Home
Resource
New results for optimization in stochastic networks
PDF
PDF format is widely accepted and good for printing.
Plug-in required
PDF-1(2.22 MB)
Citation & Export
View Usage Statistics
Staff View

Citation & Export
Hide

Simple citation

Unuvar, Merve. New results for optimization in stochastic networks. Retrieved from https://doi.org/doi:10.7282/T39P30D9

Export

Click here for information about Citation Management Tools at Rutgers.

Statistics
Hide

Description

TitleNew results for optimization in stochastic networks
NameUnuvar, Merve (author); Prekopa, Andras (chair); Boros, Endre (internal member); BEN-ISRAEL, ADI (internal member); Baykal-Gursoy, Melike (internal member); Subasi, Mine (outside member); Rutgers University; Graduate School - New Brunswick
Date Created2012
Other Date2012-10 (degree)
SubjectOperations Research, Stochastic processes, Structural optimization
Extentviii, 141 p. : ill.
DescriptionWe are interested in single commodity stochastic network design problems under probabilistic constraint with discrete and continuous random variables. We use a stochastic programming model under probabilistic constraint (also called a chance-constrained model) to study these problems. The problem addressed in this research is how to find minimum cost optimal capacities at the nodes and/or arcs subject to the constraint that the demands should be met on a prescribed probability level (reliability constraint). In our first problem formulation, we formulate the reliability constraint in terms of the Gale-Hoffman feasibility inequalities. In latter formulations, we allow system to meet the demand at least $k$-out-of-$n$ and consecutive $k$-out-of-$n$ periods. The number of reliability constraints, in both cases, increases exponentially with the size of the nodes and therefore we identify the redundant constraints and reduce their number with elimination methods. Even with the reduced number of inequalities, it is not simple to solve probabilistic constrained stochastic network problems due to the large number of efficient points that satisfy the probabilistic condition. To overcome the size limitation of the problem, we develop a new theorem for efficient point generation in the case when the random variables are discrete, and we use hybrid cutting plane / supporting hyperplane algorithm in the case when the random variables are continuous.
NotePh.D.
NoteIncludes bibliographical references
NoteIncludes vita
Noteby Merve Unuvar
Genretheses, ETD doctoral
Persistent URLhttps://doi.org/doi:10.7282/T39P30D9
Languageeng
CollectionGraduate School - New Brunswick Electronic Theses and Dissertations
Organization NameRutgers, The State University of New Jersey
RightsThe author owns the copyright to this work.
Version 8.5.5
Rutgers University Libraries - Copyright ©2024