Big data in the stochastic model of the passengers flow at the megalopolis transport system stops

Abstract

The problem of the passengers flow model development is proposed as the subsystem of the general municipal passengers transport system operation model of megalopolis. The specific features of the application subject (Volgograd city, Russia) were detected to simplify the big data simulation problem. The difficulties caused by the high dimensionality were overcome by means of the double time scaling in passengers’ flow estimation. The hour time scale was accepted to the computation of the hourly flow from each departure stop to the city district of destination without the pointing of the specific destination stop. The minute time scale was accepted to distribute the hourly flow between the destinations stops located in this district. The algorithms of the destination stops choice simulation were carried out. The follows examples of simulation results are presented: hourly passengers flow directed to the departure stops; daily variations of districts population caused by the inter-district passengers’ flows; influence of the of competition on the municipal transport system operation; destination stops choice variants according to the stops’ attractiveness scores designed by experts.