Nonconvex functions optimization using an estimation of distribution algorithm based on a multivariate extension of the clayton copula

Abstract

This paper presents a copula-based estimation of a distribution algorithm with parameter updating for numeric optimization problems. This model implements an estimation of a distribution algorithm using a multivariate extension of Clayton's bivariate copula (MEC-EDA) to estimate the conditional probability for generating a population of individuals. Moreover, the model uses traditional mutation and elitism operators jointly with a heuristic for a population restarting in the evolutionary process. We show that these approaches improve the overall performance of the optimization compared to other copula-based EDAs. © 2014 Springer International Publishing Switzerland.