A new Look into the Imperialist Competitive Algorithm

Genetic Algorithm is currently used as a solution to various problems in a wide range of disciplines. In order to improve the convergence rates of Genetic Algorithms, a new branch of evolutionary computation called “cultural algorithms” has been introduced that provides the possibility of exchanging in-formation in the population component of a conventional genetic algorithm. As expected in some applications the convergence rates obtained by cultural algo-rithms such as Imperialist Competitive Algorithm (ICA) were better or at least similar to those obtained by applying the genetic algorithms. In this paper, we aim to propose a new perspective, which is assumed to increase the capability of exchanging information in the population component of the evolutionary al-gorithms by providing an infrastructure for dialogues. In other words, we divide a population into several regions equivalent to the empires in ICA where instead of the competition among regions we introduce the notion of Dialogue among regions, which is assumed to improve the convergence rate towards the absolute minimum.