett-Luce sampling process to address issues such as ranking instability and vanishing gradient during the network training process. AGPS realizes credit-based decision order determination by establishing a bridge between the significance of agents' local observations and their decision credits, thus facilitating order optimization and dependency management. Integrating AGPS with the Multi-Agent Transformer, we propose the Prioritized Multi-Agent Transformer (PMAT), a sequential decision-making MARL algorithm with decision order optimization. Experiments on benchmarks including StarCraft II Multi-Agent Challenge, Google Research Football, and Multi-Agent MuJoCo show that PMAT outperforms state-of-the-art algorithms, greatly enhancing coordination efficiency.

@article{hu2025_2502.16496,
  title={ PMAT: Optimizing Action Generation Order in Multi-Agent Reinforcement Learning },
  author={ Kun Hu and Muning Wen and Xihuai Wang and Shao Zhang and Yiwei Shi and Minne Li and Minglong Li and Ying Wen },
  journal={arXiv preprint arXiv:2502.16496},
  year={ 2025 }
}