Dissertation/Thesis Abstract

Linear programming with the simplex algorithm: Parallelization and other optimizations
by Ankony, Robert C., M.S.C.S., California State University, Long Beach, 2012, 88; 1520889
Abstract (Summary)

The current trend in processor architectures towards multiple cores has led to a shift in program design emphasis to take advantage of architecture parallelization. The simplex algorithm presents an interesting case for parallelizing as it has many different components. As problem size increases to more and more nodes, the effect of even small performance improvements can be seen in the overall throughput of the problem.

In this study we look at parallelizing different parts of the simplex algorithm, and compare those changes to optimizations done to the serial program. We also look at other research related to parallelization and optimization of the simplex algorithm. Looking at a literature search, it is seen that there is active research in the field, using different approaches to improve performance of the problem, as well as better matching problem type to algorithm type. While the research done in this paper is sound, the active research in the field shows that there is much to be done and much still being done to improve the simplex problem.

Indexing (document details)
Advisor: Lam, Shui
Commitee: Englert, Burkhard, Hoffman, Michael, Lam, Shui
School: California State University, Long Beach
Department: Computer Science
School Location: United States -- California
Source: MAI 51/03M(E), Masters Abstracts International
Subjects: Computer Engineering, Computer science
Publication Number: 1520889
ISBN: 978-1-267-70286-9
Copyright © 2021 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy