文章导读
总览 评价 王培 , 荆燕飞 * ( 电子科技大学数学科学学院,成都611731; ) 摘要: 针对右端向量包含有错误数据的不适定问题离散化后的线性系统,本文研究了求解正规方程的扩充共轭梯度方法(CGLS)。进而通过同时扩充对应右端向量对应的Krylov 子空间和用户给
王培, 荆燕飞*
(
电子科技大学数学科学学院,成都611731; )
摘要:
针对右端向量包含有错误数据的不适定问题离散化后的线性系统,本文研究了求解正规方程的扩充共轭梯度方法(CGLS)。进而通过同时扩充对应右端向量对应的Krylov 子空间和用户给定的一个子空间的基向量组,得到了一种混合扩充CGLS方法。数值试验表明所提方法与通常使用的CGLS-型方法相比,可以得到具有更高精度的近似解。
关键词:
计算数学;共轭梯度方法;不适定问题;扩充子空间方法。
WANG Pei, JING Yan-Fei*
(
School of Mathematical Sciences/Institute of Computational Science, University of Electronic Science and Technology of China, Chengdu 611731; )
Abstract:
A comparative study of augmented conjugate gradient method for normal equations (CGLS) is carried out for solving linear systems arising from the discretization of ill-posed problems with error-contaminated data represented by the right-hand side.As a result, a new hybrid augmented CGLS method is presented by enriching the Krylov subspace simultaneously with the right-hand side and a basis of a user-supplied subspace. Numerical experiments demonstrate our proposed method performs well to generate approximate solutions of higher accuracy in comparison with commonly used CGLS-type methods.
Tag:
点此返回栏目查看更多>>>参考论文