Genetic optimization of reliability design of machine element

Canonical genetic algorithms have the defects of prematurity and stagnation when applied in optimization problems. The causes resulting in such phenomena were analyzed and a class of improved genetic algorithm with niche implemented by crossover of similar individuals and ( μ+λ ) selection was proposed. According to the reliability design theory of machine components, the genetic optimization model of jack clutch was obtained. An optimization instance and some results calculated by improved genetic algorithm were presented. The results of emulations and application show that the improved genetic algorithm with the niche technique can achieve the reliable global convergence and stable convergent velocity almost without any additional calculation expense. [

Canonical genetic algorithms have the defects of prematurity and stagnation when applied in optimization problems. The causes resulting in such phenomena were analyzed and a class of improved genetic algorithm with niche implemented by crossover of similar individuals and (mu + lambda) selection was proposed. According to the reliability design theory of machine components, the genetic optimization model of jack clutch was obtained. An optimization instance and some results calculated by improved genetic algorithm were presented. The results of emulations and application show that the improved genetic algorithm with the niche technique can achieve the reliable global convergence and stable convergent velocity almost without any additional calculation expense.