PANDA：面向可扩展容错多智能体系统的去中心化架构
原标题：PANDA: A Decentralized Architecture with Flexible Orchestration for Scalable, Fault-Tolerant Multi-Agent Systems
AI 摘要
该论文提出 PANDA，一种面向 LLM 多智能体系统的去中心化架构，通过将集体通信与团队通信解耦，使大量异构智能体能够互相发现能力并按任务自组织成小型专业团队。PANDA 将底层架构与编排策略分离，支持 star、chain、mesh 三种规划执行模式，并利用 web-of-trust 模型在无中心服务的情况下实现治理。在 HotPotQA 基准上，PANDA 可扩展至数千个智能体、毫秒级组队，并在故障下保持 100% 任务完成率。
正文节选
PANDA: A Decentralized Architecture with Flexible Orchestration for Scalable, Fault-Tolerant Multi-Agent Systems Abstract Existing architectures for LLM-based multi-agent systems (MAS) cannot reliably and efficiently solve multi-step tasks at scale: they struggle to support large numbers of agents and concurrent tasks, tolerate failures, govern agent interactions, and accommodate the diverse planning and execution patterns different tasks require. We present panda, a decentralized architecture t