Upper bounds for the infinity norm of the inverse of SDD and \(\mathcal S\)-SDD matrices (Q2370680): Difference between revisions

Let \(N=\{1,\dots, n\}\). A matrix \(A=[a_{ij}]\) is called strictly diagonally dominant (SDD) if \(| a_{ii}| > r_i(A)\) for all \(i \in N\), where \(r_i(A):=\sum_{j\in N\backslash\{i\}}| a_{ij}| \). Then the so-called Varah bound \([\min_{i\in N}(| a_{ii}| -r_i(A))]^{-1}\) is an upper bound for \(\| A^{-1}\| _\infty\). The similar notion of \({\mathcal S}\)-\(SDD\) was defined by \textit{Y. M. Gao} and \textit{X. H. Wang} [Linear Algebra Appl. 169, 257--268 (1992; Zbl 0757.15010)]. The author gives an easily calculable upper bound for the infinity norm of the inverse of an \({\mathcal S}\)-\(SDD\) matrix and improve the Varah bound for \(SDD\) matrices. The author also introduces an iterative procedure for obtaining tighter bounds and poses some open questions.