Bridging Synthetic and Real Routing Problems via LLM-Guided Instance Generation and Progressive Adaptation

TL;DR

This paper introduces EvoReal, a method that uses LLM-guided evolutionary instance generation and progressive adaptation to improve neural solvers' ability to generalize from synthetic to real-world routing problems, significantly reducing performance gaps.

Contribution

EvoReal is a novel approach combining LLM-guided evolutionary synthesis and progressive fine-tuning to enhance neural combinatorial optimization for real-world routing problems.

Findings

Significantly reduces performance gap on TSPLib and CVRPLib benchmarks.

Generates structurally realistic synthetic instances that mimic real-world data.

Improves generalization of neural solvers across diverse problem scales.

Abstract

Recent advances in Neural Combinatorial Optimization (NCO) methods have significantly improved the capability of neural solvers to handle synthetic routing instances. Nonetheless, existing neural solvers typically struggle to generalize effectively from synthetic, uniformly-distributed training data to real-world VRP scenarios, including widely recognized benchmark instances from TSPLib and CVRPLib. To bridge this generalization gap, we present Evolutionary Realistic Instance Synthesis (EvoReal), which leverages an evolutionary module guided by large language models (LLMs) to generate synthetic instances characterized by diverse and realistic structural patterns. Specifically, the evolutionary module produces synthetic instances whose structural attributes statistically mimics those observed in authentic real-world instances. Subsequently, pre-trained NCO models are progressively refined, firstly aligning them with these structurally enriched synthetic distributions and then further adapting them through direct fine-tuning on actual benchmark instances. Extensive experimental evaluations demonstrate that EvoReal markedly improves the generalization capabilities of state-of-the-art neural solvers, yielding a notable reduced performance gap compared to the optimal solutions on the TSPLib (1.05%) and CVRPLib (2.71%) benchmarks across a broad spectrum of problem scales.