LLM-Driven Co-Evolutionary Automated Heuristic Design for Bi-Component Coupled Combinatorial Optimization

2026-05-30

AI Summary

This work addresses the challenge of effectively handling the strong coupling between decision substructures—such as route planning and item selection—in the Traveling Thief Problem (TTP) and the Traveling Purchaser Problem (TPP), which existing automated heuristic design methods struggle to manage. To this end, we propose CoEvo-AHD, a novel framework that integrates large language models with a dual-population co-evolutionary algorithm. The framework jointly optimizes two complementary sets of heuristic operators through collaborative evaluation, pairwise scoring, and co-evolutionary crossover operations. It further incorporates a standardized tool-calling library and an incremental local search mechanism to enable efficient generation and evolution of high-quality heuristic combinations within the coupled decision space. Experimental results demonstrate that the heuristics automatically generated by CoEvo-AHD achieve solution quality on TTP and TPP benchmarks comparable to that of handcrafted approaches.

Abstract

While Large Language Models (LLMs) have recently shown promise in Automated Heuristic Design (AHD), existing methods typically generate and evolve heuristics as a single operator or search strategy, limiting their ability to model strong coupling among multiple decision substructures in problems such as the Traveling Thief Problem (TTP) and the Traveling Purchaser Problem (TPP). In this work, we propose CoEvo-AHD, an LLM-driven dual-population co-evolutionary framework for automated heuristic design in coupled combinatorial optimization. Unlike prior methods that evolve individual heuristics in isolation, CoEvo-AHD leverages LLMs to co-evolve two closely related operator populations. A cooperative evaluation mechanism explicitly captures interactions between route and selection operators, while pairwise scoring and synergistic joint crossover help discover complementary operator logic for joint improvement across coupled decision subspaces. We further design a tool-invocation environment library that encapsulates frequently used core operations, such as local-search delta computation, into callable functions, enabling LLM-generated operators to use standardized interfaces instead of reimplementing inefficient and error-prone problem-specific loops. Experiments on TTP and TPP show that CoEvo-AHD automatically discovers cooperative heuristic combinations and achieves competitive solution quality against traditional heuristics.

Problem

• coupled combinatorial optimization
• automated heuristic design
• bi-component coupling
• Large Language Models
• co-evolutionary framework

Innovation

• Co-evolutionary optimization
• Automated Heuristic Design
• Large Language Models
• Coupled combinatorial optimization
• Operator synergy