Micro-macro transformation of railway networks

2210-9706/$ see front matter 2011 Elsevier Ltd. A doi:10.1016/j.jrtpm.2011.09.001 ⇑ Corresponding author. Tel.: +49 30 84185244; fax E-mail addresses: [email protected] (T. Sch (R. Borndörfer), [email protected] (B. Erol), thomas.graffagn [email protected], [email protected] (E. Swarat). This paper presents a bottom-up approach to automatic railway network simplification. Starting from a detailed microscopic level as it is used in railway simulation, the network is transformed by an algorithm to an aggregated level, i.e., to a macroscopic network, that is sufficient for long-term planning and optimization. Running and headway times are rounded to a user defined precision by a special cumulative method. After this ‘‘macrotization’’ trains from a given set of requests are added to the existing timetable by solving an optimal train path allocation problem. The objective of this problem is to maximize a sum of utilities of the allocated trains; the utility can be a constant, some monetary value, etc. The optimized schedule is re-transformed back to the microscopic level in such a way that it can be simulated without any conflicts between the train paths. We apply this algorithm to macrotize a microscopic network model of the highly frequented Simplon corridor in the Alps between Switzerland and Italy. To the best knowledge of the authors and confirmed by several railway practitioners this was the first time that track allocations that have been produced in a fully automatic way on a macroscopic scale fulfill the requirements of the originating microscopic model and withstand an evaluation in the microscopic simulation tool OpenTrack. Our micro–macro transformation method allows for a much faster planning and provides solutions of a quality that are at least comparable to the most sophisticated manual schedules. In this way meaningful scenario analyses can be carried out that pave the way towards a new level of decision support in railway planning. 2011 Elsevier Ltd. All rights reserved.