Research On Passenger Travel Route Choice Behavior Of Urban Rail Transit Considering Travelling Back

The ultra-large-scale urban rail transit system carries a huge passenger flow,and passengers’ travel route choice behavior also shows a trend of diversification and abnormality.Especially during periods of heavy load of road network,passengers may choose special travel routes to improve travel experience,and travelling back behavior is one of the typical phenomena.Travelling back behavior means that in some stations where the passenger flow load is relatively large,passengers take the opposite direction train at the origin station,and then turn back at the non-crowded station and go to the destination station.This behavior can avoid staying at the platform,shorten waiting time on the platform or get seats after getting on the train.In this paper,the travel route choice behavior of urban rail transit passengers in the presence of travelling back is studied,and methods for extracting stations where travelling back occur,identifying passengers with travelling back at stations level,and generating travelling back routes are proposed.A travel route choice model considering the travelling back behavior is constructed,which corrects the error caused by the traditional travel route choice model not considering the travelling back path.The model estimation results are more consistent with the actual situation of passenger travel.Firstly,taking Beijing metro as an example,based on passenger flow data such as AFC,analyze the distribution of passenger flow and its imbalance from the perspective of time and space.Then through the analysis of O-D’s travel time,the travelling back behavior is described and portrayed.The typical stations are selected to analyze the travel time characteristics of travelling back behavior.And the characteristics of the travelling back stations are analyzed from the perspective of passengers’ travel purpose and the geographical environment.Secondly,the data-driven method is used to carry out research on the problems of the travelling back behavior,involving occurring station extraction,travelling back passengers’ identification,travelling back routes generation,etc.A method for identifying travelling back occurring stations based on TF-IDF index quantification method and FCM clustering algorithm is proposed.Based on the stations where the travelling back behavior occurs,a method for identifying and generating travelling back routes based on the Gaussian mixture model and linear programming model is proposed.The EM algorithm and implicit enumeration algorithm are used to solve the model.The method solves the problem of passenger identification of travelling back and the generation of travelling back routes,and complements the effective route set in the study of travel route choice.In order to further reveal the mechanism of passengers’ travelling back behavior,a travel route choice model considering the travelling back is constructed.Firstly,it qualitatively analyzes the influencing factors of travel route choice behavior from three perspectives of individual passenger attributes,route attributes and operational service attributes,and selects the main influencing factors through a questionnaire survey.Then a generalized cost function considering the cost of travelling back is constructed,and a path size Logit model based on relative utility is established.The genetic algorithm is used to calibrate the model parameters.The model considers the influence of travelling back factors,supplements and revises the existing travel route choice model,and solves the problem of estimating the probability of passenger travel route choice with travelling back behavior.Finally,taking Beijing metro as an example to analyze and verify the method and model proposed in this paper.The results show that the travelling back passenger identification method and the travelling back route generation method can effectively reflect the actual travel route of the passengers,which is basically consistent with the survey data.Compared with the travel route choice model that does not consider the travelling back and adopts absolute utility,the calculation results of the model proposed in this paper have smaller errors with the survey data,and are more in line with the actual travel route choice of passengers.