A Kind of Hybrid Genetic Algorithm Based on Numerical Function Optimization
This paper introduces some new crossover and mutation probability in generic algorithm related to the evolution times. We also improve the Powell direct method to reduce its time requirement of Onedimensional search. We developed a hybrid genetic algorithm for global optimization of numerical function according the combination of the genetic algorithm and Powell. This algorithm also includes the process of isolated point by the best-preserved strategy. Compared with the simple genetic algorithm, the hybrid genetic algorithm improved the performance of local search, effectively solved the prematurity problem and signifi cantly enhanced the probability to capture the global solution. Moreover, because the algorithms only use function information, its a common genetic algorithm in function optimization and a effective algorithm in engineering calculation.
Jie Xiao Jiarong Liang Weizhe Long Yin Li Shuang Xu
School of Computer and Electronics and Informatin of Guangxi Univeristy Nanning, Guangxi 530004,P.R. China
国际会议
南宁
英文
2007-07-20（万方平台首次上网日期，不代表论文的发表时间）