A Railway Scheduling Method using Probabilistic Model Checking

Abstract

Trains scheduling is an important problem in railway transportation. Many companies use fixed train timetabling to handle this problem. Train delays can affect the pre-defined timetables and postpone destination arrival times. Besides, delay propagation may affect other trains and degrade the performance of a railway network. An optimal timetable minimizes the total propagated delays in a network. In this paper, we propose a new approach to compute the expected propagated delays in a railway network. As the main contribution of the work, we use Discrete-time Markov chains to model a railway network with a fixed timetable and use probabilistic model checking to approximate the expected delays and the probability of reaching destinations with a desired delay. We use PRISM model checker to apply our approach for analyzing the impact of different train scheduling in double line tracks.