一类基于共轭梯度法的下降算法

A class of descent methods based on conjugate gradient method

共轭梯度法是求解大规模无约束优化问题的有效算法之一 ,近年来出现了很多共轭梯度法收敛性的相关文献。本文研究基于共轭梯度法的下降算法 ,证明了算法的收敛性 ,并对算法进行了数值试验 。

The conjugate gradient method is one of effective methods for soluing large scale uvconstrained optimization problem.Mang developments about this method appeared in recent literature.This paper presents a dass of descent method based on conjugate gradient mdthod,proves its global convergence and reports its numerical performance.The numerical esperiments show that the method in this paper is efficient.