Search

With our Periodic Event Scheduling Problem (PESP) based timetabling method we are able to produce a passenger robust timetable for all 88 hourly passenger trains running on tracks managed by the Danish Infrastructure Manager Banedanmark. The objective function of our model is the total expected passenger journey time in practice and is minimised. The result of this is that the produced timetable reduces the expected journey time of all corresponding train passengers together by 2.9% compared to the original timetable dened by Banedanmark. Our simulations show that the average probability of missing a transfer is also reduced from 11.34% to 2.45%. The computation of this timetable takes only 65 minutes. The major innovations of our approach are the addition of a complete objective function to the PESP model and the addition of a particular cycle constraint set that reduces computation times. In this paper, we demonstrate that these combined innovations result in a method that quickly generates cyclic timetables for a train network spanning an entire country and that these timetables also reduce the expected passenger travel time in practice. no isbn ispartof: Proceedings of the Conference on Advanced Systems in Public Transport ispartof: Conference on Advanced Systems in Public Transport location:Rotterdam date:19 Jul - 23 Jul 2015 status: published