A CONTRIBUTION TO THE CONDITIONING OF THE MIXED LEAST-SQUARES SCALED TOTAL LEAST-SQUARES PROBLEM

被引：0
|
作者：
Meng, Lingsheng [1 ]
机构：
[1] Northwest Normal Univ, Coll Math & Stat, Lanzhou 730070, Peoples R China
来源：
ELECTRONIC TRANSACTIONS ON NUMERICAL ANALYSIS | 2022年 / 55卷
关键词：
mixed least-squares scaled total least-squares; condition numbers; mixed least-squares total least-squares;
D O I：
10.1553/etna_vol55s618
中图分类号：
O29 [应用数学];
学科分类号：
070104 ;
摘要：
A new closed formula for first-order perturbation estimates for the solution of the mixed least-squares scaled total least-squares (MLSSTLS) problem is presented. It is mathematically equivalent to the one by [Zhang and Wang, Numer. Algorithms, 89 (2022), pp. 1223-1246]. With this formula, new closed formulas for the relative normwise, mixed, and componentwise condition numbers of the MLSSTLS problem are derived. Compact forms and upper bounds for the relative normwise condition number are also given in order to obtain economic storage and efficient computations.
引用
收藏
页码：618 / 626
页数：9
相关论文
共 50 条
12345