arXiv：cs.AI· Zhezheng Hao, Tianfu Wang, Huanshuo Dong, Ziyan Liu, Hong Wang, Xiankun Lin, Qiang Lin, Can Wang, Hande Dong, Jiawei Chen·· 20 小时前
Meta-Team：面向 LLM 多智能体系统的协作式自进化框架
Evolve as a Team: Collaborative Self-Evolution for LLM-based Multi-Agent Systems
AI 导读
Meta-Team 是一个基于协作自进化的经验驱动 LLM 多智能体系统（MAS）进化框架，通过保留每个智能体的执行上下文并协调任务后通信，让智能体交换分布式证据以支持进化。该框架将执行经验转化为对智能体行为、智能体间协作和团队级组织的可复用改进。在六个长时程智能体基准上，Meta-Team 持续优于单智能体系统、手工设计的 MAS 及此前的 MAS 进化方法。
正文
Abstract:LLM-based multi-agent systems (MAS) have emerged as an effective paradigm for complex and long-horizon tasks. However, in real-world tasks, MAS often exhibit various failures during execution and such failures are difficult to eliminate during design. This motivates experience-driven MAS evolution, where a system improves based on its own execution experience. Yet such evolution is challenging because MAS experience is prolonged and intricate, interleaving multiple agents' execution chains and communication messages, which makes it difficult to identify what should be improved. To address this challenge, we propose Meta-Team, an experience-driven MAS evolution framework based on collaborative self-evolution. Meta-Team preserves the execution context of each agent and coordinates post-task communication, enabling agents to exchange distributed evidence for evolution. Building on this design, Meta-Team conducts multi-scale self-evolution, transforming execution experience into reusable improvements to agent behaviors, inter-agent coordination, and team-level organization. Across six long-horizon agent benchmarks, Meta-Team consistently outperforms single-agent systems, hand-crafted MAS, and prior MAS evolution methods; further analyses demonstrate that Meta-Team enables more reliable and scalable MAS self-evolution.
| Subjects: | Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2605.29790 [cs.MA] |
| (or arXiv:2605.29790v2 [cs.MA] for this version) | |
| https://doi.org/10.48550/arXiv.2605.29790 arXiv-issued DOI via DataCite |
Submission history
From: Zhezheng Hao [view email]
[v1]
Thu, 28 May 2026 11:40:16 UTC (1,317 KB)
[v2]
Tue, 13 Oct 2026 07:59:39 UTC (1,378 KB)
来源：arXiv：cs.AI · arxiv.org