Forecast of Passenger Flow of Urban Rail Transit Based on the DNNC Model

The dynamics of the passenger flow for urban rail transit stations are characterized by complexity, nonlinearity, and periodicity. Short-term passenger flow prediction is the basis for rail transit to carry out operational and dispatching management. A single prediction method is hard to fully describe the changing characteristic of the passenger flow, and is not suitable for daily passenger flow prediction of urban rail transit. Aiming to predict the passenger flow for the Beijing rail transit, this paper proposes a Decomposition Neural Network Composition (DNNC) passenger flow prediction model. The simulation results show that the proposed prediction model could effectively improve the accuracy of passenger flow forecasting, and can well describe the changing law of passenger flow.