Sexual selection for genetic algorithms

Abstract

Genetic Algorithms (GA) have been widely used in operations research and optimization since first proposed. A typical GA comprises three stages, the encoding, the selection and the recombination stages. In this work, we focus our attention on the selection stage of GA, and review a few commonly employed selection schemes and their associated scaling functions. We also examine common problems and solution methods for such selection schemes. We then propose a new selection scheme inspired by sexual selection principles through female choice selection, and compare the performance of this new scheme with commonly used selection methods in solving some well-known problems including the Royal Road Problem, the Open Shop Scheduling Problem and the Job Shop Scheduling Problem.