EVOLUTIONARY ALGORITHM PARAMETER FITNESS: AN EXPLORATORY STUDY

Andrew Manikas and Michael Godfrey

International Journal of Management and Marketing Research, 2011, vol. 4, issue 3, 35-44

Abstract: Genetic algorithms were the first evolutionary algorithm designed. These algorithms simulate natural selection to produce good solutions quickly for complex problems. Job shop scheduling with sequence dependent setup times is an NP-hard problem – any algorithm to optimize this problem has an exponential time. We explore which parameters for a genetic algorithm allow it to solve these job shop problems in limited time trials. Prior literature assumes the use of these parameters is beneficial, and the parameter values are selected either based on prior research values or from design of experiments on a limited range of parameter values. Multiple linear regression is used to determine which parameters can significantly improve solutions. Our results show that a proportional 50/50 crossover parameter and a large population size are the two parameters to obtain good solutions in a constrained time environment.

Keywords: Evolutionary Algorithms; Genetic Algorithm; Regression; Job Shop Scheduling (search for similar items in EconPapers)
JEL-codes: C20 C60 (search for similar items in EconPapers)
Date: 2011
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://www.theibfr2.com/RePEc/ibf/ijmmre/ijmmr-v4n3-2011/IJMMR-V4N3-2011-3.pdf (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:ibf:ijmmre:v:4:y:2011:i:3:p:35-44

Access Statistics for this article

International Journal of Management and Marketing Research is currently edited by Terrance Jalbert

More articles in International Journal of Management and Marketing Research from The Institute for Business and Finance Research
Bibliographic data for series maintained by Mercedes Jalbert ( this e-mail address is bad, please contact ).