Disposition des Eisenbahnbetriebs unter Einbeziehung von zufallsbedingten Unsicherheiten im künftigen Betriebsablauf
Zusammenfassung der Projektergebnisse
The practical significance of this project stemmed from the limited expansion and increasing travel demand of railway network. In order to handle the various interferences during the actual railway operation and automatically generate robust dispatching solutions as an assistance for nowadays manual dispatching, an operational risk-based dynamic dispatching approach was developed in this project. Within the algorithm framework, the future potential random disturbances were taken into consideration in the dispatching process. Operational risk analysis of railway block sections is conducted as the foundation of the dispatching algorithm. An innovative approach has been developed to evaluate the operational risk of block sections based on artificially generated disturbances and railway simulation tools. The proposed dispatching algorithm is a hybrid dispatching model which combines both heuristic and simulation approaches and is carried out within a rolling time horizon framework. One of the typical heuristic method, Tabu search algorithm, is employed to find a near-optimal dispatching solution in each round of dispatching, and Railsys® is employed to simulate the railway operation based on the generated dispatching solutions. With continuously updates and adjustments of the railway system, the potential conflicts could be recognized in time and the generated dispatching solutions are able to handle future stochastic disturbances. Three metrics, NRR, TotalwWT, and Capacity were utilized to assess the performance of the proposed algorithm and to determine the most appropriate values of related parameters. The comparisons with FCFS principle exhibited the superior performance of the prosed algorithm. Moreover, the results of the capacity research indicated that the proposed algorithm achieved the best balance between capacity and operation quality than FCFS and conventional dispatching methods. Several advantages were observed for the proposed algorithm. The simulation-based approach was free from the limit of real data, and thus multiple groups of simulation were able to be performed for calculation of the expected value of the operational risk indicator, for provision of the various surrogate of the real-world railway operation, and for implementation of the sensitivity analysis of different related parameters. In addition, the proposed algorithm framework was not restricted to the selected probability distribution functions of stochastic disturbances. As long as the related data and information were provided, this dynamic dispatching system could be easily employed on any railway network according to different research objective system of specific studies. Furthermore, the implementation of this dispatching algorithm liberated dispatchers from burdensome priority calculation tasks. The computation time of this algorithm depends on the scale of the investigation area, the number of disturbances scenarios, the execution time of running a single simulation in the simulation tool, the upper limit of searching loops during TS process as its termination condition, etc.
Projektbezogene Publikationen (Auswahl)
-
Disposition von verkehrlichen Prozessen unter Einbeziehung von zufallsbedingten Unsicherheiten. Eisenbahntechnische Rundschau, 67(10), pp. 22–25, 2018
Tideman, M., Martin, U. & Zhao, W.
-
Proaktive Disposition luftverkehrlicher Prozesse. Internationales Verkehrswesen, 70(4), pp. 60–63, 2018
Tideman, M. & Martin, U.
-
Proactive Dispatching of Railway Operation. Proceedings of RailNorrköping 2019 – 8th International Conference on Railway Operations Modelling and Analysis (ICROMA), Norrköping, Sweden, pp. 1069–1078, 2019
Tideman, M., Martin, U. & Zhao, W.
-
Generation of Dispatching Solutions in a Rolling Time Horizon Framework with the aid of Tabu Search Algorithm. Proceedings of the 17th International Conference on Railway Engineering Design & Operation (COMPRAIL 2020)
Martin, U., Tideman, M. & Zhao, W.
