SC-MAS: Constructing Cost-Efficient Multi-Agent Systems with Edge-Level Heterogeneous Collaboration

2026-01-14

AI Summary

This work addresses the limitations of existing large language model–based multi-agent systems, which predominantly rely on homogeneous collaboration and struggle to balance performance with inference cost. Inspired by social capital theory, the authors model the system as a directed graph and introduce, for the first time, an edge-level heterogeneous collaboration mechanism. This approach enables customized communication strategies between distinct agent pairs and employs a unified controller to dynamically construct task-efficient execution structures. Evaluated on the MMLU and MBPP benchmarks, the proposed method improves accuracy by 3.35% and 3.53%, respectively, while simultaneously reducing inference costs by 15.38% and 12.13%, significantly outperforming conventional homogeneous collaboration paradigms.

Technology Category

• Multiagent Systems: Coordination and Collaboration
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Large Language Model (LLM)-based Multi-Agent Systems (MAS) enhance complex problem solving through multi-agent collaboration, but often incur substantially higher costs than single-agent systems. Recent MAS routing methods aim to balance performance and overhead by dynamically selecting agent roles and language models. However, these approaches typically rely on a homogeneous collaboration mode, where all agents follow the same interaction pattern, limiting collaboration flexibility across different roles. Motivated by Social Capital Theory, which emphasizes that different roles benefit from distinct forms of collaboration, we propose SC-MAS, a framework for constructing heterogeneous and cost-efficient multi-agent systems. SC-MAS models MAS as directed graphs, where edges explicitly represent pairwise collaboration strategies, allowing different agent pairs to interact through tailored communication patterns. Given an input query, a unified controller progressively constructs an executable MAS by selecting task-relevant agent roles, assigning edge-level collaboration strategies, and allocating appropriate LLM backbones to individual agents. Experiments on multiple benchmarks demonstrate the effectiveness of SC-MAS. In particular, SC-MAS improves accuracy by 3.35% on MMLU while reducing inference cost by 15.38%, and achieves a 3.53% accuracy gain with a 12.13% cost reduction on MBPP. These results validate the feasibility of SC-MAS and highlight the effectiveness of heterogeneous collaboration in multi-agent systems.

Problem

• Multi-Agent Systems
• Heterogeneous Collaboration
• Cost Efficiency
• Large Language Models
• Agent Interaction

Innovation

• heterogeneous collaboration
• edge-level strategy
• multi-agent system
• cost-efficient LLM
• Social Capital Theory