Differential evolution strategies for multi-objective optimization

Ashish M. Gujarathi*, B. V. Babu

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

Multi-objective optimization (MOO) using evolutionary algorithms has gained popularity in the recent past due to its ability of producing number of solutions in a single run and handling multiple objectives simultaneously. In this effort, several MOO algorithms are developed. In this manuscript several strategies of multi-objective differential evolution algorithm (namely, MODE-I, MODE-III, elitist MODE and hybrid MODE) are briefly discussed. Three important unconstrained test problems are considered for validating the performance (in terms of Pareto front and convergence & diversity metrics) of strategies of MODE algorithm with other popular algorithms from literature. It is observed that the strategies of MODE algorithm are in general able to produce Pareto front with good convergence to the true Pareto front.

Original languageEnglish
Title of host publicationProceedings of the International Conference on Soft Computing for Problem Solving, SocProS 2011
Pages63-71
Number of pages9
EditionVOL. 1
DOIs
Publication statusPublished - 2012
Externally publishedYes
EventInternational Conference on Soft Computing for Problem Solving, SocProS 2011 - Roorkee, India
Duration: Dec 20 2011Dec 22 2011

Publication series

NameAdvances in Intelligent and Soft Computing
NumberVOL. 1
Volume130 AISC
ISSN (Print)1867-5662

Other

OtherInternational Conference on Soft Computing for Problem Solving, SocProS 2011
Country/TerritoryIndia
CityRoorkee
Period12/20/1112/22/11

Keywords

  • Differential Evolution
  • Evolutionary Algorithms (EAs)
  • Multi-objective Differential Evolution (MODE)
  • Multi-objective optimization (MOO)
  • Pareto front

ASJC Scopus subject areas

  • General Computer Science

Fingerprint

Dive into the research topics of 'Differential evolution strategies for multi-objective optimization'. Together they form a unique fingerprint.

Cite this