J4 ›› 2013, Vol. 51 ›› Issue (01): 34-40.

• 数学 • 上一篇    下一篇

两类无约束优化的充分下降共轭梯度法

孙中波1, 段复建2, 高海音3, 于海鸥1   

  1. 1. 东北师范大学人文学院 数学系, 长春 130117;2. 桂林电子科技大学 数学与计算科学学院, 广西 桂林 541004|3. 长春大学 理学院, 长春 130022
  • 收稿日期:2012-03-14 出版日期:2013-01-26 发布日期:2013-01-31
  • 通讯作者: 孙中波 E-mail:szb21971@yahoo.com.cn

Two Sufficient Descent Conjugate Gradient Methodsfor Unconstrained Optimization

SUN Zhongbo1, DUAN Fujian2, GAO Haiyin3, YU Haiou1   

  1. 1. Department of Mathematics, College of Humanities and Sciences of Northeast Normal University,Changchun 130117, China|2. College of Mathematics and Computational Science, Guilin University of Electronic Technology, Guilin 541004, Guangxi Zhuang Autonomous Region,China|3. College of Science, Changchun University, Changchun 130022, China
  • Received:2012-03-14 Online:2013-01-26 Published:2013-01-31
  • Contact: SUN Zhongbo E-mail:szb21971@yahoo.com.cn

摘要:

对无约束优化问题提出两类新的充分下降共轭梯度法. 在每次迭代过程中, 算法均可得到充分下降方向. 在适当条件下, 证明了算法的全局收敛性. 数值结果表明算法可行、 有效.

关键词: 共轭梯度法, 全局收敛, 无约束优化

Abstract:

We proposed two sufficient descent conjugate gradient methods for unconstrained optimization. The sufficient descent direction is always obtained at each iteration. Under some suitable conditions, the global convergence can be induced. Numerical results show that these methods are feasible and effective.

Key words: conjugate gradient method, global convergence, unconstrained optimization

中图分类号: 

  • O224.2