D3MAS: Decompose, Deduce, and Distribute for Enhanced Knowledge Sharing in Multi-Agent Systems

2025-10-12

AI Summary

Multi-agent systems commonly suffer from knowledge redundancy, leading to redundant computations during retrieval and reasoning; existing architectures lack mechanisms to ensure stage-wise sharing of “minimal sufficient information.” This paper proposes D3MAS, a hierarchical collaborative framework comprising three layers: task decomposition, cooperative reasoning, and distributed memory. Crucially, it introduces a structured message-passing mechanism built upon a unified heterogeneous graph, enabling cross-layer semantic alignment and eliminating redundancy at the architectural level—rather than relying on post-hoc optimization. Experiments across four benchmark datasets demonstrate that D3MAS improves reasoning accuracy by 8.7–15.6%, reduces average knowledge redundancy by 46%, and significantly enhances both collaborative efficiency and scalability of multi-agent systems.

Technology Category

• Multiagent Systems: Distributed Problem Solving
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Knowledge Representation and Reasoning: Computational Complexity of Reasoning

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Multi-agent systems powered by large language models exhibit strong capabilities in collaborative problem-solving. However, these systems suffer from substantial knowledge redundancy. Agents duplicate efforts in retrieval and reasoning processes. This inefficiency stems from a deeper issue: current architectures lack mechanisms to ensure agents share minimal sufficient information at each operational stage. Empirical analysis reveals an average knowledge duplication rate of 47.3% across agent communications. We propose D3MAS (Decompose, Deduce, and Distribute), a hierarchical coordination framework addressing redundancy through structural design rather than explicit optimization. The framework organizes collaboration across three coordinated layers. Task decomposition filters irrelevant sub-problems early. Collaborative reasoning captures complementary inference paths across agents. Distributed memory provides access to non-redundant knowledge. These layers coordinate through structured message passing in a unified heterogeneous graph. This cross-layer alignment ensures information remains aligned with actual task needs. Experiments on four challenging datasets show that D3MAS consistently improves reasoning accuracy by 8.7% to 15.6% and reduces knowledge redundancy by 46% on average.

Problem

• Multi-agent systems suffer from high knowledge redundancy in collaborative problem-solving
• Current architectures lack mechanisms for minimal sufficient information sharing
• Agents duplicate efforts in retrieval and reasoning processes significantly

Innovation

• Hierarchical framework coordinates three structured layers
• Task decomposition filters irrelevant sub-problems early
• Distributed memory provides non-redundant knowledge access