Multi-Agent Systems: From Classical Paradigms to Large Foundation Model-Enabled Futures

2026-04-20

AI Summary

Traditional multi-agent systems exhibit limited semantic understanding and flexible coordination in complex, open environments, hindering their adaptability to diverse task requirements. This work pioneers the integration of large foundation models into the analytical framework of multi-agent systems, establishing a closed-loop collaborative mechanism centered on perception, communication, decision-making, and control. Through a systematic review and comparative analysis, the study elucidates key differences between classical architectures and large-model-driven systems in terms of structure, coordination mechanisms, adaptability, and application scenarios. It advances multi-agent collaboration from mere state exchange toward semantic reasoning, highlighting breakthroughs enabled by large models in adaptability, scalability, and semantic-aware coordination, while outlining critical challenges and promising directions for future research.

Technology Category

• Multiagent Systems: Coordination and Collaboration
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Intelligent Robots: Multi-Robot Systems

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

With the rapid advancement of artificial intelligence, multi-agent systems (MASs) are evolving from classical paradigms toward architectures built upon large foundation models (LFMs). This survey provides a systematic review and comparative analysis of classical MASs (CMASs) and LFM-based MASs (LMASs). First, within a closed-loop coordination framework, CMASs are reviewed across four fundamental dimensions: perception, communication, decision-making, and control. Beyond this framework, LMASs integrate LFMs to lift collaboration from low-level state exchanges to semantic-level reasoning, enabling more flexible coordination and improved adaptability across diverse scenarios. Then, a comparative analysis is conducted to contrast CMASs and LMASs across architecture, operating mechanism, adaptability, and application. Finally, future perspectives on MASs are presented, summarizing open challenges and potential research opportunities.

Problem

• Multi-Agent Systems
• Large Foundation Models
• Classical Paradigms
• Coordination
• Adaptability

Innovation

• Large Foundation Models
• Multi-Agent Systems
• Semantic-Level Reasoning
• Closed-Loop Coordination
• Adaptability