一类共轭梯度法的全局收敛性

J4 ›› 2010, Vol. 45 ›› Issue (5): 101-105.

一类共轭梯度法的全局收敛性

程李晴

新乡医学院, 河南新乡 453003

收稿日期:2008-12-27 出版日期:2010-05-16 发布日期:2010-05-24
作者简介:程李晴（1971-），女，讲师，硕士，研究方向为最优化理论和方法. Email:chengliqing@xxmu.edu.cn

The global convergence of a class of conjugate gradient methods

CHENG Li-qing

Xinxiang Medical University, Xinxiang 453003, Henan, China

Received:2008-12-27 Online:2010-05-16 Published:2010-05-24

摘要/Abstract

摘要：

将具有某种性质的一类共轭梯度法与一种Armijo型线搜索方法相结合,得到了一类新的共轭梯度算法,证明了这类新算法是全局收敛的,PRP方法为其一个特例。数值试验表明,新算法是有效的。

关键词: 无约束优化;PRP共轭梯度法; Armijo型线搜索;全局收敛

Abstract:

Conjugate gradient methods with a certain kind of Property under an Armijo-type line search were investigated. These methods include the well-known PRP method as a special case, and the global convergence of the methods is proved. Preliminary numerical results show that these methods are efficient.

Key words: unconstrained optimization;the Polak-Ribière-Polyak conjugate gradient method;Armijo-type line search;global convergence

程李晴. 一类共轭梯度法的全局收敛性[J]. J4, 2010, 45(5): 101-105.

CHENG Li-qing. The global convergence of a class of conjugate gradient methods[J]. J4, 2010, 45(5): 101-105.

参考文献

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed