Time-tabling for the future

Important results from the project

The aim of the project has been to show that it is possible to use a combination of dual optimization and Rapid Branching to produce an executable timetable. Although, after many iterations, the model has failed to prove that the finished solution is (near) optimal, the timetable that has been prepared contains all the trains that are in demand. The project is a pilot study whose most important purpose is to establish whether it is meaningful to continue the development work. The answer to this question is affirmative.

Expected long term effects

Instead of manual handling, an approach has been tried which is based on proposals for a new timetable based on formal optimization. In addition to seminars with representatives of the Trafikverket and the operators on the track, work was started on getting information about driving times on all the smallest units of the track, ie. for all blocks and all types of trains. In addition, a development work was carried out to program the optimization algorithm used. The project gained access to the operators´ desired departure times in April, and a completed timetable was available in June 2018.

Approach and implementation

A pilot study testing a computer-based support for scheduling has been tested on the Malmbanan, the Malmberget - Kiruna - Narvik section. The study was carried out at the same time as the Swedish Transport Administration manually produced a timetable (T20). The agency presented a first draft timetable in mid-June and after comments the timetable was set in September 2019. The first results of the pilot study were available in early May. As all requests could not be met, programming was updated, and new versions tested. A ready timetable for all trains was available at the beginning of June.