Exploring the Improvement of Evolutionary Computation via Large Language Models

TL;DR

Evolutionary Computation (EC) struggles with large search spaces, local optima, and heavy reliance on problem-specific human design. The paper surveys integrating large language models (LLMs) into EC, focusing on LLM-driven EA algorithms, LLM-assisted population and individual design, and broader improvements such as multimodal and interactive EC. Concrete insights include LMEA for strategy selection and the use of LLMs as, or to guide, evolutionary operators, with noted benefits for combinatorial optimization and reduced expert tuning. The proposed integration aims to make EC more automated, scalable, and adaptable to dynamic, language-rich tasks, potentially expanding EC's applicability across domains.

Abstract

Evolutionary computation (EC), as a powerful optimization algorithm, has been applied across various domains. However, as the complexity of problems increases, the limitations of EC have become more apparent. The advent of large language models (LLMs) has not only transformed natural language processing but also extended their capabilities to diverse fields. By harnessing LLMs' vast knowledge and adaptive capabilities, we provide a forward-looking overview of potential improvements LLMs can bring to EC, focusing on the algorithms themselves, population design, and additional enhancements. This presents a promising direction for future research at the intersection of LLMs and EC.