Detection of the patterns in the daily route choices of the urban social transport system clients based on the decoupling of passengers’ preferences between the levels of uncertainty

Abstract

The ideas of data mining techniques were applied for the problem of municipal passengers transport system simulation and its results interpreting and generalization. The purpose of the presented work is to propose and justify the passengers flow model suitable for the detection of hidden patterns in the processes of flow forming with the application of the available sources for model identification. The patterns of the daily route choices detection are based on the decoupling of the general model between the sub-models according to the different levels of uncertainty of passengers intentions in route choice, and on the following joining of the computational results received for the sub-models. The availability of the approach was illustrated by the examples of the typical patterns in the destination stops choice and in hourly passengers’ flow from the departure stops. The model testing shows the high correlation of the simulated passengers’ flow with the results of the real observations.