LLaMEA is a groundbreaking framework that uses Large Language Models to create and evolve metaheuristic algorithms automatically, eliminating the need for human algorithm designers.

• Applies evolutionary principles to algorithm generation using LLMs like GPT-4
• Creates increasingly performant algorithms through feedback cycles of generation, mutation, and selection
• Enables the automatic discovery of new optimization techniques without domain expertise
• Demonstrates how LLMs can move beyond code generation to become creative engineering partners

This research transforms how optimization algorithms are developed for engineering applications, reducing development time while potentially discovering novel approaches human experts might overlook.

LLaMEA: A Large Language Model Evolutionary Algorithm for Automatically Generating Metaheuristics