| Global annealing genetic algorithm and its convergence analysis |
| |
| 作者姓名： | 张讲社 徐宗本 梁怡 |
| |
| 作者单位： | Research Center for Applied Mathematics and Institute for Information and Systems Sciences，Faculty of Science，Xi'an Jiaotong University，Xi'an 710049，China，Research Center for Applied Mathematics and Institute for Information and Systems Sciences，Faculty of Science，Xi'an Jiaotong University，Xi'an 710049，China，The Centre for Environmental Studies，The Chinese University of Hong Kong，Shatin，N.T.，Hong Kong，China |
| |
| 基金项目： | Project supported by the Hi-Tech Project of China and the National Natural Science Foundation of China |
| |
| 摘 要： | A new selection mechanism termed global annealing selection (GAnS) is proposed for the genetic algorithm. It is proved that the GAnS genetic algorithm converges to the global optimums if and only if the parents are allowed to compete for reproduction, and that the variance of population's fitness can be used as a natural stopping criterion. Numerical simulations show that the new algorithm has stronger ability to escape from local maximum and converges more rapidly than canonical genetic algorithm.
|
Global annealing genetic algorithm and its convergence analysis |
| |
| Jiangshe Zhang,Zongben Xu,Yee Leung.Global annealing genetic algorithm and its convergence analysis[J].Science in China(Technological Sciences),1997,40(4):414-424. |
| |
| Authors: | Jiangshe Zhang Zongben Xu Yee Leung |
| |
| Affiliation: | 1. Research Center for Applied Mathematics and Institute for Information and Systems Sciences, Faculty of Science, Xi'an Jiaotong University, 710049, Xi'an, China
2. The Centre for Environmental Studies, The Chinese University of Hong Kong, Shatin, N. T., Hong Kong, China
|
| |
| Abstract: | A new selection mechanism termed global annealing selection (GAnS) is proposed for the genetic algorithm. It is proved that the GAnS genetic algorithm converges to the global optimums if and only if the parents are allowed to compete for reproduction, and that the variance of population's fitness can be used as a natural stopping criterion. Numerical simulations show that the new algorithm has stronger ability to escape from local maximum and converges more rapidly than canonical genetic algorithm. |
| |
| Keywords: | genetic algorithm simulated evolutionary computation computational intelligence annealing selection Markov chain |
| 本文献已被 CNKI SpringerLink 等数据库收录！ |
