Modeling Uncertainty in the Timetable-Based Railway Network Design Problem

Many European countries plan their railway infrastructure according to strategic timetables, using them as input for further strategic and tactical planning steps, including network design. While both tactical timetabling and network design are well covered by the research, there exists, to the best of the authors’ knowledge, no model which focuses on network design based on strategic timetables. In this short paper, we present an overview of our research. The traditional network design problem is extended to incorporate railway-specific features such as headway-based capacity estimations and demand derived from a strategic timetable. This includes trains represented by integral flows with start, destination, and time bounds as well as timetabling constraints for line frequencies and transfers. Since strategic timetabling is done many years in advance, the strategic timetable and as such the demand for the network design problem are subject to uncertainty. To account for this, we aim to calculate networks which are robust towards changing input timetables. We describe two approaches to model this: optimizing the network for a timetable family (a set of discrete scenarios) and varying demands within scenarios, which is modelled by a set of optional trains. The paper describes basic modelling decisions, details the approach to incorporate the uncertainty and shows the main features of the optimization model as well as a case study and further research.