A Novel Approach for Analyzing Passengers’ Train Choice Behavior in Urban Rail Transit System

Abstract

AbstractIn recent years, with the improvement of information management, a large number of passenger travel data and train operation data are collected automatically in urban rail transit (URT) system. A large amount of archived data about when and where passengers travel is obtained through smart cards, which may be used to research passenger behavior and traffic demand management. Combined with Automatic fare collection (AFC) and Automatic Vehicle Location (AVL) data, it can assist the management department to make planning, scheduling, performance evaluation and so on. This study proposes a data-driven approach to understand passenger’s train choice behavior in congested subway system. First, the model gives a simple and reasonable assumption, that is, “get off-leave now” (GO-LN). The advantage of this assumption is that it does not rely on manual investigation. Second, an adaptive model by using AFC and AVL data is developed to estimate the probability of passenger’s train choice behavior, and the expectation maximization (EM) algorithm is used to solve the problem. Finally, a case study carried out on the Ba Tong Line of Beijing Metro is discussed to illustrate the effectiveness and robustness of the proposed model. The method proposed in this study is helpful to fully understand passenger travel behavior from incomplete data sets, infer the distribution of network passenger flow, formulate traffic regulations and provide suggestions for subway operators to implement corresponding measures. KeywordsUrban rail transit systemTrain choice behaviorData-driven approachEM algorithm