摘要
在LS方法基础上,提出了一种新的求解无约束最优化问题的共轭梯度法.新方法通过一个新的公式计算参数,克服了LS方法的数值效果不稳定和收敛性弱的缺点,并且在强Wolfe线搜索下证明了该方法具有充分下降性和全局收敛性.大量的数值试验表明新方法是稳定的、有效的.
A new conjugate gradient mehod is proposed to solve unconstrained optimization problems on the basis of LS method.We adopt a new formula for calculating parameter in the new mehod which can overcome the effects of the LS method in numerical instability and weak convergence of the deficiencies.And under the strong Wolfe line search,the sufficient descent property and the global convergence of the new mehod was proved.A large number of numerical experiments show that the new mehod is stable and effective.
出处
《数值计算与计算机应用》
CSCD
北大核心
2009年第4期247-254,共8页
Journal on Numerical Methods and Computer Applications
关键词
无约束优化
共轭梯度法
强Wolfe线搜索
充分下降性
全局收敛性
Unconstrained optimization
Conjugate gradient method
Strong Wolfe line search
Sufficient descent property
Global convergence
