Algorithms and models for railway optimization

Abstract

Mobility is increasing in a way that calls for systematic traffic planning in a broad context. In Europe the railways are requested to play a central role in this development. Future developments and improvements of European railways will have an impact on people's lives and therefore on society in general. The problems arising in this context are large and highly complex. Here are many interesting and challenging algorithmic problems waiting to be studied. Research topics include the network design, line planning, time table generation, crew scheduling, rolling stock rostering, shunting, time table information and delay management. In this talk we present models and algorithmic methods for several of these problems. We will discuss the interplay between algorithmic aspects and practical issues like availability and quality of data. The focus will be on two topics from network design and time table information respectively where we have ongoing cooperation with railway companies. As an example from network design, we will consider a scenario, where the effects of introducing new train stops in the existing railway network is studied. For time table information whose algorithmic core problem is the computation of shortest paths we discuss new algorithmic issues arising from the huge size of the underlying data. © Springer-Verlag Berlin Heidelberg 2003.