MetaAgent: Automatically Constructing Multi-Agent Systems Based on Finite State Machines

2025-07-30

AI Summary

Current automated multi-agent design methods suffer from poor scenario generalization, limited tool integration, rigid communication architectures, and heavy reliance on external training data. To address these limitations, we propose a finite state machine (FSM)-based automated framework construction method that enables end-to-end system generation without training data, leveraging programmable state transitions and action control. The approach integrates large language models (LLMs) for task understanding and agent role orchestration, supporting dynamic tool invocation and adaptive communication topology reconfiguration. Experimental evaluation across multiple open-ended tasks demonstrates that systems generated by our method significantly outperform mainstream automated design baselines and achieve collaborative performance on par with manually fine-tuned systems. To the best of our knowledge, this is the first work to unify high flexibility, strong generalization, and zero-shot deployability in automated multi-agent system design.

Technology Category

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

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Large Language Models (LLMs) have demonstrated the ability to solve a wide range of practical tasks within multi-agent systems. However, existing human-designed multi-agent frameworks are typically limited to a small set of pre-defined scenarios, while current automated design methods suffer from several limitations, such as the lack of tool integration, dependence on external training data, and rigid communication structures. In this paper, we propose MetaAgent, a finite state machine based framework that can automatically generate a multi-agent system. Given a task description, MetaAgent will design a multi-agent system and polish it through an optimization algorithm. When the multi-agent system is deployed, the finite state machine will control the agent's actions and the state transitions. To evaluate our framework, we conduct experiments on both text-based tasks and practical tasks. The results indicate that the generated multi-agent system surpasses other auto-designed methods and can achieve a comparable performance with the human-designed multi-agent system, which is optimized for those specific tasks.

Problem

• Automates multi-agent system design for diverse tasks
• Overcomes limitations of rigid communication and tool integration
• Enhances performance comparable to human-designed systems

Innovation

• Automated multi-agent system generation
• Finite state machine control
• Optimization algorithm refinement