一种区域转移实数遗传算法

Field Division Real-Coded Genetic Algorithm

在讨论实数遗传算法基本遗传算子及参数选择的基础上 ,针对算法能快速找到较好解的特点及恰当的搜索范围能保证算法找到满意解的现象 ,将区域划分与转移思想应用到算法结构改进中 ,设计了一种改进的实数遗传算法。文中对新算法进行了细致描述 ,并与类似方法进行了比较。最后对多个实验函数进行寻优 ,优化实验结果证明 ,新方法在寻找复杂问题的全局解、提高搜索精度方面比基本实数遗传算法有较大的改进。文章最后对新算法的优缺点进行了总结。

The characteristics of basic real-coded genetic operators and the selection of parameters are discussed. A basic real-coded genetic algorithm has a fast searching speed in the early stage of optimization. It can easily find the optimum if domain area is properly selected, an improved real-coded genetic algorithm is put forward. The strategy of field partition and changing is used for the new algorithm, and the algorithm is compared with some existed similar algorithms. Numerical experiments show that the new algorithm can find the global solution of complex functions and can improve the precision of searching result.