| 一种基于复合交叉的实数编码遗传算法 |
| |
| 引用本文： | 李勇,曹广益,朱新坚. 一种基于复合交叉的实数编码遗传算法[J]. 计算机仿真, 2006, 23(6): 166-170 |
| |
| 作者姓名： | 李勇 曹广益 朱新坚 |
| |
| 作者单位： | 上海交通大学自动化系燃料电池研究所,上海,200030;上海交通大学自动化系燃料电池研究所,上海,200030;上海交通大学自动化系燃料电池研究所,上海,200030 |
| |
| 摘 要： | 提出了一种基于复合交叉的实数编码遗传算法。通过对父代染色体间的区域进行多次交叉操作,复合交叉操作保留了被搜索区域的信息。算法首先对父代染色体进行复合交叉操作,然后利用包含在新个体集合中的信息进行信息最大化选择,对每一代个体进行基于适应度的选择。集合中信息冗余的个体被从种群中删除,位于欠搜索区域中的个体被保留。由于算法能够始终保持种群的多样性,算法不仅能搜索到全局最优点,同时也能找到尽量多的局部极值点。利用算法对多极值函数进行了寻优,仿真结果表明了算法的有效性。
|
| 关 键 词： | 复合交叉 信息最大化选择 种群多样性 实数编码遗传算法 |
| 文章编号： | 1006-9348(2006)06-0166-05 |
| 收稿时间： | 2005-05-23 |
| 修稿时间： | 2005-05-23 |
A Multiple Crossover Real-coded Genetic Algorithm |
| |
| LI Yong,CAO Guang-yi,ZHU Xin-jian. A Multiple Crossover Real-coded Genetic Algorithm[J]. Computer Simulation, 2006, 23(6): 166-170 |
| |
| Authors: | LI Yong CAO Guang-yi ZHU Xin-jian |
| |
| Affiliation: | Department of Automation, Fuel Cell Institute, Shanghai Jiaotong University, Shanghai 200030,China |
| |
| Abstract: | A new genetic algorithm based on multiple crossover(MCGA) is presented, which makes crossover operation several times on the same pair of parent individuals.The individuals generated by every multiple crossover are immediately selected according to a new usefulness-based selection rule,then the chromosomes of a generation is selected by a fitness-based selection rule.The chromosomes in an over-searched region of state space are deleted and the chromosomes in a region that is not searched are kept.This algorithm limits the reproduction of redundant individuals,and the diversity of population is perfectly maintained.The global and several local optimal solutions can be kept in the final generation.This algorithm is used for optimizing the multi-model functions and the results illustrate that this algorithm is very effective. |
| |
| Keywords: | Multiple crossover Usefulness-based selection Population diversity Real-coded genetic algorithm |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
