Empirical Analysis of Traveling Backwards and Passenger Flows Reassignment on a Metro Network with Automatic Fare Collection (AFC) Data and Train Diagram

Abstract

Metro (subway) systems are becoming overcrowded in some of China’s mega-cities, such as Beijing, Shanghai, and Guangzhou. As many passengers are unable to board trains on an overcrowded metro network during peak periods, some of them are willing to spend more time and energy by traveling backwards in order to secure a seat or even room for standing. Traditional studies on travel behavior analysis and transit assignment models seldom deal with this situation. We propose a methodology including the affinity propagation cluster method with between-within-proportion (BWP) index and an adaptive “0-1” model named the traveling backwards model (TBM) to identify the phenomenon of “traveling backwards” and to reassign passenger flows on a metro network using automatic fare collection (AFC) data and actual train diagrams. As a numerical example, this integrated approach is applied to the Beijing metro system. Our research shows that the affinity propagation cluster method with BWP index and implicit enumeration algorithm for TBM work well. TBM is a good replacement for the existing assignment model and travel behavior analysis, especially for those mega-cities’ networks in peak hours.