arXiv:2506.02935v4 Announce Type: replace 
Abstract: Multi-Task Learning (MTL) in Neural Combinatorial Optimization (NCO) is a promising approach to train a unified model capable of solving multiple Vehicle Routing Problem (VRP) variants. However, existing Reinforcement Learning (RL)-based multi-task methods can only train light decoder models on small-scale problems, exhibiting limited generalization ability when solving large-scale problems. To overcome this limitation, this work introduces a novel multi-task learning method driven by knowledge distillation (MTL-KD), which enables the efficient training of heavy decoder models with strong generalization ability. The proposed MTL-KD method transfers policy knowledge from multiple distinct RL-based single-task models to a single heavy decoder model, facilitating label-free training and effectively improving the model's generalization ability across diverse tasks. In addition, we introduce a flexible inference strategy termed Random Reordering Re-Construction (R3C), which is specifically adapted for diverse VRP tasks and further boosts the performance of the multi-task model. Experimental results on 6 seen and 10 unseen VRP variants with up to 1000 nodes indicate that our proposed method consistently achieves superior performance on both uniform and real-world benchmarks, demonstrating robust generalization abilities.

تقدم الأبحاث الجديدة حول التعلم متعدد المهام لمحللات توجيه المركبات العصبية نهجًا مبتكرًا لمعالجة عدة متغيرات لمشكلة توجيه المركبات. من خلال معالجة قيود الأساليب الحالية، تهدف هذه الدراسة إلى تعزيز قدرات التعميم للنماذج، مما يجعلها أكثر فعالية في التحديات الكبيرة.

La nueva investigación sobre el aprendizaje multi-tarea para los solucionadores de enrutamiento de vehículos neuronales presenta un enfoque innovador para abordar varias variantes del problema de enrutamiento de vehículos. Al abordar las limitaciones de los métodos existentes, este estudio busca mejorar las capacidades de generalización de los modelos, haciéndolos más efectivos para desafíos a gran escala.

La nouvelle recherche sur l'apprentissage multi-tâches pour les solveurs de routage de véhicules neuronaux présente une approche innovante pour s'attaquer à diverses variantes du problème de routage de véhicules. En abordant les limitations des méthodes existantes, cette étude vise à améliorer les capacités de généralisation des modèles, les rendant plus efficaces pour des défis à plus grande échelle.

The new research on Multi-Task Learning for Neural Vehicle Routing Solvers presents an innovative approach to tackle various Vehicle Routing Problem variants. By addressing the limitations of existing methods, this study aims to enhance the generalization capabilities of models, making them more effective for larger-scale challenges.

MTL-KD: Multi-Task Learning Via Knowledge Distillation for Generalizable Neural Vehicle Routing Solver

Was this article worth reading? Share it

Ready to build your own newsroom?