LaT：用于多任务车辆路径规划求解器的大语言模型训练器
LaT: LLM-as-Trainer for Multi-Task Vehicle Routing Solvers
浏览论文内容
中文总结 AI 辅助
研究针对多任务车辆路径规划求解器中VRP变体优化难度不同及现有方法不足的问题，提出LaT训练范式，用预训练大语言模型作外部训练器，定期分析指标生成指导向量注入编码器层，实验证明该范式能提升求解器质量。
中文摘要 AI 辅助
多任务神经求解器旨在在统一模型中处理多个车辆路径规划问题（VRP）变体，避免对每个约束组合进行单独训练。然而，VRP变体在优化难度上存在差异，现有方法缺乏对训练状态的阶段性反馈，导致模型偏向某些特定变体。虽然元学习可以支持自适应训练，但通常需要双层优化和额外的梯度更新，增加了计算成本。为解决这一限制，我们提出了LLM-as-Trainer（LaT），一种即插即用的训练范式，它使用预训练的大语言模型作为外部训练器。LaT定期分析跨任务验证指标以生成阶段性指导向量。该向量与当前任务的约束向量相结合，并注入到每个编码器层，在后续策略优化期间为神经求解器提供额外的训练信息。在16个VRP变体上的实验表明，LaT提高了几种先进多任务神经求解器在已训练和未见过的变体上的求解质量，支持了所提出训练范式的有效性和通用性。
英文摘要
Multi-task neural solvers aim to handle multiple Vehicle Routing Problem (VRP) variants within a unified model, avoiding separate training for each constraint combination. However, VRP variants differ in optimization difficulty, while existing methods lack stage-wise feedback on their training status, making the model biased to some specific variants. Although meta-learning can support adaptive training, it typically requires bi-level optimization and additional gradient updates, increasing computational cost. To address this limitation, we propose LLM-as-Trainer (LaT), a plug-and-play training paradigm that uses a pretrained large language model as an external trainer. LaT periodically analyzes cross-task validation metrics to generate a stage-wise guidance vector. This vector is combined with the current task's constraint vector and injected into each encoder layer, providing the neural solver with additional training information during subsequent policy optimization. Experiments on 16 VRP variants show that LaT improves the solution quality of several state-of-the-art multi-task neural solvers on both trained and unseen variants, supporting the effectiveness and generality of the proposed training paradigm.
发表机构
- School of Future Technology, South China University of Technology（华南理工大学未来技术学院）
- School of Computer Science and Engineering, South China University of Technology（华南理工大学计算机科学与工程学院）
- School of Engineering and Computer Science, Victoria University of Wellington（惠灵顿维多利亚大学工程与计算机科学学院）
- Institute of Marine Science and Technology, Shandong University（山东大学海洋科学与技术研究院）
- School of Computing and Information Systems, Singapore Management University（新加坡管理大学计算与信息系统学院）
机构由 AI 辅助整理，请以论文原文为准。
