arXiv cs.MA· Hannes B\"uchi, Manon Flageat, Eduardo Sebasti\'an, Amanda Prorok·· 11 小时前AI 评分31
以事件为触发器的多智能体强化学习行为多样性框架
Events as Triggers for Behavioral Diversity in Multi-Agent Reinforcement Learning
AI 导读
研究者提出一种将智能体身份与行为解耦的多智能体强化学习框架，通过事件触发行为切换，让智能体在特定时刻承担不同角色。该框架包含 Neural Manifold Diversity（NMD）距离度量和基于事件的超网络，后者在共享团队策略上生成 LoRA 模块，实现智能体策略的即时重配置。实验显示其性能超越现有基线，具备零样本泛化能力，且是唯一能解决需要顺序行为重分配任务的方法。
正文 · 原文
Abstract:Effective multi-agent cooperation requires agents to adopt diverse behaviors as task conditions evolve-and to do so at the right moment. Yet, current Multi-Agent Reinforcement Learning (MARL) frameworks that facilitate this diversity are still limited by the fact that they bind fixed behaviors to fixed agent identities. Consequently, they are ill-equipped for tasks where agents need to take on different roles at very specific moments in time. We argue that, to define these behavioral transitions, the missing ingredient is $\textbf{events}$. Events are changes in the state of the system that induce qualitative changes in the task. Based on this view, we introduce a framework that decouples agent identity from behavior, capturing a continuous manifold from which agents instantiate their behaviors in response to events. This framework is based on two elements. First, to build an expressive behavior manifold, we introduce Neural Manifold Diversity (NMD), a formal distance metric that remains well-defined when behaviors are transient and agent-agnostic. Second, we use an event-based hypernetwork that generates Low-Rank Adaptation (LoRA) modules over a shared team policy, enabling on-the-fly agent-policy reconfiguration in response to events. We prove that this construction ensures that diversity does not interfere with reward maximization by design. Empirical results demonstrate that our framework outperforms established baselines across benchmarks while exhibiting zero-shot generalization, and being the only method that solves tasks requiring sequential behavior reassignment.
| Subjects: | Multiagent Systems (cs.MA); Machine Learning (cs.LG) |
| Cite as: | arXiv:2605.12388 [cs.MA] |
| (or arXiv:2605.12388v3 [cs.MA] for this version) | |
| https://doi.org/10.48550/arXiv.2605.12388 arXiv-issued DOI via DataCite |
Submission history
From: Hannes Büchi [view email]
[v1]
Tue, 12 May 2026 16:51:23 UTC (7,041 KB)
[v2]
Wed, 13 May 2026 11:22:03 UTC (7,041 KB)
[v3]
Thu, 29 Oct 2026 21:34:42 UTC (6,942 KB)
来源：arXiv cs.MA · arxiv.org