EvoAgent: Towards Automatic Multi-Agent Generation via Evolutionary Algorithms

2024-06-20

arXiv.org

AI Summary

Current LLM-based agent systems face scalability challenges due to heavy reliance on manual architecture design, hindering automated construction of multi-agent systems. Method: This paper introduces the first evolutionary algorithm–based framework for automatic multi-agent system generation, leveraging mutation, crossover, and selection operators. It autonomously evolves any zero-shot single-agent into a functionally complementary and structurally diverse multi-agent system—without human intervention—and integrates seamlessly with mainstream LLM agent frameworks. Task-driven fitness evaluation and LLM-native orchestration enable adaptive optimization. Contribution/Results: This work pioneers the systematic application of evolutionary computation to multi-agent system synthesis, overcoming bottlenecks in manual architectural design. Empirical evaluation demonstrates significant improvements in problem-solving performance and generalization across diverse complex tasks. The complete codebase and toolchain are publicly released.

Technology Category

• Multiagent Systems: Multiagent Learning
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Intelligent Robots: Multi-Robot Systems

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• 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

Abstract

The rise of powerful large language models (LLMs) has spurred a new trend in building LLM-based autonomous agents for solving complex tasks, especially multi-agent systems. Despite the remarkable progress, we notice that existing works are heavily dependent on human-designed frameworks, which greatly limits the functional scope and scalability of agent systems. How to automatically extend the specialized agent to multi-agent systems to improve task-solving capability still remains a significant challenge. In this paper, we introduce EvoAgent, a generic method to automatically extend specialized agents to multi-agent systems via the evolutionary algorithm, thereby improving the effectiveness of LLM-based agents in solving tasks. Specifically, we consider the existing agent frameworks as the initial individual and then apply a series of evolutionary operators (e.g., mutation, crossover, selection, etc.) to generate multiple agents with diverse settings. Experimental results across various tasks show that EvoAgent can significantly enhance the task-solving capability of LLM-based agents, and can be generalized to any LLM-based agent framework to extend them into multi-agent systems. Resources are available at https://evo-agent.github.io/.

Problem

• Automatically extend specialized agents to multi-agent systems
• Improve task-solving capability of LLM-based agents
• Enhance scalability and functional scope of agent systems

Innovation

• Automates multi-agent system generation via evolutionary algorithms
• Enhances LLM-based agents' task-solving effectiveness
• Generalizes to any LLM-based agent framework