| 摘要: |
| 在文献[4,5]的基础上,提出求解无约束优化问题的共轭梯度公式中βk参数的一种新的计算公式:βk+1=μ(‖gk+1‖2/dkTyk),0 < μ < 1;对标准Wolf搜索条件进行推广,得到一种新的算法,并证明了算法的全局收敛性. |
| 关键词: 无约束优化 共轭梯度法 线搜索 全局收敛性 |
| DOI： |
| 投稿时间：2005-12-26修订日期：2006-06-15 |
| 基金项目:广西自然科学基金项目(0448075)资助 |
|
| A New Conjugate Gradient Method with a Global Convergent Property |
|
TIAN Ya-juan, MA Chang-feng
|
| (Deptartment of Computing Science and Mathematics, Guilin University of Electronic Technology, Guilin, Guangxi, 541004, China) |
| Abstract: |
| We propose a new formula about βk based on the paper[4] and[5],that is βk+1=μ(‖gk+1‖2/dkTyk),0 < μ < 1.This paper presents a wider line search than the standard Wolf search and give a new conjugate gradient method with global convergence. |
| Key words: unconstrained optimization conjugate gradient method line search global convergence |
