摘要
本文提出一个函数极小化的最优单纯形方法,它是用一维最优点 X~*来代替当前单纯形中函数最高值顶点 X^h.
For function minimization,this paper poses an optimal simplex method which is the further improvement of the work of Nelder and Mead.The author's main idea lies in the replacement of X^h,the vertex of highest functional value in the existing simplex method with X~*,the one-dimensional optimal point.Here, (?) If f(X^L)≥f(X~*),then X^h is replaced by X~* and a new optimal simplex is re- sulted;otherwise,the distance between the vertices of original simplex and the vertex X^1 of lowest functional value will be reduced to a half.
出处
《华侨大学学报（自然科学版）》
CAS
1989年第2期105-109,共5页
Journal of Huaqiao University(Natural Science)
