Dissertation/Thesis Abstract

Modeling alternatives to solve the connected graph k-clustering problem
by Es-Sakli, Nada, M.S., Lehigh University, 2011, 74; 1488515
Abstract (Summary)

In this thesis we explore modeling alternatives for the graph k-clustering problem. For k ≥ 2, the problem is to find a minimum cost partition of k connected subgraphs of a graph G=(V, E). We used the AMPL modeling framework to develop both a Mixed Integer Linear Programming model that uses a multi-commodity formulation and a column generation procedure that uses a restricted master problem and generates variables by solving a Pricing problem.

Indexing (document details)
Advisor: Belotti, Pietro, Terlaky, Tamas
Commitee:
School: Lehigh University
Department: Industrial Engineering
School Location: United States -- Pennsylvania
Source: MAI 49/03M, Masters Abstracts International
Source Type: DISSERTATION
Subjects: Industrial engineering, Operations research
Keywords: Clustering, Clustering problem, Column generation, Connected graph, K-clustering, Partitioning
Publication Number: 1488515
ISBN: 9781124454276
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest