Publication Details

Elliptical and Archimedean Copulas in Estimation of Distribution Algorithm

HYRŠ Martin and SCHWARZ Josef. Elliptical and Archimedean Copulas in Estimation of Distribution Algorithm. In: MENDEL 2015 21st International Conference on Soft Computing. Brno: Faculty of Mechanical Engineering BUT, 2015, pp. 19-26. ISBN 978-80-214-4984-8.
Czech title
Eliptické a Archimedovské kopule v algoritmu s pravděpodobnostním modelem
Type
conference paper
Language
english
Authors
Hyrš Martin, Ing., Ph.D. (DCSY FIT BUT)
Schwarz Josef, doc. Ing., CSc. (DCSY FIT BUT)
Keywords

Estimation of distribution algorithms, copula theory, multivariate copula sampling, Clayton copula, Gumbel copula, Frank copula, Gaussian copula, Student t-copula

Abstract

Estimation of distribution algorithm (EDA) is a variant of evolution algorithms, which is based on construction and sampling of probability model. Nowadays the copula theory is often utilized for the probability model estimation to simplify this process. We made comparison of two classes of copulas - elliptical and Archimedean ones - for the set of standard optimization benchmarks. The experimental results con firm our assumption that the elliptical copulas outperform the Archimedean ones namely in the case of the complex optimization problems.

Published
2015
Pages
19-26
Proceedings
MENDEL 2015 21st International Conference on Soft Computing
Conference
21st International Mendel Conference on Soft Computing, Brno, CZ
ISBN
978-80-214-4984-8
Publisher
Faculty of Mechanical Engineering BUT
Place
Brno, CZ
EID Scopus
2-s2.0-84938877457
BibTeX 
@INPROCEEDINGS{FITPUB11012,
   author = "Martin Hyr\v{s} and Josef Schwarz",
   title = "Elliptical and Archimedean Copulas in Estimation of Distribution Algorithm",
   pages = "19--26",
   booktitle = "MENDEL 2015 21st International Conference on Soft Computing",
   year = 2015,
   location = "Brno, CZ",
   publisher = "Faculty of Mechanical Engineering BUT",
   ISBN = "978-80-214-4984-8",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/11012"
}
Files
Back to top