Metaheuristics and Large Language Models Join Forces: Toward an Integrated Optimization Approach

2024/05/28 by Camilo Chacón Sartori, Christian Blum, Filippo Bistaffa +1 · 1 voice · 1 citation
Computer Science · #Natural Language Processing Techniques #cs.AI

paper · pdf · doi:10.1109/access.2024.3524176

arxiv published 2024/05/28 · openalex publication_date 2024/12/30 · openalex created_date 2024/12/31 · arxiv updated 2025/02/12 · openalex updated_date 2026/07/29

Abstract

Since the rise of Large Language Models (LLMs) a couple of years ago, researchers in metaheuristics (MHs) have wondered how to use their power in a beneficial way within their algorithms. This paper introduces a novel approach that leverages LLMs as pattern recognition tools to improve MHs. The resulting hybrid method, tested in the context of a social network-based combinatorial optimization problem, outperforms existing state-of-the-art approaches that combine machine learning with MHs regarding the obtained solution quality. By carefully designing prompts, we demonstrate that the output obtained from LLMs can be used as problem knowledge, leading to improved results. Lastly, we acknowledge LLMs’ potential drawbacks and limitations and consider it essential to examine them to advance this type of research further. Our method can be reproduced using a tool available at:https://github.com/camilochs/optipattern.

Citations

• Machine learning for combinatorial optimization: A methodological tour d’horizon
• On the Dangers of Stochastic Parrots
• Machine learning at the service of meta-heuristics for solving combinatorial optimization problems: A state-of-the-art
• On random graphs. I.
• AI as Agency Without Intelligence: on ChatGPT, Large Language Models, and Other Generative Models
• Exploiting Generative AI to Scale up Intelligent Tutoring Systems

Cited by

• RAG/LLM Augmented Switching Driven Polymorphic Metaheuristic Framework

Discussions

• Metaheuristics and Large Language Models Join Forces [hn, 2 points, 0 comments]