A Novel Simplex Hybrid Genetic Algorithm
Hongfeng Xiao, Guanzheng Tan · 2008
In simplex hybrid genetic algorithm, widely using Nelder-Mead simplex method (NMSM) would lead to precocity of genetic algorithm and increase in computation quantity, so a novel simplex hybrid genetic algorithm is proposed in this paper. First, we propose a new efficient simplex crossover operator. Second, using the successful experiences of dividing the vertexes in NMSM into the best vertex, the worst vertex and worse vertex, the population in genetic algorithm also is divided into three niches: the best niches, the worst niche and worse niches, into which NMSM, simplex multiple-direction search and the efficient simplex crossover operator are embedded respectively. By these means, the main idea of "classifying niche and then treating them by their categories" is implemented and then the novel simplex hybrid genetic algorithm (Simplex-NHGA) is built up. At last, five standard test functions are used to verify the correctness and efficiency of simplex-NHGA.