arXiv cs.AI· Peiheng Jian, Zhenwei Wang, Tiehua Zhang, Lei Zhang, Zhishu Shen, Ruibin Bai, Zhijun Ding·· 12 小时前AI 评分18
TRCM：面向鲁棒神经车辆路径规划的自适应拓扑路线一致性建模
TRCM: Topology-Aware Route-Consistency Modeling for Robust Neural Vehicle Routing
AI 导读
TRCM 通过复用自回归构造流程中的节点表示预测两两同路线概率，并将其作为路线一致性信号注入客户选择。在匹配解码预算下，CVRP50 最优性差距从 3.67% 降至 3.28%，CVRP100 从 2.11% 降至 1.99%；8 倍增强推理下 CVRP100 差距从 1.61% 降至 1.49%。在 XML100 上，TRCM*8Aug 将平均目标改进 0.24%。
正文 · 原文
Abstract:Neural combinatorial optimization has enabled fast constructive solvers for the Capacitated Vehicle Routing Problem (CVRP), yet most autoregressive models still construct solutions through sequential node selection and represent route organization only implicitly. We investigate whether route-level customer co-assignment can be modeled explicitly and used to guide this node-level decision process. To this end, we propose TopologyAware RouteConsistency Modeling (TRCM), which reuses node representations from a standard autoregressive construction pipeline to predict pairwise same-route probabilities and injects them into customer selection as a route-consistency signal. Because sameroute supervision is both class-imbalanced and structurally uneven, TRCM combines structure-preserving augmentation with Topological Entropy (TE)- based reweighting to emphasize locally ambiguous customer pairs. Under matched decoding budgets, TRCM reduces the optimality gap from 3.67% to 3.28% on CVRP50 and from 2.11% to 1.99% on CVRP100; under 8-fold augmented inference, it reduces the CVRP100 gap from 1.61% to 1.49%. On XML100, TRCM*8Aug improves the average objective by 0.24%, with larger gains on several structurally informative factors. A controlled demand-reassignment test over 500 variants further shows more concentrated objective fluctuations. These results support explicit route-consistency modeling as an effective structural principle for autoregressive neural routing.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.27568 [cs.LG] |
| (or arXiv:2610.27568v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.27568 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Zhenwei Wang [view email]
[v1]
Tue, 20 Oct 2026 03:32:20 UTC (1,313 KB)
来源：arXiv cs.AI · arxiv.org