Train Time Delay Prediction for High-Speed Train Dispatching Based on Spatio-Temporal Graph Convolutional Network

Train delay prediction can improve the quality of train dispatching, which helps dispatcher to estimate running state more accurately and make reasonable dispatching decision. The one is affected by many factors, such as passenger flow, fault, extreme weather, strategy. departure time generally determined dispatchers, limited their strategy knowledge. existing methods cannot comprehensively consider temporal spatial dependence between multiple trains routes. In this paper, we don’t try predict specific train, but collective cumulative effect over a certain period, represented total number arrival delays in station. We propose deep learning framework, spatio-temporal graph convolutional network (TSTGCN), station for emergency plans. proposed model mainly composed recent, daily weekly components. Each component contains two parts: attention mechanism convolution, effectively capture characteristics. weighted fusion three components produces final result. experiments on operation data from China Railway Passenger Ticket System demonstrate that TSTGCN clearly outperforms advanced baselines prediction.