求解数值微分的进化策略新算法  

A novel evolutionary strategy algorithm to solve numerical differentiation

作　　者：夏慧明[1] 郭德龙[2] 王志刚[1] 周永权[3] 

机构地区：[1]南京师范大学泰州学院数学科学与应用学院,江苏泰州225300 [2]黔南民族师范学院数学系,贵州都匀558000 [3]广西民族大学数学与计算机科学学院,南宁530006

出　　处：《哈尔滨商业大学学报（自然科学版）》2011年第4期627-629,共3页Journal of Harbin University of Commerce：Natural Sciences Edition

基　　金：国家自然科学基金(60461001);广西自然科学基金(0542048);广西研究生教育创新计划资助项目(2007106080701M18);广西民族大学重大项目资助课题;泰州市科技发展计划项目

摘　　要：数值导数的公式对开发求解常微分方程和偏微分方程边值问题的算法很重要.数值微分的例子通常采用已知的函数,这样数值近似值可以与精确解进行比较.主要是提出了一种求解数值微分的进化策略新算法,该算法在求解微分值时,表现出精度高、收敛速度快等优点.The formula of the numerical derivative is very important to design an algorithm on solution of the rim value problems of the ordinary differential equation and partial differential equation. The examples of the numerical differentiation are usually adopted by known functions, for this the approximate numerical value can be compared with the accurate value. In this paper, a new algorithm--evolutionary strategy algorithm was proposed for solving the function＇ s optimum differential value. This algorithm had some advantages about that the answers were precise and restrained fast when find the solution of the differential value.

关 键 词：步长 进化策略 数值微分 中心差分公式 

分 类 号：O175[理学—数学]