Publication detail

Multivariate Gaussian Copula in Estimation of Distribution Algorithm with Model Migration

HYRŠ, M. SCHWARZ, J.

Original Title

Multivariate Gaussian Copula in Estimation of Distribution Algorithm with Model Migration

Type

conference paper

Language

English

Original Abstract

The paper presents a new concept of an island-based model of Estimation of Distribution Algorithms (EDAs) with a bidirectional topology in the field of numerical optimization in continuous domain. The traditional migration of individuals is replaced by the probability model migration. Instead of a classical joint probability distribution model, the multivariate Gaussian copula is used which must be specified by correlation coefficients and parameters of a univariate marginal distributions. The idea of the proposed Gaussian Copula EDA algorithm with model migration (GC-mEDA) is to modify the parameters of a resident model respective to each island by the immigrant model of the neighbour island. The performance of the proposed algorithm is tested over a group of five well-known benchmarks.

Keywords

Estimation of distribution algorithms, Copula Theory, Sklar's theorem, multivariate Gaussian copula, island-based model, model migration, optimization problems. 

Authors

HYRŠ, M.; SCHWARZ, J.

RIV year

2014

Released

11. 12. 2014

Publisher

Institute of Electrical and Electronics Engineers

Location

Piscataway

ISBN

978-1-4799-4492-7

Book

2014 IEEE Symposium on Foundations of Computational Intelligence (FOCI) Proceedings

Pages from

114

Pages to

119

Pages count

6

BibTex

@inproceedings{BUT111681,
  author="Martin {Hyrš} and Josef {Schwarz}",
  title="Multivariate Gaussian Copula in Estimation of Distribution Algorithm with Model Migration",
  booktitle="2014 IEEE Symposium on Foundations of Computational  Intelligence (FOCI) Proceedings",
  year="2014",
  pages="114--119",
  publisher="Institute of Electrical and Electronics Engineers",
  address="Piscataway",
  doi="10.1109/FOCI.2014.7007815",
  isbn="978-1-4799-4492-7"
}