Chromosome differentiation for the application of parent-centric real-parameter crossover operators

摘要: Parent-centric real-parameter crossover operators create the offspring in the neighborhood of one of the parents, the female parent, using a probability distribution. The other parent, the male parent, is considered to define the range of the probability distribution. In this paper, we present a process that determines the individuals in the population that may become female or/and male parents. An important advantage of this proposal is that it makes possible the design of two kind of realcoded genetic algorithms: ones specialized for global search and ones that are effective local searchers. We introduce an hybrid technique that combines these real-coded genetic algorithms with the aim of achieving robust performance on a wide range of problems. The experiments show that this technique may reach high robustness levels, as compared with other real-coded genetic algorithms and hybrid real-coded genetic algorithms proposed in the literature.