Cauchy-Schwarz inequalities associated with positive semidefinite matrices (Q750564): Difference between revisions

The authors describe a very interesting method for proving inequalities of Cauchy-Schwarz type for matrices and use it to derive various inequalities for \(l_ p\)-norms, some of which are new. We state one of the main results: Suppose that the block matrix: \(A=\left( \begin{matrix} L\\ X^*\end{matrix} \begin{matrix} X\\ M\end{matrix} \right)\) is positive semidefinite \((L\in M_ m\), \(M\in M_ n\) and \(X\in M_{m,n})\). Then L and M are positive semidefinite and for all \(p>0\) and every unitarily invariant norm \(\| \cdot \|\), i.e., a norm which satisfies the condition: \(\| UAV\| =\| A\|\) for all \(A\in M_{m,n}\) and all unitary \(U\in M_ m\) and \(V\in M_ n\), one has the inequality: \(\| | X|^ p\|^ 2\leq \| L^ p\| \| M^ p\|.\) Certain applications for \(l_ p\)-norms are also given.