arXiv cs.AI· Ni Zhang, Jianan Zhou, Yue-Jiao Gong, Zhiguang Cao·· 2 天前AI 评分20
自引导 LLM 推理用于车辆路径问题错误诊断
Self-Guided LLM Reasoning for Error Diagnosis in Vehicle Routing Problems
AI 导读
研究系统考察了 LLM 检测和纠正车辆路径问题（VRP）实例错误的能力，发现其在复杂实例或大幅增加推理负担的错误上存在明显局限。为此提出自引导推理框架，让 LLM 自主生成代码提供结构化洞察以辅助错误诊断。在多种问题设置和 LLM 上的实验表明，该框架显著提升错误检测性能，达到接近完美的准确率。
正文 · 原文
Abstract:Vehicle Routing Problems (VRPs) are fundamental to many real-world applications. Recent advances in large language models (LLMs) have spurred progress in solving these NP-hard problems. However, most existing approaches assume that input instances are error-free, neglecting the presence of faults commonly encountered in real-world data. In this work, we systematically investigate the capability of LLMs to detect (and correct) errors in VRP instances, and identify key limitations, particularly in handling complex instances or errors that substantially increase the LLM's reasoning burden. To overcome these challenges, we propose a self-guided reasoning framework in which LLMs autonomously generate code that provides structured insights to aid in error diagnosis. Experiments across diverse problem settings and LLMs demonstrate that our framework significantly improves error detection performance, achieving near-perfect accuracy.
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.21393 [cs.AI] |
| (or arXiv:2610.21393v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.21393 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jianan Zhou [view email]
[v1]
Thu, 15 Oct 2026 07:08:51 UTC (187 KB)
来源：arXiv cs.AI · arxiv.org