MOC: Multi-Order Communication in LLM-based Multi-Agent Systems
2026-06-01 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors point out that while many systems using large language models focus on how agents coordinate, less attention is given to how agents share messages effectively. They show that simply combining direct neighbor messages can miss important information over multiple steps. To fix this, the authors introduce a new communication method called Multi-Order Communication (MOC) that captures deeper message connections and smartly merges information to keep key details. Their tests on different datasets and models show MOC helps agents perform better while using less communication.
Large Language ModelsMulti-Agent SystemsCommunication ProtocolsMessage PassingMulti-Hop DependenciesSemantic MergingCoordination TopologyToken ConstraintsEvidence Receptive Field
Authors
Yao Guan, Lin Wang, Zhihu Lu, Ziyi Wang, Wenzhu Yan, Qiang Duan
Abstract
Despite the remarkable progress of Large Language Model (LLM) based Multi-Agent Systems, most research focuses on optimizing coordination topology while largely underexploring the equally critical problem: how to transmit and optimize messages among agents effectively? Current communication schemes typically rely on the direct concatenation of first-order neighbor responses, which induces a restricted evidence receptive field and leads to the dilution of crucial insights over multi-hop paths. To address these limitations, we propose the Multi-Order Communication (MOC) scheme, which reconstructs the inter-agent communication to capture multi-hop dependencies and incorporates a structural message consolidation strategy to ensure efficiency. Specifically, we formalize the communication mechanism to construct a structured multi-order evidence stream, and subsequently design a Semantic-Topological Merging algorithm to optimize semantic fidelity within token constraints. Extensive experiments across six diverse datasets and LLM backbones of varying parameter scales demonstrate that MOC consistently improves task performance and reduces communication costs. The code is available at https://github.com/yao-guan/MOC.