摘要
描述了一种在新Armijo线搜索下的Fletcher-Revees(FR)共轭梯度法,并分析了其收敛性,从理论上证明了借助新的Armijo线搜索,FR共轭梯度法不仅可保证在每步迭代中都容易找出步长,而且可保证全局收敛性.
Fletcher-Reeves（FR） conjugate gradient method with a new Armijo-type line search is described, and its convergence is analyzed. The results show that by means of the new Armijo-type line search, an adequate step size at each iteration step can be chosen easily by the Fetcher-Reeves conjugate gradient method, and the global convergence can be ensured.
出处
《汕头大学学报（自然科学版）》
2008年第2期15-21,共7页
Journal of Shantou University：Natural Science Edition
关键词
无约束优化
FR共轭梯度法
全局收敛
R-线性收敛
unconstrained optimization
FR conjugate gradient method
global convergence
R-linearly convergence
