摘要
基于Hager-Zhang提出的共轭梯度法,构造了一种新的谱βk,证明了该方法不依赖于任何线搜索就具有充分下降性,并且在Armijo搜索下证明了算法的全局收敛性。数值试验表明,该方法明显优于谱DY、谱FR、谱PRP算法。
Based on Hager-Zhang＇s conjugate gradient formula has sufficient descent without relying on any line search. And , a new spectral is proposed to prove that this method the global convergence of corresponding algorithm by Armijo line search was validated. Numerical results show that this method is more efficient and suitable than DY spectrum algorithm, FR spectrum algorithm and PRP spectrum algorithm.
出处
《太原科技大学学报》
2011年第2期153-156,共4页
Journal of Taiyuan University of Science and Technology
基金
山西省自然科学基金(2008011013)
关键词
无约束优化
谱共轭梯度法
ARMIJO搜索
全局收敛性
unconstrained optimization, spectral conjugate gradient method, Armijo line search, global convergence
