Preconditioning Linear Least-Squares Problems by Identifying a Basis Matrix

Preconditioning Linear Least-Squares Problems by Identifying a Basis Matrix
复制标题

通过识别基矩阵来预处理线性最小二乘问题

DOI:
10.1137/140975358
复制
发表时间:
2015
影响因子:
3.1
通讯作者:
Arioli M
Arioli M
中科院分区:
数学2区
文献类型:
--
作者:
Arioli M

文献摘要

参考文献

被引文献

相似文献

本文研究了线性最小二乘问题的解,其中矩阵()具有秩,矩阵()是大而稀疏的。我们假设它是一个矩阵,而不是一个运算符。这个问题的预处理是困难的,因为矩阵不具有使标准预处理有效的微分问题的性质。不完整的Cholesky技术应用于正规方程不产生一个良好的条件问题。我们试图绕过病态找到一个非奇异子矩阵,减少欧几里德范数。我们用来对一个对称拟定线性方程组进行预处理,使其条件数与条件数无关,并且与原最小二乘问题有相同的解。我们说明了我们的方法在一些标准的测试问题上的性能,并表明它与其他方法是有竞争力的。
We study the solution of the linear least-squares problemwhere the matrix() has rankand is large and sparse. We assume thatis available as a matrix, not an operator. The preconditioning of this problem is difficult because the matrixdoes not have the properties of differential problems that make standard preconditioners effective. Incomplete Cholesky techniques applied to the normal equations do not produce a well-conditioned problem. We attempt to bypass the ill-conditioning by finding annonsingular submatrixofthat reduces the Euclidean norm of. We useto precondition a symmetric quasi-definite linear system whose condition number is then independent of the condition number ofand has the same solution as the original least-squares problem. We illustrate the performance of our approach on some standard test problems and show it is competitive with other approaches.
DOI: --
发表时间: 1985
期刊:
影响因子: --
作者:
D. E. Knuth
通讯作者: D. E. Knuth
DOI: --
发表时间: 2000
影响因子: 1.5
作者:
A. George;K. Ikramov;A. B. Kucherov
通讯作者: A. B. Kucherov
如何让Lanczos算法收敛缓慢
DOI: --
发表时间: 1979
期刊:
影响因子: --
作者:
D. Scott
通讯作者: D. Scott
DOI: --
发表时间: 1990
期刊: Conference on Algorithms and Hardware for Parallel Processing
影响因子: --
作者:
M. Hegland
通讯作者: M. Hegland