A converse of a matrix inequality (Q5929766): Difference between revisions
From MaRDI portal
Created a new Item |
Added link to MaRDI item. |
||
links / mardi / name | links / mardi / name | ||
Revision as of 01:27, 30 January 2024
scientific article; zbMATH DE number 1586480
Language | Label | Description | Also known as |
---|---|---|---|
English | A converse of a matrix inequality |
scientific article; zbMATH DE number 1586480 |
Statements
A converse of a matrix inequality (English)
0 references
18 January 2002
0 references
Let \(X = (x_{ij})\) denote a nonnegative \(m \times n\)-matrix whose entries do not all vanish. Let \(s(X)\) denote the sum of all entries of \(X\) and let \(\|X \|\) be the Frobenius norm of \(X.\) Then the following inequality holds: \[ \frac{m \sum_{i=1}^m \left( \sum_{j=1}^n x_{ij} \right)^2 + n \sum_{j=1}^n \left( \sum_{i=1}^m x_{ij} \right)^2}{(s(X))^2 + mn \|X \|^2} \geq \frac{m+n}{mn + \text{ min}(m,n)}. \tag \(*\) \] This estimate is best possible; conditions for resp. a characterization of the case of equality are given explicitly. That the quotient on the left hand side of \((*)\) admits the upper bound 1 was established earlier by \textit{E. R. van Dam} [Linear Algebra Appl. 280, No. 2-3, 163-172 (1998; Zbl 0934.15021)].
0 references
matrix inequalities
0 references
Cauchy inequality
0 references
nonnegative matrices
0 references
Frobenius norm
0 references