Automated Data Driven Method For Route Choice Estimation In Metro Systems

The rapid growth of population in major cities increases the demand for subway systems,which cost transit agencies operating at over capacity with crowding issues.Better understanding of passenger route choice behavior is the prerequisite for policy maker and operation planers to make informed decisions to solve the crowding problems.Previous research on path choice mostly used survey,which are constrained by coverage,accuracy and cost.The emerging technologies of Automated Fare Collection(AFC)and Automatic Vehicle Location(AVL)data collection provides opportunity to address this problem from new perspective.After reviewing and discussing current research on route choice,this thesis proposes an estimation method for OD path fraction using AFC and AVL data.This method consists of two main parts:the first part uses Gaussian mixture model to estimate passengers denied boarding probability at key stations(boarding and transfer),the second part take denied boarding into consideration and formulates maximum likelihood estimation problem to estimate the path fraction for different OD pairs at different times.The method is validated using both synthetic data and actual survey data.The results show that the crowding impact on passenger's route choice is heterogeneous across times and space.The simulation of different scenarios indicates the importance of travel guidance in relieving crowding impact in daily lives,and explored the possibility to combine discrete choice model in this direction.