A clustering method using new modularity indices and a genetic algorithm with extended chromosomes

Research output: Contribution to conferencePaper

5 Citations (Scopus)

Abstract

Module definition entails clustering an original product architecture into independent or coordinated modules. Clustering algorithms based on Design Structure Matrices (DSMs) for defining modules have been widely studied. After reviewing existing clustering algorithms, we introduce simple new metrics that can be used as modularity indices bounded between 0 and 1 and also utilized as the objective functions to obtain optimal DSMs including the maximized interactions within modules and the minimized interactions between modules. As a search strategy for clustering modules, a combinatorial genetic algorithm using a new extended chromosome approach and modified operators for the chromosome is suggested. The module definition results indicated that the proposed clustering method using new modularity indices and genetic algorithm helps obtain optimal modular product architectures more logically.

Original languageEnglish (US)
Pages167-176
Number of pages10
StatePublished - Jan 1 2014
Event16th International Dependency and Structure Modelling Conference, DSM 2014 - Paris, France
Duration: Jul 2 2014Jul 4 2014

Other

Other16th International Dependency and Structure Modelling Conference, DSM 2014
CountryFrance
CityParis
Period7/2/147/4/14

Fingerprint

Chromosomes
Clustering algorithms
Genetic algorithms
Optimal design

All Science Journal Classification (ASJC) codes

  • Building and Construction
  • Civil and Structural Engineering

Cite this

Jung, S., & Simpson, T. W. (2014). A clustering method using new modularity indices and a genetic algorithm with extended chromosomes. 167-176. Paper presented at 16th International Dependency and Structure Modelling Conference, DSM 2014, Paris, France.
Jung, Sangjin ; Simpson, Timothy William. / A clustering method using new modularity indices and a genetic algorithm with extended chromosomes. Paper presented at 16th International Dependency and Structure Modelling Conference, DSM 2014, Paris, France.10 p.
@conference{5140b5ed8a2c4b52b141213c8f732974,
title = "A clustering method using new modularity indices and a genetic algorithm with extended chromosomes",
abstract = "Module definition entails clustering an original product architecture into independent or coordinated modules. Clustering algorithms based on Design Structure Matrices (DSMs) for defining modules have been widely studied. After reviewing existing clustering algorithms, we introduce simple new metrics that can be used as modularity indices bounded between 0 and 1 and also utilized as the objective functions to obtain optimal DSMs including the maximized interactions within modules and the minimized interactions between modules. As a search strategy for clustering modules, a combinatorial genetic algorithm using a new extended chromosome approach and modified operators for the chromosome is suggested. The module definition results indicated that the proposed clustering method using new modularity indices and genetic algorithm helps obtain optimal modular product architectures more logically.",
author = "Sangjin Jung and Simpson, {Timothy William}",
year = "2014",
month = "1",
day = "1",
language = "English (US)",
pages = "167--176",
note = "16th International Dependency and Structure Modelling Conference, DSM 2014 ; Conference date: 02-07-2014 Through 04-07-2014",

}

Jung, S & Simpson, TW 2014, 'A clustering method using new modularity indices and a genetic algorithm with extended chromosomes' Paper presented at 16th International Dependency and Structure Modelling Conference, DSM 2014, Paris, France, 7/2/14 - 7/4/14, pp. 167-176.

A clustering method using new modularity indices and a genetic algorithm with extended chromosomes. / Jung, Sangjin; Simpson, Timothy William.

2014. 167-176 Paper presented at 16th International Dependency and Structure Modelling Conference, DSM 2014, Paris, France.

Research output: Contribution to conferencePaper

TY - CONF

T1 - A clustering method using new modularity indices and a genetic algorithm with extended chromosomes

AU - Jung, Sangjin

AU - Simpson, Timothy William

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Module definition entails clustering an original product architecture into independent or coordinated modules. Clustering algorithms based on Design Structure Matrices (DSMs) for defining modules have been widely studied. After reviewing existing clustering algorithms, we introduce simple new metrics that can be used as modularity indices bounded between 0 and 1 and also utilized as the objective functions to obtain optimal DSMs including the maximized interactions within modules and the minimized interactions between modules. As a search strategy for clustering modules, a combinatorial genetic algorithm using a new extended chromosome approach and modified operators for the chromosome is suggested. The module definition results indicated that the proposed clustering method using new modularity indices and genetic algorithm helps obtain optimal modular product architectures more logically.

AB - Module definition entails clustering an original product architecture into independent or coordinated modules. Clustering algorithms based on Design Structure Matrices (DSMs) for defining modules have been widely studied. After reviewing existing clustering algorithms, we introduce simple new metrics that can be used as modularity indices bounded between 0 and 1 and also utilized as the objective functions to obtain optimal DSMs including the maximized interactions within modules and the minimized interactions between modules. As a search strategy for clustering modules, a combinatorial genetic algorithm using a new extended chromosome approach and modified operators for the chromosome is suggested. The module definition results indicated that the proposed clustering method using new modularity indices and genetic algorithm helps obtain optimal modular product architectures more logically.

UR - http://www.scopus.com/inward/record.url?scp=84907075048&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84907075048&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:84907075048

SP - 167

EP - 176

ER -

Jung S, Simpson TW. A clustering method using new modularity indices and a genetic algorithm with extended chromosomes. 2014. Paper presented at 16th International Dependency and Structure Modelling Conference, DSM 2014, Paris, France.