摘要
提出了一种改进的全局优化进化算法.该算法采用实数编码,通过对可行域量子化用正交设计产生初始种群,用正交设计和因素分析设计杂交算子.在进行杂交之前,根据两个个体变量之间的距离恰当地应用高斯变异,平衡了算法的局部搜索能力和全局搜索能力,从而提高了算法的效率.最后的数值结果显示了该算法的有效性.
An improved evolutionary algorithm for global optimization was proposed in this paper. The algorithm is realeoded and initial population is generated using orthogonal design via quantizing the solution space. Orthogonal design together with factor analysis is applied to design crossover operator. Gaussian mutation is applied to the two individuals if necessary in order to balance the laeal search ability and the global search abihy before crossover, thus efficiency of algorithm is improved, which can be showed by the numerical results.
出处
《安阳师范学院学报》
2007年第2期12-15,共4页
Journal of Anyang Normal University
关键词
进化算法
高斯变异
局部搜索
全局搜索
Evolutionary algorithm
Gaussian mutation
Laeal search
Global search
