dc.contributor.advisor |
Jamali, Prof Dr. A. R. M. Jalal Uddin |
|
dc.contributor.author |
Ali, Md. Ishaque |
|
dc.date.accessioned |
2018-08-14T06:48:52Z |
|
dc.date.available |
2018-08-14T06:48:52Z |
|
dc.date.copyright |
2015 |
|
dc.date.issued |
2015-06 |
|
dc.identifier.other |
ID 1151554 |
|
dc.identifier.uri |
http://hdl.handle.net/20.500.12228/387 |
|
dc.description |
dc.description This thesis is submitted to the Department of Mathematics, Khulna University of Engineering & Technology in partial fulfillment of the requirements for the degree of Master of Philosophy in Mathematics, June 2015. |
en_US |
dc.description |
Cataloged from PDF Version of Thesis. |
|
dc.description |
Includes bibliographical references (pages 82-90). |
|
dc.description.abstract |
Design of Experiment (DoE) is an important issue for developing mathematical model of
any physical problem especially when there exist large numbers of factors. Optimal Latin
Hypercube Design (LHD) is one of the well-known and used tools among the experimental
designs. For obtaining optimal LHD, Iterated Local Search (ILS) is one of the best way
among the heuristic approaches. Grosso et al. (2009) showed that ILS approach has the
ability to obtain a large number of maximin (Optimizations by maximizing minimum pairwise
distance) LHD where distances are measured in terms of Euclidian distance measure.
Several authors showed that rather than Euclidean distance measure other measures may
suitable for good DoE. Manhattan distance measure is one of them [Morris and Mitchell
(1995)]. In this research work, the main objective is to study the optimality of the maximin
LHD obtained by ILS approach regarding Manhattan distance measure. For this purpose,
ILS approach is implemented in windows environment (rather than Sun cluster, as Gross et
al. (2009) done). Extensive experiments are performed to obtain maximin LHD measured
in Euclidian distance measure. Then further experiments are reformed on those LHDs to
find the minimum pair-wise distance of each LHD measured in Manhattan distance. Those
values are compared with available one in the literature. It is noted that few values
(maximin LHD measured in Manhattan distance measure) are available in the literature. It
seems that the minimum pair-wise distance measured in Manhattan distance measure of the
maximin LHDs obtained by ILS approach, optimized in Euclidian distance measure are
comparable with those maximin LHDs obtained through other approaches but optimized in
the Manhattan distance measure. Moreover some further experiments are performed to find
out some new characteristics of those LHDs which may be used for further study. Some
improved maximin LHDs are also obtained in this experimental arena and are presented in
the thesis. |
en_US |
dc.description.statementofresponsibility |
Md. Ishaque Ali |
|
dc.format.extent |
90 pages |
|
dc.language.iso |
en_US |
en_US |
dc.publisher |
Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh. |
en_US |
dc.rights |
Khulna University of Engineering & Technology (KUET) thesis/dissertation/internship reports are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission. |
|
dc.subject |
Optimality Analysis |
en_US |
dc.subject |
Rectangular (Manhattan) Distance |
en_US |
dc.subject |
Maximin LHDs |
en_US |
dc.title |
Optimality Analysis on the Basis of Rectangular (Manhattan) Distance Measure for Maximin LHDs Obtained by the Iterated Local Search Heuristics Approach |
en_US |
dc.type |
Thesis |
en_US |
dc.description.degree |
Master of Philosophy in Mathematics |
|
dc.contributor.department |
Department of Mathematics |
|