---
title: "多智能体LLM系统的组织科学：解耦谁、如何与哪个算法"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2607.25446v1"
permalink: "https://feed.astralor.com/items/itm_8409a158dfbfc610ae21ffa4"
published: "2026-07-28T08:35:21.000Z"
score: 81
channel: "ai"
tags: ["IMACS","LLM","上下文赌博机","多智能体系统","组织理论","自适应路由"]
---

# 多智能体LLM系统的组织科学：解耦谁、如何与哪个算法

## AI 点评

> 这篇论文把管理学里的组织理论真枪实弹地搬进了多智能体系统，不是简单类比，而是可执行可验证。对搞Agent编排的工程师和研究者来说，这是一份既讲原理又给代码的实操地图，尤其注意它证明了组织设计不能一招鲜——换了模型就得重新调。

## 速览

多智能体框架常将组织、协调和协作协议三个维度纠缠在一起。本文提出IMACS框架，将三者解耦为独立可互换的层，使经典组织理论（Belbin角色、Mintzberg协调、RACI责任）成为可执行的配置，并将六种已有协作协议统一接口。进一步提出自适应组织路由（Adaptive Org Routing）——一种基于上下文赌博机的元协议，能在质量-成本权衡下为每个任务动态选择协议，在受控研究中优于所有固定协议。消融实验发现，责任分配的效果取决于协议是否将交付物路由给负责智能体，且最优分配因模型族而异，证明组织设计必须针对每个模型重新验证或学习。

## 关键点

- 提出IMACS框架，解耦多智能体系统的组织（Who）、协调（How）和协作协议（Which Algorithm）三个层面。
- 将Belbin角色、Mintzberg协调机制、RACI责任矩阵等经典组织理论映射为可执行配置。
- 自适应组织路由（Adaptive Org Routing）作为上下文赌博机元协议，动态选择最佳协作协议。
- 实验揭示责任分配的效果依赖于协议路径，且最优组织设计因模型族而异，不可硬编码。

## 全文

Multi-agent frameworks built on large language models (LLMs) routinely entangle three logically distinct concerns: who is on the team (organization), how members align (coordination), and which algorithm fuses their work (collaboration protocol). IMACS (Intelligent Multi-Agent Collaboration System) separates the three into orthogonal, independently swappable layers. Classic organizational theory (Belbin roles, Mintzberg coordination, RACI accountability) becomes executable, validated configuration, and the framework places six published collaboration algorithms behind a common interface while exposing roles, coordination, and accountability as independently configurable factors. We use this separation to conduct controlled comparisons in which organizational assignments vary while the collaboration protocol is held fixed. It also turns protocol choice into a variable that can be learned: Adaptive Org Routing, a contextual-bandit meta-protocol, selects a protocol per task under an explicit quality-cost tradeoff, outperforms every fixed protocol in a controlled study, and trains online on real benchmark and LLM-judge rewards. The ablations expose a mechanism. Accountability placement changes outcomes exactly when the protocol routes the deliverable through the accountable agent, and the winning placement flips across model families, so organizational design cannot be hard-coded; it must be revalidated, or learned, for each model binding.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_8409a158dfbfc610ae21ffa4