DecArena：评测大规模去中心化多智能体协作
DecArena: Evaluating Large-Scale Decentralized Multi-Agent Collaboration
AI 导读
DecArena 是一个面向去中心化多智能体协作的基准，在图约束通信下提供可扩展的分布式工具调用任务，包含 100K+ 工具。实验显示现有 LLM 智能体难以发现非本地资源并协作，标准强化学习也无法可靠解决该问题。作者提出的 GraphHarness 基线结合拓扑感知预热与自改进 LLM 路由，使任务成功率提升 +48.5、协作质量提升 +30.6、成本降低 +34.0。
正文 · 原文
Abstract:Emerging personal AI agents raise fundamental scalability and privacy concerns for centralized multi-agent systems (MASs), which rely on a global orchestrator with broad access to agents, resources, and communication. In decentralized environments, global visibility is unavailable. Agents need to proactively discover non-local resources (e.g. tools, skills), and collaborate with other agents to solve tasks. We study decentralized multi-agent collaboration under graph-constrained communication, a natural reflection of human structures. Each agent observes only its local state and can interact only with neighboring agents. We introduce a benchmark with synthetic scalable distributed tool calling with 100K+ tools. Our experiments reveal that recent LLM agents struggle to discover non-local relevant resources and collaborate in decentralized setting. Moreover, standard reinforcement learning fails to reliably overcome this limitation, suggesting that decentralized MAS needs more explicit harness mechanisms for effective collaboration, even using SOTA LLMs. We then present a strong baseline \textbf{GraphHarness}, combining topology-aware warmup with self-improving LLM-based routing. Results show that GraphHarness substantially improves task success (+48.5) and produces higher quality collaboration (+30.6), while using fewer costs (+34.0) than baselines.
| Subjects: | Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2611.15521 [cs.MA] |
| (or arXiv:2611.15521v1 [cs.MA] for this version) | |
| https://doi.org/10.48550/arXiv.2611.15521 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jiali Cheng [view email]
[v1]
Wed, 11 Nov 2026 14:13:26 UTC (234 KB)
来源：arXiv cs.AI · arxiv.org