Scholarlink

The integration of Large Language Models (LLMs) with evolutionary computation has introduced a promising paradigm for automating the design of metaheuristic algorithms. However, existing frameworks, such as the Large Language Model Evolutionary Algorithm (LLaMEA), often lack precise control over mutation mechanisms, leading to inefficiencies in solution space exploration and potentially suboptimal convergence. This paper introduces a novel approach to mutation control within LLM-driven evolutionary frameworks, inspired by theory of genetic algorithms. Specifically, we propose dynamic mutation prompts that adaptively regulate mutation rates, leveraging a heavy-tailed power-law distribution to balance exploration and exploitation. Experiments using GPT-3.5-turbo and GPT-4o models demonstrate that the former fails to adhere to the specific mutation instructions. At the same time, the latter is able to control its mutation rate under the instruction of well-constructed mutation prompts. Further experiments show that the introduction of these dynamic rates can improve the convergence speed and adaptability of LLaMEA when using GPT-4o. This work sets a starting point for better controlled LLM-based mutations in code optimization tasks, paving the way for further advancements in automated metaheuristic design.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Controlling the Mutation in Large Language Models for the Efficient Evolution of Algorithms

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Haoran Yin,
  • Anna V. Kononova,
  • Thomas Bäck,
  • Niki van Stein

摘要

The integration of Large Language Models (LLMs) with evolutionary computation has introduced a promising paradigm for automating the design of metaheuristic algorithms. However, existing frameworks, such as the Large Language Model Evolutionary Algorithm (LLaMEA), often lack precise control over mutation mechanisms, leading to inefficiencies in solution space exploration and potentially suboptimal convergence. This paper introduces a novel approach to mutation control within LLM-driven evolutionary frameworks, inspired by theory of genetic algorithms. Specifically, we propose dynamic mutation prompts that adaptively regulate mutation rates, leveraging a heavy-tailed power-law distribution to balance exploration and exploitation. Experiments using GPT-3.5-turbo and GPT-4o models demonstrate that the former fails to adhere to the specific mutation instructions. At the same time, the latter is able to control its mutation rate under the instruction of well-constructed mutation prompts. Further experiments show that the introduction of these dynamic rates can improve the convergence speed and adaptability of LLaMEA when using GPT-4o. This work sets a starting point for better controlled LLM-based mutations in code optimization tasks, paving the way for further advancements in automated metaheuristic design.