Genetic Algorithms for Optimization
Adrian A. Hopgood · 2021
This chapter discusses the Genetic algorithms are the most popular type of evolutionary algorithm. On the other hand, if the selection method is too weak, less-fit individuals are given too much opportunity to reproduce and evolution may become too slow. The first type of approach, fitness-proportionate selection, is prone to both premature convergence and stalled evolution. The application of fitness scaling at these late stages of evolution is intended to strengthen the selection pressure in order to converge near the exact optimum. Eight approaches to fitness scaling in fitness-proportional selection are described in the following subsections. The subsequent section presents tournament selection, which has a different basis from fitness-proportional selection. Fitness scaling would have no effect on tournament selection, as it does not alter the rank ordering of fitnesses. Tournament selection is computationally cheap, as it requires neither the calculation of scaling parameters nor the application of a roulette wheel.