ICAM: Rethinking Instance-Conditioned Adaptation in Neural Vehicle Routing Solver
Keywords: Vehicle Routing Problem, Reinforcement Learning, Instance-Conditioned Adaptation, Neural Combinatorial Optimization, Large-scale Generalization
TL;DR: ICAM efficiently leverages instance-conditioned information to generate promising solutions for cross-scale instances with a very fast inference time and significantly improves the generalization performance of neural vehicle routing solver.
Abstract: The neural combinatorial optimization (NCO) has shown great potential for solving routing problems without requiring expert knowledge. However, existing constructive NCO methods still struggle to directly solve large-scale instances, which significantly limits their application prospects. To address these crucial shortcomings, this work proposes a novel Instance-Conditioned Adaptation Model (ICAM) for better large-scale generalization of neural routing solvers. In particular, we design a simple yet efficient instance-conditioned adaptation function to significantly improve the generalization performance of existing NCO models with a very small time and memory overhead. In addition, with a systematic investigation on the performance of information incorporation between different attention mechanisms, we further propose a powerful yet low-complexity instance-conditioned adaptation module to generate better solutions for instances across different scales. Experimental results show that our proposed method is capable of obtaining promising results with a very fast inference time in solving Traveling Salesman Problems (TSPs), Capacitated Vehicle Routing Problems (CVRPs) and Asymmetric Traveling Salesman Problems (ATSPs). To the best of our knowledge, our model achieves state-of-the-art performance among all RL-based constructive methods for TSPs and ATSPs with up to 1,000 nodes and extends state-of-the-art performance to 5,000 nodes on CVRP instances, and our method also generalizes well to solve cross-distribution instances.
Primary Area: applications to robotics, autonomy, planning
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 14049
Loading