Train rescheduling algorithm which minimizes passengers' dissatisfaction

Abstract

Although computer systems which assist human experts in rescheduling disrupted train traffic is being practically used recently, they are not so helpful in decreasing the workload of human experts. This is because they are lacking in intelligence such as to automatically make rescheduling plans. In this paper, we propose an algorithm for automatic train rescheduling. Firstly, we propose to use passengers' dissatisfaction as a. criterion of rescheduling plans and to regard the train rescheduling problem as a constraint optimization problem in which dissatisfaction of passengers should be minimized. Then we introduce an algorithm for train rescheduling designed as a combination of PERT and meta-heuristics. We also show some experimental results of the algorithm using actual train schedule data. © Springer-Verlag Berlin Heidelberg 2005.