Abstracts

Many subway systems around the world are experiencing large increases in demand, especially during peak periods. As a result, they operate at (or near) capacity, with crowding an increasingly important issue that operators must deal with. Systems can be very congested with passengers frequently unable to board the first train.; This dissertation addresses the important problem of evaluating the performance of subway systems operating near capacity, especially from the passenger's point of view. It develops a key building block towards this goal, the Passenger-to-Itinerary Assignment Model (PIAM), to identify the boarding train(s) of each passenger using Automatic Fare Collection (AFC) and Automatic Vehicle Location (AVL) data.; PIAM is a model system, with a number of individual modules that interact but can also be used independently: the access/egress time model, left behind model, route choice model, and assignment model. The overall problem is challenging because of the large number of feasible itineraries a passenger may have. To deal with this, PIAM first estimates the left behind probabilities by station and time interval at the aggregate level and then assigns individual passengers to itineraries. The model is extended to incorporate trips involving route choice by estimating the route choice fractions and integrating them into the assignment model.; The research also proposes an alternative, computationally efficient method using principles from queuing theory to estimate time dependent aggregate crowding levels at stations and on platforms.; The methodology is validated using synthetic data and the performance compares favorably to a recent model (Horcher et al., 2017). It is also applied using actual data from a congested, subway system during peak hours. A series of applications are developed using PIAM output to assess the capacity utilization of the network, including train load estimation, passengers left behind, journey time components, crowding at stations, performance under special events, etc.