Study On Passenger Travel Route Selection In Urban Rail Transit Network

With the rapid development of China’s urban rail transit,urban rail transit presents the characteristics of network operation,and the passenger flow of urban rail transit increases rapidly.There are many and flexible and complex travel schemes for urban rail transit.AFC data can be used to analyze the passenger travel characteristics of urban rail transit.AFC data can record the spatial and temporal information of passengers entering and leaving the station.The AFC data are preprocessed by data mining technology to study the behavior of travelers’ choice of travel path.Based on this,this paper takes the train timetable and AFC data of urban rail transit as the research object,and uses data mining technology to study the coupling degree between the individual passenger’s travel path and its possible travel path of urban rail transit,so as to improve the passenger’s travel path selection and possible path and fit degree in urban rail transit.The research of this paper can provide reference value for urban rail transit passenger travel path selection,and provide theoretical basis for the operation of urban rail transit passenger flow allocation.The specific research contents of this paper are as follows:(1)Firstly,analyze the constituent elements of the travel time of urban rail transit passengers,and reclassify the travel time of passengers.On the basis of studying the physical topology network of urban rail transit passenger travel,train timetable and AFC data of passengers in and out of the station,the passenger travel space-time network is constructed,and the passenger travel path search algorithm in the space-time network is designed.Finally,the passenger travel paths with and without transfer stations are searched respectively under different constraints,and the depth-first traversal search method is used to obtain the set of feasible routes for passengers to travel,and the Logit model is used to calculate the choice probability of passenger travel routes.(2)The passenger and train are matched,and the passenger probability estimation model is established in the matching process.By studying the effective train set,this paper constructs the passenger riding probability estimation model,which estimates the selection probability of each train in the effective train set,and then assigns passengers to different trains;for the study of the estimation method of passenger travel time distribution,the influencing factors of travel time distribution characteristics can be integrated,which provides a reference for using EM algorithm to estimate the constructed travel time distribution probability The solution of the model provides a reference.This method solves the time-consuming and laborious problem of dispatchers’ actual tracking and research,and provides a reasonable calibration method for passenger flow assignment related research.(3)Through the comparative analysis of the number of effective space-time paths,the number of available trains in the effective train set,the time-space node of the train arriving at the platform and the transfer times of the effective space-time path,the passengers are classified.Through the analysis of the travel time of the passengers entering and leaving the station,the train number chain and the probability of selecting the train are estimated,and the train number chain under the passenger travel choice path is obtained.Finally,based on the operation data of Tianjin urban rail transit,three representative stations are selected for on-the-spot follow-up survey to obtain the travel time distribution characteristics of passengers entering and leaving the station,and the results are compared and verified;the AFC data of several passengers and the itinerary of the only effective travel path and multiple travel paths are selected to verify the urban rail transit system based on AFC data and train timetable The method of passenger travel path estimation and passenger travel probability estimation.