U.S. National Research Council, Transportation Research Board
An intercity passenger rail is built to connect several major cities. To provide satisfactory services to passengers, railway operators plan different stop schedules, such as all-stop, skip-stop, and express services. However, stopping patterns determined by empirical rules or political arguments are generally not optimal. This paper aims at developing a decision support system to generate the optimal combination of stopping patterns for minimizing total passenger in-vehicle time. This problem was first formulated by using mixed integer programming, but this method is intractable when dealing with large-scale problems because of the complexity of model structure and the nature of the problem. A genetic algorithm was then developed to search for the optimal or near-optimal solution efficiently within a reasonable computation time. The proposed algorithm was successfully implemented on Taiwan High Speed Rail. The resulting solution is better than the current practice, and the proposed algorithm is capable of finding the optimal solution in seconds. The present case study demonstrates that the decision support tool can tackle large-scale problems and can help operators efficiently and effectively design an optimal combination of stopping patterns.