Real-coded ECGA for solving decomposable real-valued optimization problems
Journal
Studies in Computational Intelligence
Journal Volume
157
Pages
61-86
Date Issued
2008
Author(s)
Abstract
We present the real-coded extended compact genetic algorithm (rECGA) for solving decomposable real-valued optimization problems. Mutual information between real-valued variables is employed to measure variable interaction or dependency, and the variable clustering and aggregation algorithms are proposed to identify the substructures of a problem through partitioning variables. Then, mixture Gaussian probability density function is estimated to model the promising individuals for each substructure, and the sampling of multivariate Gaussian probability density function is carried out by adopting Cholesky decomposition. Facet analyses are made about the population sizing and sampling of the factorization based on the Gaussian probability density function. Finally, experiments on decomposable test functions are conducted. The results illustrate that the rECGA is able to correctly identify the substructure of decomposable problems with linear or nonlinear correlations, and achieves a good scalability. © 2008 Springer-Verlag Berlin Heidelberg.
Type
book