Interpolating the Rotfel'd inequality for unitarily invariant norms (Q2419047)

Let \(\mathbb{M}_n\) be the \(C^*\)-algebra of all \(n\times n\) complex matrices. For Hermitian matrices \(A, B\in \mathbb{M}_n\), we write \(A\geq 0\) if \(A\) is positive semidefinite, \(A>0\) if \(A\) is positive definite, and \(A\geq B\) if \(A-B\geq0\). A norm \(\|\,.\,\|\) is called unitarily invariant norm if \(\|UAV\|=\|A\|\) for all \(A\in\mathbb{M}_n\) and all unitary matrices \(U, V\in\mathbb{M}_n\). The Rotfel'd inequality shows that for \(A, B\in\mathbb{M}_n\) and all non-negative concave functions \(f(t)\) on \([0, \infty)\), there holds \[\|f(|A + B|)\| \leq \|f(|A|)\| + \|f(|B|)\|,\] where \(\|\,.\,\|\) is the spectral norm. Under the same assumptions, \textit{M. Uchiyama} [Proc. Am. Math. Soc. 134, No. 5, 1405--1412 (2006; Zbl 1089.47010)] extended this inequality to every unitarily invariant norm. Furthermore, \textit{J.-C. Bourin} and \textit{M. Uchiyama} [Linear Algebra Appl. 423, No. 2--3, 512--518 (2007; Zbl 1123.15013)] generalized the right side of this inequality to \(\|f(A)+f(B)\|\) for all positive matrices \(A, B\in\mathbb{M}_n\) and all unitarily invariant norm. In the present paper, the author shows some refinement of the Rotfel'd inequality and some related results by using the singular value decomposition of a matrix. In particular, it is shown that for \(A, B\in\mathbb{M}_n\) and for all non-negative concave functions \(f(t)\) on \([0, \infty)\), there holds \[ \|f(|A + B|)\| \leq\left\|f\left(\frac{1}{2}\begin{pmatrix} |A|+|B| & A^*+B^* \\ A+B & |A^*|+|B^*| \end{pmatrix}\right)\right \|\leq\|f(|A|)\| + \|f(|B|)\|, \] where \(\|\,.\,\|\) is a unitarily invariant norm.