Evaluating and improving the passenger punctuality of a timetable

Nederlandse Spoorwegen (NS) measures the reliability of their passenger train operations with passenger punctuality. Passenger punctuality is defined as the percentage of passengers whose journey was successful in terms of travel time and is measured after the execution of the timetable. Improving the reliability is a crucial point of focus for NS, and therefore, this research focuses on improving timetables using passenger punctuality as measure. Firstly, we develop an evaluation model to give a measure of the passenger punctuality for any timetable before its execution. The data, obtained through the evaluation model, is also used for further analysis of the timetable on, for example, missed transfers. Secondly, we use the existing Stochastic Optimization Model (SOM) to improve timetables. Originally, the objective of SOM is to minimize the train delay. In this research, we adapt SOM to make customized objective functions possible. By using an objective function based on the passengers in a case study, we obtain timetables which show significant improvements with respect to the passenger punctuality. In the case study, we increased the 5-minute passenger punctuality from 57.31% to 68.90%. Even though the current model is not applicable on the whole railway network, this thesis provides a proof-of-concept for the used approach.