摘要
提出了一种新的无约束优化下降算法 .在每步迭代中算法以当前点负梯度和前一点负梯度的线性组合为搜索方向 ,用Armijo搜索定义步长 .
A new gradient descent algorithm for unconstrained optimization problem is proposed. In each iteration, the linear combination of negative gradient and its previous gradient are used as a search direction, and stepsize is defined by Armijo's line search. The convergence of the algorithm is proved under some mild conditions.
出处
《曲阜师范大学学报(自然科学版)》
CAS
2002年第4期13-16,共4页
Journal of Qufu Normal University(Natural Science)
基金
TheworkissupportedbyNationalNaturalScienceFoundationofChina (10 1710 5 4)