A Discrete Event Simulation-Based Optimization Approach for Train Timetabling Problem

In this paper, a simulation-based optimization approach is proposed to solve train scheduling problem in single-line and double-track rail networks. In the proposed approach, the simulation model is used to generate feasible train timetable. The simulation model is designed in an object software environment called Enterprise Dynamics (ED). The input data of the problem includes the type of blocks, the order and location of the stations, the number of lines and platform inside the stations, the station facilities (dewatering, refueling, and prayer room), maintenance intervals and praying intervals. The decision variable is the departure time of trains from the origin and the objective of the problem is to minimize the total weighted non-planned stopping times of trains. To find a near-optimal timetable, a simulation-based optimization method based on the evolutionary path relinking algorithm is designed. To evaluate the efficiency of the proposed algorithm, the scheduling of passenger trains in the Tehran-Razi route has been performed. The results of the research show that the implementation of the proposed model has led to significant improvement in non-planned train stopping times as against to the current schedule of the Iranian railways. The proposed algorithm is also able to generate a train timetable in a reasonable time according to the intervals allowed for prayer.