New models and applications for railway timetabling Defended on Friday, 18 December 2020
A timetable is a crucial element for the daily operations of a railway operator. At the same time, designing such a timetable is an extremely complex puzzle, and years of investigations are necessary to design a timetable from scratch. Amongst several other aspects, planners should take the travel demand, connections between trains, capacity on the tracks and in the train, and daily disturbances into account when designing a timetable. Next to this, there are often too many restrictions that a timetable has to satisfy, such that no longer a timetable can exist satisfying all these restrictions.
In this thesis, methods are developed that can support the design of a timetable such that the timetable is as good as possible. For example, we can compute an ideal timetable, matching with travel demand as good as possible. Using this ideal timetable, one can make clear whether regular departure patterns are useful or not, and how this is related with the expected travel time of passengers. This can also be used to determine infrastructure-bottlenecks. Other methods in this thesis aim at relaxing a set of restrictions, if there are too many restrictions, or to make a timetable that is robust against minor disturbances.
Keywords
Public transportation, optimization, timetabling, large-scale optimization, passengers, operations research