Skip to main navigation Skip to search Skip to main content

A mean shift assisted differential evolution algorithm

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

It is well known that Differential Evolution (DE) algorithm has been widely applied to solve global optimization problems during the last decades. DE is usually criticized for the slow convergence. To improve the algorithm performance, we propose an algorithm called MSDE that utilizes a local search operator based on mean shift. In MSDE, one offspring solution is generated by the mean shift based search operator, and the others are created by the DE search operator. A test suite of 12 benchmark functions with different characteristics are chosen to evaluate our approach. The experimental results suggest that MSDE can successfully improve the performance of DE and have a faster convergence rate on the given test suite.

Original languageEnglish
Title of host publicationBio-inspired Computing – Theories and Applications - 11th International Conference, BIC-TA 2016, Revised Selected Papers
EditorsLinqiang Pan, Maoguo Gong, Tao Song, Gexiang Zhang, Tao Song
PublisherSpringer Verlag
Pages163-172
Number of pages10
ISBN (Print)9789811036132
DOIs
StatePublished - 2016
Event11th International Conference on Bio-inspired Computing – Theories and Applications, BIC-TA 2016 - Xian, China
Duration: 28 Oct 201630 Oct 2016

Publication series

NameCommunications in Computer and Information Science
Volume682
ISSN (Print)1865-0929

Conference

Conference11th International Conference on Bio-inspired Computing – Theories and Applications, BIC-TA 2016
Country/TerritoryChina
CityXian
Period28/10/1630/10/16

Keywords

  • Differential evolution
  • Global optimization
  • Mean shift
  • Search operator

Fingerprint

Dive into the research topics of 'A mean shift assisted differential evolution algorithm'. Together they form a unique fingerprint.

Cite this