齐次函数模型极小化的Greville方法

AN UNCONSTRAINED MINIMIZATION METHOD BASED ON HOMOGENEOUS FUNCTION (Ⅰ)——GREVILLE'S METHOD

本文提出了根据齐次函数模型的无约束极小化方法——Greville方法,它是Jacobson-Oksman方法的推广。由于它利用了广义逆技术,故方法的适应性较强,作者证明了对于齐次函数,Greville方法在n+2步得到广义极小点。文章还给出了算法和数值结果。

An unconstrained minimization algorithm based on the homogenous function model is presented. This algorithm using a generalized inverse technique is a generalized Jaeobson-Oksman method. It is shown that Greville's method gains a generalized minimum in n+2 iterations for homogeneous functions. This paper gives algorithm's steps and its numerical results.