摘要
构造了一种混合共轭梯度法,并将其与Rosen投影梯度法相结合运用于求解线性等式约束优化问题.这种新的混合共轭梯度投影法有效改善了Rosen投影梯度法收敛性速度较慢的情况,并在Wolfe线搜索下具有全局收敛性.
In this paper, a class of hybrid conjugate gradient methods were proposed, and they were applied to solve the linear equality constrained optimization problem combined with the Rosen gradient projection method. This new hybrid conjugate gradient projection method was effectivly improved the convergence rate of the Rosen gradient projection method, and the global convergence of the algorithm was proved under the Wolfe line search conditions.
出处
《安徽大学学报(自然科学版)》
CAS
北大核心
2013年第4期10-13,共4页
Journal of Anhui University(Natural Science Edition)
基金
国家自然科学基金资助项目(10671057)
关键词
约束优化问题
共轭梯度法
Rosen投影梯度法
WOLFE线搜索
全局收敛性
constrained optimization problems
conjugate gradient method
Rosen gradient projection method
Wolfe line search
global convergence