Back Propagation Neural Network (BPNN) in Passenger Demand Forecast for Moscow-Kazan HSR

摘要

Forecasting passenger demand accurately is a vital issue for management and operation of high-speed railway. This paper propose to use back propagation neural network (BPNN) in order to predict future passenger demand on Moscow-Kazan HSR by considering socio-economic factors, such as: gross regional product (GRP), population, real income, and passenger demand for the last 20 years in 7 studied regions. This approach includes 3 stages: 1) data collection of the influenced factors; 2) travel modes division (comparison analysis of all travel modes including future HSR in order to define possible percentage of passengers); 3) BPNN method application. The paper presents a forecast of passenger demand until 2027. From the travel modes division was found close relationship between air and HSR modes. The paper contributes to the empirical literature on HSR passenger demand forecast. Results indicate that BPNN method is a reliable method, which is able to predict the demand of future HSR.