论文部分内容阅读
遗传算法是一类借鉴生物进化规律的随机化搜索算法,实质为一类程序框架,其参数设定后转化为算法实例,用于解决实际问题.算法参数的设置方式可分为静态参数设置和动态参数设置.静态参数设置是在算法实例运行之前已经选定且在算法实例运行过程中保持不变;动态参数设置是在算法实例运行过程中动态调整参数.遗传算法框架的最优参数组合会随着算法实例运行而动态改变.所以,研究参数动态设置具有理论和现实意义.本文提出通过增减性选择、典型形态选择、分段化处理三步确定参数“优化基准函数”的遗传算法参数动态选择方案;并提出了函数的分段化处理方法.在实验部分,首先验证了参数动态优化较之参数静态选择的优势,继而通过选择高斯变异算子标准差的“优化基准函数”进行试验.
Genetic algorithm is a kind of randomized search algorithm that draws lessons from the rule of biological evolution, which is essentially a kind of program framework, whose parameters are set and converted into algorithm instances for solving practical problems.The setting of algorithm parameters can be divided into static parameter setting and Dynamic parameter settings Static parameter settings are selected prior to the running of the algorithm instance and remain unchanged during the running of the algorithm instance Dynamic parameters are set to dynamically adjust the parameters during the running of the algorithm instance The optimal parameters of the genetic algorithm framework It will change dynamically with the operation of the algorithm instance.Therefore, it is of theoretical and practical significance to study the dynamic setting of the parameters.This paper proposes to determine the genetic parameters of the parameter “optimized reference function” through three steps: addition and subtraction, typical morphological selection and segmentation processing The method of dynamic selection of the parameters of the algorithm is proposed.At the same time, the segmentation method of the function is proposed.In the experimental part, the superiority of the dynamic selection of the parameters is verified firstly, and then the optimal benchmark function of the standard deviation of the Gaussian mutation operator "experimenting.