Sexual Selection with Competitive/Co-operative Operators for Genetic Algorithms

In a standard genetic algorithm (GA), individuals reproduce asexually: any two organisms may be parents in crossover. Gender separation and sexual selection here inspire a model of gendered GA in which crossover takes place only between individuals of opposite sex and the GA’s evaluation, selection, and mutation strategies depend on gender. Consequently, a pattern of cross-gender cooperation and intra-gender competition emerges. A symbiotic relation between the selection and crossover operators also arises. Experimental results prove this strategy to be advantageous, significantly outperforming the standard GA, both in number of generations required and in the quality of solutions. KEY WORDS Genetic Algorithms, Sexual Selection, Evolutionary Computation

• Type: Conference
• Year: 2003
• Where: NCI
• Authors: Jose Sánchez-Velazco, John A. Bullinaria
• Keywords: Crossover Takes Place, Genetic Algorithms, NCI 2003, Neural Networks, Sexual Selection