Pre-processing, repairing and transfer functions can help binary electromagnetism-like algorithms

RICARDO JAVIER SOTO DE GIORGIS, BRODERICK CRAWFORD LABRIN, Alexis Muñoz, Franklin Johnson, Fernando Paredes

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

13 Scopus citations

Abstract

The Electromagnetism-like algorithm is a relatively modern metaheuristic based on the attraction-repulsion mechanism of particles in the context of electromagnetism theory. This paper focuses on improving performance of this metaheuristic when solving binary problems. To this end, we incorporate three elements: pre-processing, repairing, and transfers functions. The pre-processing allows to reduce the size of instances, while repairing eliminates those potential solutions that violate the constraints. Finally, the incorporation of a transfer function adapts the solutions to a binary domains. We illustrate experimental results where the incorporation of these elements improve the resolution phase, when solving a set of 65 non-unicost set covering problems.

Original languageEnglish
Title of host publicationArtificial Intelligence Perspectives and Applications - Proceedings of the 4th Computer Science On-line Conference 2015, CSOC 2015
EditorsRadek Silhavy, Roman Senkerik, Zuzana Kominkova Oplatkova, Zdenka Prokopova, Petr Silhavy
PublisherSpringer Verlag
Pages89-97
Number of pages9
ISBN (Print)9783319184753
DOIs
StatePublished - 1 Jan 2015
Event4th Computer Science On-line Conference, CSOC 2015 - Zlin, Czech Republic
Duration: 27 Apr 201530 Apr 2015

Publication series

NameAdvances in Intelligent Systems and Computing
Volume347
ISSN (Print)2194-5357

Conference

Conference4th Computer Science On-line Conference, CSOC 2015
CountryCzech Republic
CityZlin
Period27/04/1530/04/15

Keywords

  • Electromagnetism-like Algorithm
  • Metaheuristics
  • Pre-processing

Fingerprint Dive into the research topics of 'Pre-processing, repairing and transfer functions can help binary electromagnetism-like algorithms'. Together they form a unique fingerprint.

Cite this