Inequalities involving unitarily invariant norms and operator monotone functions (Q5957190)

Let \(\|\cdot\|\) be a unitarily invariant norm on matrices. For matrices \(A,B,X\) with \(A,B\) positive semidefinite and \(X\) arbitrary, the authors prove that the function \(t\mapsto\||A^tXB^{1-t} |^r\|\cdot\||A^{1-t}X B^t|^r\|\) is convex on \([0,1]\) for each \(r>0\). This is a generation of the results of \textit{R. Bhatia} and \textit{C. Davis} [Linear Algebra Appl. 223/224, 119-129 (1995; Zbl 0824.47006)] and \textit{A. W. Marshall} and \textit{I. Olkin} [Pac. J. Math. 15, 241-247 (1965; Zbl 0166.30001)]. The authors prove that if \(f(t)\) is a nonnegative operator monotone function on \([0,\infty)\) and \(\|\cdot \|\) is a normalized unitarily invariant norm, then \(f(\|X\|) \leq\|f(|X|) \|\) for every matrix \(X\). The special case when \(f(t)=t^r (0<r\leq 1)\) is used to consider the monotonicity of \(p\mapsto\|A^p +B^p\|^{1/p}\) as well as \(p\mapsto\|( A^p+B^p)^{1/p}\|\). Furthermore, the authors obtain some norm inequalities of Hölder and Minkowski types related to the expression \(\||A|^p+ |B|^p \|^{1/p}\). The paper includes the following sections: Introduction; Convexity of certain functions involving unitarily invariant norms; A norm inequality for operator monotone functions with applications; Norm inequalities of Hölder and Minkowski types.