Multi-objective periodic railway timetabling on dense heterogeneous railway corridors

• Published in: Transportation research. Part B: methodological Vol. 125; pp. 52 - 75
• Main Authors: Yan, Fei, Bešinović, Nikola, Goverde, Rob M.P.
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01.07.2019; Elsevier Science Ltd
• Subjects: Algorithms; Computer applications; Euclidean geometry; Flexible overtaking; Linear programming; Mixed integer; Multi-objective optimization; Multiple objective analysis; Periodic timetable; Railroads; Timetable robustness; Timetables; Transportation corridors; ε-constraint; ε-constraint; Multi-objective optimization; Timetable robustness; Flexible overtaking; Periodic timetable
• ISSN: 0191-2615; 1879-2367
• DOI: 10.1016/j.trb.2019.05.002

•We propose a multi-objective periodic timetable optimization problem.•Multiple overtakings are allowed to improve infrastructure occupation and robustness.•The model optimizes a trade-off between journey times, regularity, vulnerability, and the number of overtakings.•We apply the epsilon-constraint method and explore the resulting 4-dimensional Pareto frontier.•The method is demonstrated on a real-world Dutch railway corridor. This paper proposes a new multi-objective periodic railway timetabling (MOPRT) problem with four objectives to be minimized: train journey time, timetable regularity deviation, timetable vulnerability and the number of overtakings. The aim is to find an efficient, regular and robust timetable that utilizes the infrastructure capacity as good as possible. Based on the Periodic Event Scheduling Problem, we formulate the MOPRT problem as a Mixed Integer Linear Program (MILP). The ε-constraint method is applied to deal with the multi-objective property, and algorithms are designed to efficiently create the Pareto frontier. By solving the problem for different values of ε, the four-dimensional Pareto frontier is explored to uncover the trade-offs among the four objectives. The optimal solution is obtained from the Pareto-optimal set by using standardized Euclidean distance, while capacity utilization is used as an additional indicator to chose between close solutions. Computational experiments are performed on a theoretical instance and a real instance in one direction of a Dutch railway corridor, demonstrating the efficiency of the model and approach.