Title

MODELS AND SOLUTION ALGORITHMS FOR ASYMMETRIC TRAFFIC AND TRANSIT ASSIGNMENT PROBLEMS

Date of Award:

2014

Document Type:

Dissertation

Degree Name:

Doctor of Philosophy (PhD)

Department:

Civil and Environmental Engineering

Advisor/Chair:

Kevin Heaslip

Co-Advisor/Chair:

Anthony Chen

Abstract

Modeling the transportation system is important because it provides a “common ground” for discussing policy and examining the future transportation plan required in practices. Generally, modeling is a simplified representation of the real world; however, this research added value to the modeling practice by investigating the asymmetric interactions observed in the real world in order to explore potential improvements of the transportation modeling.

The Asymmetric Transportation Equilibrium Problem (ATEP) is designed to precisely model actual transportation systems by considering asymmetric interactions of flows. The enhanced representation of the transportation system by the ATEP is promising because there are various asymmetric interactions in real transportation such as intersections, highway ramps, and toll roads and in the structure of the transit fares.

This dissertation characterizes the ATEP with an appropriate solution algorithm and its applications. First, the research investigates the factors affecting the convergence of the ATEP. The double projection method is applied to various asymmetric types and complexities in the different sizes of networks in order to identify the influential factors including demand intensities, network configuration, route composition between modes, and sensitivity of the cost function. Secondly, the research develops an enhancement strategy for improvement in computational speed for the double projection method. The structural characteristics of the ATEP are used to develop the convergence enhancement strategy that significantly reduces the computational burdens.

For the application side, instances of asymmetric interactions observed in in-vehicle crowding and the transit fare structure are modeled to provide a suggestion on policy approach for a transit agency. The direct application of the crowding model into the real network indicates that crowd modeling with multi user classes could influence the public transportation system planning and the revenue achievement of transit agencies. Moreover, addition of the disutility factor, crowding, not always causes the increase of disutility from the transit uses. The application of the non-additive fare structure in the Utah Transit Authority (UTA) network addresses the potential of the distance-based fare structure should the UTA make a transition to this fare structure from their current fare model. The analysis finds that the zero base fare has the highest potential for increasing the transit demand. However, collecting less than $0.50 with a certain buffer distance for the first boarding has potential for attracting the users to UTA's transit market upon the fare structure change.

This document is currently not available here.

Share

COinS