| 摘要: |
| 根据PRP和HS公式具有相同分子只是分母不同的相似性,通过适当的结合和构造,给出一个新的共轭梯度公式.证明该公式的新方法在强Wolfe-Powell线搜索下具有充分下降性,在适当的假设和弱Wolfe-Powell线搜索下具有全局收敛性,并用数值试验证实新方法是有效的. |
| 关键词: 无约束优化 共轭梯度法 Wolfe-Powell条件 全局收敛 |
| DOI： |
| 投稿时间：2006-12-12修订日期：2007-03-26 |
| 基金项目:Supported by the Natural Science Foundation of China (10161002);the Natural Science Foundation of Guangxi Province (0542043);the Youth Foundation of Guangxi Normal University |
|
| A Conjugate Gradient Formula Generated by PRP and HS Formu |
|
CHEN Cui-ling1, ZHAO Yan2
|
| (1.College of Mathematics Science, Guangxi Normal University, Guilin, Guangxi, 541004, China;2.School of Management, University of Shanghai for Science and Technology, Shanghai, 200093, China) |
| Abstract: |
| With the similarity of the form between PRP and HS formulas which have the same numerator and different denominators, and proper combining and composing, a new conjugate gradient formula is obtained.The present method based on this formula possesses a sufficient descent property with the strong Wolfe-Powell line search.Under some suitable assumptions and the weak Wolfe-Powell line search, the global convergence result is established.The preliminary numerical results show that the proposed method is efficient. |
| Key words: unconstrained optimization conjugate gradient method Wolfe-Powell conditions global convergence |
