| 一种基于超群体的并行遗传算法 |
| |
| 引用本文： | 蒙祖强,蔡自兴. 一种基于超群体的并行遗传算法[J]. 计算机工程与应用, 2001, 37(21): 28-30,33 |
| |
| 作者姓名： | 蒙祖强 蔡自兴 |
| |
| 作者单位： | 中南大学信息科学与工程学院 |
| |
| 基金项目： | 国家自然科学基金资助(编号:69974043)，国家博士点基金(编号:99053317)，湖南省自然科学基金(编号:99JJY20062)资助 |
| |
| 摘 要： | 文章首次提出了空间交配的慨念,构造了一种基于超群体的并行遗传算法。它把每一子群体(sub-group)看作一个特殊的个体,称为超个体(super-individual);该算法就是对由若干超个体组成的群体———超群体(super-group)施加遗传运算,从而实现遗传算法的并行化。它不但较好地克服了早熟问题,而且开拓遗传算法研究的新方向。最后,给出了实验的对比分析,证实了算法的有效性。
|
| 关 键 词： | 并行遗传算法 空间交配 早熟收敛 孤岛模型 |
| 文章编号： | 1002-8331-(2001)21-0028-03 |
A Parallel Genetic Algorithm Based on Super-group |
| |
| Meng Zuqiang Cai Zixing. A Parallel Genetic Algorithm Based on Super-group[J]. Computer Engineering and Applications, 2001, 37(21): 28-30,33 |
| |
| Authors: | Meng Zuqiang Cai Zixing |
| |
| Abstract: | This paper is the first to bring out the concept of Spatial Crossover and devises a parallel genetic algorithm based on super-group.And each sub-group is regard as a special individual,named super-individual.The genetic opera-tions of the super-group which consists of some super-individuals,result in parallelization of genetic algorithm.The algo-rithm not only performs well in overcoming the problem of premature convergence,but also carves out a new method of researching genetic algorithm.Experiments are also comparatively analyzed in this paper,which confirm its validity. |
| |
| Keywords: | Parallel Genetic Algorithms Spatial Crossover Premature convergence Isolate model |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
