Conjugate Gradient Methods with Sufficient Descent Condition for Large-scale Unconstrained Optimization

被引:0
|
作者
Ling, Mei Mei [1 ]
Leong, Wah June [1 ]
机构
[1] Univ Putra Malaysia, Fac Sci, Dept Math, Serdang, Selangor, Malaysia
关键词
sufficient descent condition; conjugate gradient method; large-scale unconstrained optimization; CONVERGENCE PROPERTIES; GLOBAL CONVERGENCE;
D O I
10.1063/1.4903647
中图分类号
O59 [应用物理学];
学科分类号
摘要
In this paper, we make a modification to the standard conjugate gradient method so that its search direction satisties the sufficient descent condition. We prove that the modified conjugate gradient method is globally convergent under Annijo line search. Numerical results show that the proposed conjugate gradient method is efficient compared to some of its standard counterparts for large-scale unconstrained optimization.
引用
收藏
页码:629 / 633
页数:5
相关论文
共 50 条