Reschedule the Shunting and Service Scheduling Plan After Disturbances at a Chinese High-Speed Railway Depot

Jintang Shi, Haodong Li, Daoqi Wang, Pieter VansteenwegenIEEE Access 2024 · OpenReview

Abstract: Electrical Multiple Units (EMUs) need to undergo services like repairing, cleaning and wheel turning according to a train shunting and service scheduling (SSS) plan at the high-speed railway depot. Disturbances in the timetable at the station and overtime services affect the original SSS plan. Therefore, frequent modifications to the original SSS plan and rescheduling of some EMUs are required to recover the feasibility of the plan. This paper considers a rescheduling method for integrating train shunting, service scheduling and routing at EMU depots with various layouts. In this paper, the movement process of EMUs at depots is represented with the help of a two-layer time-space network. A heuristic algorithm for rescheduling is designed. Case studies based on a real-world depot show that the heuristic can repair the original SSS plan in 180 seconds while ensuring that the new SSS plan is as close as possible to the original SSS plan.

The Challenge of Depot Management

High-speed rail networks rely on Electrical Multiple Units (EMUs) that require rigorous nightly maintenance. This "shunting and service" process includes cleaning, sewage suction, wheel turning, and various repair levels. These operations take place in specialized depots where space is limited and tracks are often shared, creating complex logistical constraints.

Layout of a typical EMU depot showing various functional yards including storage, cleaning, and repair areas.

Disturbances, such as late arrivals from stations or maintenance overruns, can cause a ripple effect. Because many depot tracks operate on a Last-In-First-Out (LIFO) basis, a single delayed train can block several others. Currently, dispatchers often fix these plans manually, a slow and stressful process that can lead to further inefficiencies. Rapid rescheduling is essential to recover the feasibility of the plan and maintain operational flow.

Methodology: Two-Layer Time-Space Network

The researchers introduced a multi-faceted approach to automate rescheduling. The core innovation is a two-layer time-space network that separates depot operations into a "shunting layer" for movement and a "dwelling layer" for services and parking. This allows the system to track train locations and activities with second-by-second precision.

Physical network representation for short and long EMUs, illustrating the nodal structure used to model depot movements.

Key components of the method include:

  • Saw-Move Modeling: The model explicitly accounts for time-consuming direction changes required to navigate the depot's layout.
  • Heuristic Local Search: Instead of recalculating the entire schedule from scratch, the algorithm identifies a "pool" of affected EMUs and optimizes their sequences locally to save time.
  • Stability Focus: A cost-adjustment parameter is used to "attract" the algorithm toward using the tracks assigned in the original plan, minimizing disruption for crew and staff.

Key Results and Insights

The study validated the method using real-world data from the Jinan Depot in China, a major hub with 24 storage tracks and multiple specialized service tracks.

  • Speed: The heuristic found feasible solutions in 48 to 157 seconds across various test scenarios, comfortably meeting the 180-second real-time requirement for railway operations.
  • Efficiency: The system successfully resolved conflicts in high-pressure scenarios involving up to 30 trains, effectively managing LIFO bottlenecks by proactively moving "trapped" units to storage tracks.
  • Minimal Disruption: In test cases, the algorithm often fixed schedules with minimal changes to track assignments, primarily adjusting service times to absorb delays rather than reassigning physical resources.

Conclusion

By automating the rescheduling process, this research provides a tool that can handle the dynamic nature of high-speed rail operations. Future work aims to integrate crew scheduling directly into the model and enhance the proactive robustness of original schedules to mitigate the impact of disturbances before they occur.

Summary updated

Abstract

Electrical Multiple Units (EMUs) need to undergo services like repairing, cleaning and wheel turning according to a train shunting and service scheduling (SSS) plan at the high-speed railway depot. Disturbances in the timetable at the station and overtime services affect the original SSS plan. Therefore, frequent modifications to the original SSS plan and rescheduling of some EMUs are required to recover the feasibility of the plan. This paper considers a rescheduling method for integrating train shunting, service scheduling and routing at EMU depots with various layouts. In this paper, the movement process of EMUs at depots is represented with the help of a two-layer time-space network. A heuristic algorithm for rescheduling is designed. Case studies based on a real-world depot show that the heuristic can repair the original SSS plan in 180 seconds while ensuring that the new SSS plan is as close as possible to the original SSS plan.
BibTeX
@inproceedings{shi2024reschedule,
  author = {Jintang Shi and Haodong Li and Daoqi Wang and Pieter Vansteenwegen},
  title = {{Reschedule the Shunting and Service Scheduling Plan After Disturbances at a Chinese High-Speed Railway Depot}},
  booktitle = {IEEE Access},
  year = {2024},
  url = {https://openreview.net/forum?id=DRv9BvYHxg},
}

Loading related content...

Figure