关键词: 遗传算法, 正交试验, 寻优性能, 影响因素

Abstract: Genetic algorithm is widely applied to function optimization. It’s very important to  select the appropriate parameters for the performance of function optimization. This paper takes four classical functions as examples, and analyzes five parameters’ impacts on genetic algorithm performance on function optimization based on orthogonal test. The results show that：the greatest influencing factor on searching time is mutation probability, the second is population size, the other three crossover operator, crossover probability and encoding length are less; the three factors that influence iteration times are mutation probability, population size and encoding length, and the influence of crossover probability and crossover operator are less. This paper analyzes the principles of parameters combination for genetic algorithm optimal performance.

Key words: genetic algorithm, orthogonal experiment, optimization performance, influencing factors