arXiv cs.MA· Wenwen Zheng, Yuzhe Yang, Helen Qu, Xin Eric Wang, Haewon Jeong·· 10 天前AI 评分22
多智能体搜索中的吸收态相变
Absorbing State Phase Transitions in Multi-Agent Search
AI 导读
研究用吸收态相变形式化方法预测 LLM 多智能体搜索任务的成功率，将搜索任务分为四类，并从理论上推导出临界通信度 d_c——每个智能体可通信的最小邻居数，超过该值错误假设便不会失控扩散、搜索进入已解状态。在软件配置调试与物理机制发现等真实任务上评估前沿 LLM 多智能体系统，结果与理论吻合程度不一：智能体可能不与邻居通信，并发展出对个体有利但限制协作收益的策略。
正文 · 原文
Abstract:Nontrivial dynamics can emerge in large language model (LLM)-based multi-agent systems, and preliminary evidence exists that formalisms from statistical mechanics can be effective at modeling and predicting such behaviors. In parallel, designing multi-agent communication topology for optimal task-solving is an active research question. In this paper, we focus on predicting the success of multi-agent search tasks using the formalism of absorbing state phase transitions. We first taxonomize search tasks into four types, informed by classical results in combinatorial search. We then theoretically derive a critical communication degree $d_c$, the minimum number of agents each agent can communicate with, above which incorrect hypotheses do not proliferate uncontrollably and the search enters the solved state. Finally, we evaluate frontier LLM-based multi-agent systems on real-world search and discovery tasks, software configuration debugging and physical mechanism discovery, and find that agreement with theory is mixed. LLM agents may not communicate with their neighbors and can develop strategies that are individually beneficial but limits the benefits of collaboration.
| Comments: | Code: this https URL |
| Subjects: | Multiagent Systems (cs.MA) |
| Cite as: | arXiv:2609.38327 [cs.MA] |
| (or arXiv:2609.38327v2 [cs.MA] for this version) | |
| https://doi.org/10.48550/arXiv.2609.38327 arXiv-issued DOI via DataCite |
Submission history
From: Helen Qu [view email]
[v1]
Tue, 29 Sep 2026 18:00:49 UTC (784 KB)
[v2]
Wed, 7 Oct 2026 17:19:56 UTC (784 KB)
来源：arXiv cs.MA · arxiv.org