A matrix inequality with weights and its applications (Q2365727)

Let \((a_{ij})\) be a positive semidefinite symmetric \(n \times n\) matrix of rank \(n-1\) and \(A_ i\) the complementary minor of \(a_{ii}\), \(1 \leq i \leq n\). The authors prove the following inequality: if \(x_ 1,\dots,x_ n\) are positive real numbers, then \((\sum^ n_{i=1} a_{ii}x_ i^{-1})^{n-1}(\sum^ n_{i=1} x_ iA_ i)^{-1} \geq (n-1)^{n-1}(\prod^ n_{i=1} x_ i)^{-1}\), with equality if and only if all the nonzero eigenvalues of \(\text{diag}(x_ 1\dots x_ n)^{-1}A\) are equal, and derive some inequalities for the vertex angles of an \(n\)-dimensional simplex.