Towards 3D Architectures: A Comparative Study on Cellular GAs Dimensionality

Cellular genetic algorithms (cGAs) have shown their ability to solve not only difficult optimization problems, but also outperform centralized genetic algorithms (GAs) in terms of efficiency and efficacy. The study herein presented aims to analyze and compare 2D and 3D cellular GAs, while maintaining in general their configuration constraints such as population size, neighbourhood radius, local selection method, replacement polices, among others. A primary objective of this paper is to provide a wide insight into the advantages of increasing cellular dimensionality for future development of 3D adaptive optimization engine architectures.