Tailoring ε-MOEA to concept-based problems

Amiram Moshaiov*, Yafit Snir

*Corresponding author for this work

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


Concept-based MOEAs are tailored MOEAs that aim at solving problems with a-priori defined subsets of solutions that represent conceptual solutions. In general, the concepts' subsets may be associated with different search spaces and the related mapping into a mutual objective space could have different characteristics from one concept to the other. Of a particular interest are characteristics that may cause premature convergence due to local Pareto-optimal sets within at least one of the concept subsets. First, the known ε-MOEA is tailored to cope with the aforementioned problem. Next, the performance of the new algorithm is compared with C 1-NSGA-II. Concept-based test cases are devised and studied. In addition to demonstrating the significance of premature convergence in concept-based problems, the presented comparison suggests that the proposed tailored MOEA should be preferred over C 1-NSGA-II. Suggestions for future work are also included.

Original languageEnglish
Title of host publicationParallel Problem Solving from Nature, PPSN XII - 12th International Conference, Proceedings
Number of pages10
EditionPART 2
StatePublished - 2012
Event12th International Conference on Parallel Problem Solving from Nature, PPSN 2012 - Taormina, Italy
Duration: 1 Sep 20125 Sep 2012

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume7492 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference12th International Conference on Parallel Problem Solving from Nature, PPSN 2012


Dive into the research topics of 'Tailoring ε-MOEA to concept-based problems'. Together they form a unique fingerprint.

Cite this