An Exact Integer Linear Programming Formulation for the Passenger Oriented Timetabling Problem

Abstract: We present a new mathematical formulation for tactical railway timetabling that aims at minimizing total passenger perceived travel time. This new formulation for the POT problem of Polinder et al. (2022), uses as input a railway network, an existing line plan, and a demand matrix, and outputs a timetable. Contrary to general tactical timetabling models, we relax the assumption that line frequency is given as input in the mathematical model. Instead, we consider a maximum frequency, such that some lines' frequencies can be decreased. We come up with solution methods to solve the problem formulation and expect experimental results to improve on timetables created using current state-of-the-art methods that take line frequency as input. We will test the instances using the most utilised parts of the Dutch railway network as input and compare methods for the perceived travel time (including waiting time). In particular, we expect to see better results in specific cases where the trade-off between a lower average travel time and fewer trains running is possible.