标题:
一个三项共轭梯度算法及其收敛性A Three- Terms Conjugate Gradient Algorithm and Its
Convergence
作者:
陈海
关键字:
共轭梯度, 充分下降, 收敛性Conjugate Gradient; Sufficient Descent Property; Convergence.
期刊名称:
《Pure Mathematics》, Vol.1 No.1, 2011-05-05
摘要:
本文给出一个三项共轭梯度算法，搜索方向在不需要任何线搜索的条件下，拥有充分下降性条件，在此方向的定义中，不但拥有梯度值信息还拥有函数值信息，证明了全局收敛性并给出数值检验结果。
In this paper, a three- terms conjugate gradient method is proposed .The search direction possesses the sufficient descent property without any line search. Moreover, the search direction has not only the gradient value but also function value. The global convergence will be established and the numerical results are reported.