求一类多维函数总极值点的综合数值方法

AN SYNTHETICALLY NUMERICAL METHOD FOR SEARCHING GLOBAL MINIMUM POINT OF SOME KIND OF MULTDIMENSIONAL FUNCTIONS

本文综合三种较为有效的求总极值的确定型方法和随机型方法，提出自动寻找好的初始迭代点以较为方便地获取一类多维函数的总极值点的数值方法。这种方法只需在求局部极值算法程序中加入一个初值点选择模块就可获得总极值点求解程序。多个算例表明，该方法对一类多维函数的总极值点求解是很有效的。

This paper presents a new numerical method for searching global minimum of some kind of multidimensional functions by synthesizirig three determinate and random methods.The synthesis produces better initial iteration point automatically,and this makes the searching of global minimum point more effectively and conveniently.The program for searching global minimum point can be made by adding an initial iteration point selection module to the program of searching local minimum point.Calculation examples show that this method is very effectively for searching the global minimum point of some kind of multidimensional functions.