论文部分内容阅读
为了有效找到全局最优，将“进化方向”引入到遗传算法中被认为是非常有用的．文中提出了一个新的进化方向算子，用于对标准遗传算法进行改进．该算子依据父代个体和当代个体的进化方向产生子代个体．进化方向的求取无须计算梯度．算子的作用对象是个体中的变量而不是染色体中的基因，因而适合于采用任何形式编码的遣传算法场合．包含该方向进化算子的遗传算法被用于二维多峰函数的极大值搜索问题，进行了对比验证．将新算法进一步应用于离心压缩机扩压器叶片逆命题设计问题，数值结果显示，改进遗传算法比标准遗传算法有好得多的收敛特性
In order to effectively find the global optimum, it is considered to be very useful to introduce the “evolutionary direction” into the genetic algorithm. In this paper, a new evolution direction operator is proposed to improve the standard genetic algorithm. The operator generates offspring individuals according to the evolution direction of the father individual and the contemporary individual. Evolution of the direction of the need to calculate the gradient. The role of the operator is an individual variable rather than a gene in the chromosome, and thus suitable for use in any form of coding deportation algorithm. The genetic algorithm containing this direction evolution operator is used to search the maximum value of two-dimensional multimodal functions, and the comparison is validated. The new algorithm is further applied to the inverse design of centrifugal compressor diffuser vanes. Numerical results show that the improved genetic algorithm has much better convergence than the standard genetic algorithm