Development and evaluation of a synchronous heuristic approach for the automatic assignment of trains to standardised train paths: Demand oriented rescheduling of train paths and their allocation in railway networks (ATRANS)

• Applicant: Professor Dr.-Ing. Andreas Oetting
• Subject Area: Traffic and Transport Systems, Intelligent and Automated Traffic
• Term: from 2015 to 2020
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 257931669

Introduction: Train operating companies apply for and pay for the right to use railway infrastructure for each train. This right is called train path. The annual German schedule 13 contains about 60.000 train paths. Up to now, for each train a specific train path is designed. New European and national legal guidelines as well as the increasing level of utilisation of the rail network require an entirely new approach for the allocation of train paths: the train paths for freight traffic will have to be systematised and standardised. Therefore, traffic supply (train paths without a train assigned) and demand (train ride) are approached separately. First, standardised train paths are designed. Later, demand is assigned to these train paths. Algorithms for the design of train paths already exist. However, there is only one optimisation approach for assignment of demand to the train paths. Unfortunately, this approach systematically reduces capacity of railway infrastructure.Objective: Currently experts assign the demand to the standardized pre-defined train paths successfully but manually with a high workload. Therefore, subproject ATRANS 2.2 aims to develop and evaluate a heuristic automatic approach for the assignment of demand. This approach should enable to easily generate an optimized timetable with an efficient use of the railway infrastructure. During the assignment of demand, conflicting train registrations for the same standardised train path occur. Each resolution of such a conflict must be compliant and therefore non-discriminatory and traceable. For this reason an optimization approach solving all conflicts at a time, is unfeasible. The selected approach needs to terminate in order to generate the required conflict-free timetable. Due to the size of the problem the new algorithms have to feature a very low computation time.Method: In order to fulfil the requirements compliant with legal guidelines and non-discriminatory only the more complex synchronous algorithms apply. They assign all train rides at a time (synchronous) stepwise in time and space. A major problem of synchronous algorithms is the missing projection. One key-research issue in this subproject is therefore the development of algorithms for a projection with low computation time. At least, three complementary ways are investigated:-A macroscopic modelling based on the systemized train paths using minimum headway times allows a projection up to 10-20 km with a low data volume.-The extensive assessment of alternative conflict solutions includes resulting conflicts with further trains in the future.-The number of trains considered in the conflict resolution depends on the conflict investigated. This number depends on pre-defined parameters to adjust computation time.Use: The resulting algorithms are necessary to fulfil the legal guidelines, they generate more capacity for the increasing demand and they lead to higher efficiency of the trans-European railway network.

• DFG Programme: Research Grants