A Mixed-Curvature based Pre-training Paradigm for Multi-Task Vehicle Routing Solver

Suyu Liu, Zhiguang Cao, Shanshan Feng, Yew-Soon Ong
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:38066-38101, 2025.

Abstract

Solving various types of vehicle routing problems (VRPs) using a unified neural solver has garnered significant attentions in recent years. Despite their effectiveness, existing neural multi-task solvers often fail to account for the geometric structures inherent in different tasks, which may result in suboptimal performance. To address this limitation, we propose a curvature-aware pre-training framework. Specifically, we leverage mixed-curvature spaces during the feature fusion stage, encouraging the model to capture the underlying geometric properties of each instance. Through extensive experiments, we evaluate the proposed pre-training strategy on existing neural multi-task solvers across a variety of testing scenarios. The results demonstrate that the curvature-aware pre-training approach not only enhances the generalization capabilities of existing neural VRP solvers on synthetic datasets but also improves solution quality on real-world benchmarks.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-liu25b, title = {A Mixed-Curvature based Pre-training Paradigm for Multi-Task Vehicle Routing Solver}, author = {Liu, Suyu and Cao, Zhiguang and Feng, Shanshan and Ong, Yew-Soon}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {38066--38101}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/liu25b/liu25b.pdf}, url = {https://proceedings.mlr.press/v267/liu25b.html}, abstract = {Solving various types of vehicle routing problems (VRPs) using a unified neural solver has garnered significant attentions in recent years. Despite their effectiveness, existing neural multi-task solvers often fail to account for the geometric structures inherent in different tasks, which may result in suboptimal performance. To address this limitation, we propose a curvature-aware pre-training framework. Specifically, we leverage mixed-curvature spaces during the feature fusion stage, encouraging the model to capture the underlying geometric properties of each instance. Through extensive experiments, we evaluate the proposed pre-training strategy on existing neural multi-task solvers across a variety of testing scenarios. The results demonstrate that the curvature-aware pre-training approach not only enhances the generalization capabilities of existing neural VRP solvers on synthetic datasets but also improves solution quality on real-world benchmarks.} }
Endnote
%0 Conference Paper %T A Mixed-Curvature based Pre-training Paradigm for Multi-Task Vehicle Routing Solver %A Suyu Liu %A Zhiguang Cao %A Shanshan Feng %A Yew-Soon Ong %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-liu25b %I PMLR %P 38066--38101 %U https://proceedings.mlr.press/v267/liu25b.html %V 267 %X Solving various types of vehicle routing problems (VRPs) using a unified neural solver has garnered significant attentions in recent years. Despite their effectiveness, existing neural multi-task solvers often fail to account for the geometric structures inherent in different tasks, which may result in suboptimal performance. To address this limitation, we propose a curvature-aware pre-training framework. Specifically, we leverage mixed-curvature spaces during the feature fusion stage, encouraging the model to capture the underlying geometric properties of each instance. Through extensive experiments, we evaluate the proposed pre-training strategy on existing neural multi-task solvers across a variety of testing scenarios. The results demonstrate that the curvature-aware pre-training approach not only enhances the generalization capabilities of existing neural VRP solvers on synthetic datasets but also improves solution quality on real-world benchmarks.
APA
Liu, S., Cao, Z., Feng, S. & Ong, Y.. (2025). A Mixed-Curvature based Pre-training Paradigm for Multi-Task Vehicle Routing Solver. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:38066-38101 Available from https://proceedings.mlr.press/v267/liu25b.html.

Related Material