An Approach for the Railway Multiterritory Dispatching Problem

Abstract

In this research, we address the railway multiterritory dispatch planning (RMTDP) problem. The goal of the RMTDP problem is to find the optimal movement of trains across consecutive dispatch territories, and it is one of the major challenges that decision makers face on a daily basis. It ideally takes into account the correct placement of maintenance windows, remaining capacity of terminals, and availability of train crews, among other critical aspects such as locomotive balance, fueling locations, and inspections. Although these train movement plans are made at the corridor level, which comprises several dispatch territories, when it comes to execution, the meet–pass decisions are made at the individual dispatch territories. This notion causes disruptions and misalignment at the boundaries of dispatch territories. The approach in this paper aims at finding a holistic conflict-free master plan by optimally matching train lineups at territory boundaries and smoothly routing trains through bottlenecks. We propose an efficient solution approach that iteratively constructs a master scheduling plan while minimizing the amount of train delay within a given planning horizon. This is accomplished by designing a time–space network model to identify feasible schedules and developing a mathematical programming–based heuristic to solve the underlying model. A thorough computational study shows the effectiveness of our heuristic approach, as we report reasonable average run times of 3.0 and 6.5 minutes to solve instances of moderate to large size problems, respectively. The results obtained from the algorithm using test snapshots from a Class I railroad company have been shown to assistant chief dispatchers and have received encouraging feedback for applicability.