Constructing a good conflict-free timetable is hard. To further complicate matters, the quality of a timetable depends on many different aspects, e.g. transport goal fulfillment, robustness and capacity usage. In TTK we will analyse different measurements for timetable quality, and develop a framework for multi-objective optimization.
Aim and goal
The purpose of the project is to improve railway capacity allocation. A framework for multi-objective optimization could boost the Swedish Infrastructure manager’s work with timetable quality, and subsequently lead to the better timetables. The framework could also be used to get a common picture of the capacity allocation situation, and is a step toward automatic allocation of train paths.
Timetable planning is an inherently hard problem. The deregulation of the railway market and increased demand for railway transport put further pressure on the timetable process. Currently the planners at the Swedish infrastructure manager construct timetables by hand, without any automatic timetable generation support. However, advances in computer science and optimization open up for the possibility of using optimization techniques to support timetable planners. This project focuses on the different objectives that are relevant when constructing a timetable, and how these different objectives interact and can be managed in a multi-objective optimization framework.
The project makes a thorough review of timetable quality measurements, both in the research literature and in practice. To select a sensible set of measurements to include in the framework, the dependencies and effects of the measurements are analysed. Then the framework for multi-objective optimization is designed and tested on a number of use-cases.
The effect is a better timetable process and timetable, and in the long run better railway transportation for both passengers and freight.