Parallel Implementations of the Frank-Wolfe Algorithms for the Traffic Assignment Problem

Date of Award

Fall 2013

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Computational Modeling & Simulation Engineering

Program/Concentration

Modeling and Simulation

Committee Director

Duc T. Nguyen

Committee Director

Man Wo Ng

Committee Member

Andrew Collins

Call Number for Print

Special Collections LD4331.E58 A45 2013

Abstract

Transportation planners seek to understand how to best invest limited resources for future transportation network development. The traffic assignment problem is one algorithm of great importance to planners because it provides insight into how traffic will flow within the network. The Frank-Wolfe algorithm is a traditional solution method for this optimization problem, but it has been characterized by its slow rate of convergence and poor computational performance. This thesis examines and implements several modern advancements in this algorithm which are designed to improve the rate of convergence.

In addition to algorithm changes, another method to improve the performance of an algorithm is to perform some or all of the calculations in parallel, thus reducing the wall clock time to obtain the result. Modern desktop computers commonly include multiple processors. To further improve the performance, the Message Passing Interface (MPI) parallelization technique is applied to the Frank-Wolfe algorithms in order to accelerate the most computationally-intensive parts of the algorithm. Numerical results on several real transportation networks are used to validate the developed parallel procedures.

Rights

In Copyright. URI: http://rightsstatements.org/vocab/InC/1.0/ This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).

DOI

10.25777/rxgk-ey49

This document is currently not available here.

Share

COinS