Skip to main navigation Skip to search Skip to main content

Distribution of computational effort in parallel MOEA/D

  • Juan J. Durillo
  • , Qingfu Zhang
  • , Antonio J. Nebro
  • , Enrique Alba

Research output: Chapters, Conference Papers, Creative and Literary WorksRGC 32 - Refereed conference paper (with host publication)peer-review

Abstract

MOEA/D is a multi-objective optimization algorithm based on decomposition, which consists in dividing a multi-objective problem into a number of single-objective sub-problems. This work presents two variants, called pMOEA/Dv1 and pMOEA/Dv2, of a new parallel model of MOEA/D that have been developed under the observation that different sub-problems may require different computational effort, and thus, demand different number of evaluations. Our interest in this paper is to analyze whether the proposed models are able of outperforming the MOEA/D in terms of the quality of the computed fronts. To cope with this issue, our proposals have been evaluated using a benchmark composed of eight problems and the obtained results have been compared against MOEA/D-DE, an extension of the original MOEA/D where new individuals are generated by an operator taken from differential evolution. Our experiments show that some configurations of pMOEA/Dv1 and pMOEA/Dv2 have been able to compute fronts of higher quality than MOEA/D-DE in many of the evaluated problems, giving room for further research in this line. © Springer-Verlag Berlin Heidelberg 2011.
Original languageEnglish
Title of host publicationLearning and Intelligent Optimization
Subtitle of host publication5th International Conference, LION 5, Selected Papers
PublisherSpringer Verlag
Pages488-502
Volume6683 LNCS
ISBN (Print)9783642255656
DOIs
Publication statusPublished - 2011
Externally publishedYes
Event5th International Conference on Learning and Intelligent Optimization, LION 2011 - Rome, Italy
Duration: 17 Jan 201121 Jan 2011

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume6683 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference5th International Conference on Learning and Intelligent Optimization, LION 2011
PlaceItaly
CityRome
Period17/01/1121/01/11

Fingerprint

Dive into the research topics of 'Distribution of computational effort in parallel MOEA/D'. Together they form a unique fingerprint.

Cite this