A Structured-Population Genetic-Algorithm based on Hierarchical Hypercube of Genes Expressions
Mohamed Belal, Prof.Mohamed Hagag · International Journal of Computer Applications · 2013
Structured-population Genetic Algorithm (GA) usually leads to more superior performance than the panmictic genetic algorithm; since it can control two opposite processes, namely exploration and exploitation in the search space.Several spatially structured-population GAs have been introduced in the literature such as cellular, patchwork, island-model, terrain-based A, graph-based, religion-based and social-based GA.All the aforementioned works did not construct the subpopulations based on the genes information of the individuals themselves.The structuring of sub-populations based on this information might help in attaining better performance and more efficient search strategy.In this paper, the structured population is represented as hierarchical hypercube of subpopulations that are dynamically constructed and adapted at search time.Each sub-population represents a sub-division of the real genes space.This structure could help in directing the search towards the promising sub-spaces.Finally, a comparative study with other known structured population GA is provided.