Railway timetabling with integrated passenger distribution

Timetabling for railway services often aims at optimizing travel times passengers. At the same time, restricting assumptions on passenger behavior and modeling are made. While research has shown that discrete choice models suitable to estimate distribution of passengers routes, this not been considered in timetabling yet. We investigate how integrate a into an optimization framework present two mixed integer linear programs problem. Both approaches design timetables simultaneously find corresponding available routes. One model uses route choices. The other integrated simulation approximate according logit model, commonly used model. compare both new with three state-of-the-art methods heuristic approach set artificial instances partial network Netherlands Railways (NS). Our experiments provide insights impact considering multiple routes instead single route, versus predetermined assignment respect solution quality.