Parameters quantification of genetic algorithm

Hari Mohan Pandey*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference proceeding (ISBN)

3 Citations (Scopus)

Abstract

This paper presents the importance of parameters tuning in global optimization algorithms. The primary objective of an experiment is to recognize the process. The experiments are carried out to learn the effect of various factors at different levels. Hence, identifying the optimal parameters setting is important for robust design. One of the most popular global optimization algorithms: genetic algorithm is considered in this study. The domain of inquiry is travelling salesman problem. The present study employs the Taguchi method that involves the use of an orthogonal array in the estimation of the factors. Taguchi approach has been widely applied in experimental design for problems with multiple factors. The use of Taguchi design is a novel idea—leads to efficient algorithms—can find a satisfactory solution in a few iterations, which improves the convergence speed and reduces the cost. Experimental results show that the Taguchi design is less sensitive to initial value of parameters. Two versions of genetic algorithms (with tuning and without tuning) are implemented. The analysis shows the superiority of genetic algorithm with tuning over genetic algorithm without tuning.

Original languageEnglish
Title of host publicationINFORMATION SYSTEMS DESIGN AND INTELLIGENT APPLICATIONS, VOL 2, INDIA 2016
EditorsSuresh Chandra Satapathy, Jyotsna Kumar Mandal, Siba K. Udgata, Vikrant Bhateja
PublisherSpringer-Verlag
Pages711-719
Number of pages9
Volume434
ISBN (Print)9788132227502
DOIs
Publication statusPublished - 3 Feb 2016
Event3rd International Conference on Information Systems Design and Intelligent Applications, INDIA 2016 - Visakhapatnam, India
Duration: 8 Jan 20169 Jan 2016

Publication series

NameAdvances in Intelligent Systems and Computing
Volume434
ISSN (Print)2194-5357

Conference

Conference3rd International Conference on Information Systems Design and Intelligent Applications, INDIA 2016
CountryIndia
CityVisakhapatnam
Period8/01/169/01/16

Keywords

  • Aguchi method
  • Genetic algorithm
  • Robust design
  • Travelling salesman problem
  • Taguchi method

Fingerprint Dive into the research topics of 'Parameters quantification of genetic algorithm'. Together they form a unique fingerprint.

Cite this