Bi-level RL-Heuristic Optimization for Real-world Winter Road Maintenance
Yue Xie, Zizhen Xu, William Beazley, Fumiya Iida · 2026
Read Bi-level RL-Heuristic Optimization for Real-world Winter Road Maintenance by Yue Xie, Zizhen Xu, William Beazley, Fumiya Iida (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-24097.txt.
Description
Yue Xie and three co-authors present a bi-level reinforcement learning and heuristic optimization framework for real-world winter road maintenance. The 2026 paper addresses large-scale gritting and snow routes on UK strategic roads, where existing methods struggle with scale and lean heavily on human planning.
At the upper level, a reinforcement learning agent partitions the road network into manageable clusters, allocating resources from multiple depots. Within each cluster, the lower level solves a multi-objective vehicle routing problem. It minimizes the longest vehicle travel time and total carbon emissions. It respects vehicle constraints, depot capacities, and segment requirements. Results on real operational data show balanced workloads, longest travel times below a two-hour target, lower emissions, and substantial cost savings. Those results demonstrate how AI-driven bi-level optimization can improve operational decisions for highway operators.
Catalog record
| CREATOR | Yue Xie, Zizhen Xu, William Beazley, Fumiya Iida |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 29 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-24097