Evolutionary Computation and Large Language Models: A Survey of Methods, Synergies, and Applications

TL;DR

The paper surveys the bidirectional synergy between Evolutionary Computation (EC) and Large Language Models (LLMs), detailing how EC can optimize LLM prompts, hyperparameters, and architectures while LLMs can automate metaheuristic design, surrogate modeling, and adaptive search strategies. It introduces and compare frameworks such as EvoPrompt, PhaseEvo, GAPPO, LLaMEA, LMEA, and MEoH, and discusses co-evolutionary paradigms, novel metaheuristic design, and decision-support enhancements. The analysis highlights significant potential for automated, data-efficient optimization across prompting, architecture search, and problem-solving, alongside substantial challenges in computational cost, benchmarking, and interpretability. The work argues for hybrid, co-adaptive architectures and a future research agenda to realize scalable, explainable, and robust EC–LLM systems with broad cross-domain impact.

Abstract

Integrating Large Language Models (LLMs) and Evolutionary Computation (EC) represents a promising avenue for advancing artificial intelligence by combining powerful natural language understanding with optimization and search capabilities. This manuscript explores the synergistic potential of LLMs and EC, reviewing their intersections, complementary strengths, and emerging applications. We identify key opportunities where EC can enhance LLM training, fine-tuning, prompt engineering, and architecture search, while LLMs can, in turn, aid in automating the design, analysis, and interpretation of ECs. The manuscript explores the synergistic integration of EC and LLMs, highlighting their bidirectional contributions to advancing artificial intelligence. It first examines how EC techniques enhance LLMs by optimizing key components such as prompt engineering, hyperparameter tuning, and architecture search, demonstrating how evolutionary methods automate and refine these processes. Secondly, the survey investigates how LLMs improve EC by automating metaheuristic design, tuning evolutionary algorithms, and generating adaptive heuristics, thereby increasing efficiency and scalability. Emerging co-evolutionary frameworks are discussed, showcasing applications across diverse fields while acknowledging challenges like computational costs, interpretability, and algorithmic convergence. The survey concludes by identifying open research questions and advocating for hybrid approaches that combine the strengths of EC and LLMs.

Figures (10)

• Figure 1: Organization of the paper.
• Figure 2: Components of a typical LLM Prompt.
• Figure 3: A General Framework of EC for Prompt Engineering.
• Figure 4: Experimental workflow.
• Figure 5: Prompt and procedure for regression task.