Use of Hybrid Methods to Enhance Real-Time Railway Traffic Control (Dispatching)
In the context of the digitalization of railway operations, hybrid methods become key assets for improving existing traffic control and dispatching (i.e. executing the schedule). Dispatching is characterized by continuous traffic monitoring and systematic adjustment of the schedule to the occurrence of stochastic events (NP-Complete problem). Due to the complexity of this process, decision-support tools have been developed to support dispatchers in real-time. To date, available tools are applicable to simple sub-tasks within the schedule adjustment process and with limited solution quality. This proposal aims to investigate different combinations of hybrid methods (exact, heuristic, machine learning) and develop practical approaches that can directly improve the schedule adjustment quality and thus, enhance the railway system’s capacity. To reach the proposed objective, two tasks are foreseen within the scope of this proposal: 1) conducting exploratory and experimental research and 2) developing a joint proposal for acquiring external funding.
Consortium: TU Darmstadt, Germany, KTH, Sweden, and UPC, Spain