摘要
基于利用修正HS方法提高算法效率和利用DY方法保证算法的全局收敛性等思想,分别在不同条件下提出两种新的混合共轭梯度法求解大规模无约束优化问题.在一般Wlolfe线搜索下不需给定下降条件,证明了两个算法的全局收敛性,数值实验表明所提出算法的有效性,特别对于某些大规模无约束优化问题,数值表现较好.
Motivated by using the computational efficiency of the modified Hestenes-Stiefel (HS) method and the global convergence of Dai-Yuan (DY) method, two new mixed conjugate gradient methods are presented under the different conditions to solve the large scale unconstrained optimization. The global convergence of the proposed methods are proved under the Wolfe line search and without the descent condition. Numerical experiments show that the two algorithms are firmly efficient.
出处
《数值计算与计算机应用》
CSCD
2012年第2期92-98,共7页
Journal on Numerical Methods and Computer Applications
基金
国家自然科学基金(11171003)
教育部科学技术重点项目(No.211039)
吉林省自然科学基金(No.201215102)资助
关键词
无约束最优化
共轭梯度法
WOLFE线搜索
全局收敛性
Unconstrained optimization
Conjugate gradient algorithm
Wolfe line search
Global convergence