摘要
为加快非线性单调方程组的运算效率,基于高效率线搜索方法和投影技术,构建了一个新型的无导数型三项共轭梯度投影算法.通过改进搜索方向,使得新算法在任何线搜索下都自动满足充分下降性条件和信赖域特性.在一定的假设下,新方法具有全局收敛性,初步数值试验结果表明,新算法比同类算法更加高效.
In our report,in order to enhance the computational speed of solving nonlinear monotone equation,based on the effective line search method and projection technique,a new three-term conjugate gradient projection algorithm without derivative was constructed. Because of improving search direction,the algorithm possesses sufficient descent condition and trust region trait under any line search. Under some reasonable assumptions,the new algorithm converges globally,and which is more efficient than some similar algorithms.
作者
李丹丹
王松华
Li Dandan;Wang Songhua(Department of Applied Mathematics,Guangzhou Huashang College,Guangzhou 511300,China;School of Mathematics and Statistics,Baise University,Baise 533000,China)
出处
《海南大学学报(自然科学版)》
CAS
2021年第2期125-131,共7页
Natural Science Journal of Hainan University
基金
广东普通高校重点科研项目(2019KZDXM042)
广西自然科学基金(2018GXNSFAA281259,2020GXNSFAA159069)
广东财经大学华商学院校内项目(2020HSDS15)。
关键词
无导数型
共轭梯度法
充分下降性
信赖域性质
全局收敛性
derivative-free
conjugate gradient
sufficient descent property
trust region trait
global convergence