Hadamard-type inequalities for \(k\)-positive matrices (Q2062819)

Let \(M_n({\mathbb R})\) be the set of \(n\times n\) matrices with real entries. For an \(n\times n\) symmetric matrix, \(A=(a_{ij})_{1\leq i,j\leq n}\in M_n({\mathbb R})\), \(\lambda(A)=(\lambda_1,\ldots,\lambda_n)\) denotes the set of its eigenvalues and \[ \mathrm{diag}(A)=\mathrm{diag}(a_{11},\ldots, a_{nn}). \] Let \(n\geq 2\) and \(1\leq k \leq n\). The \(k\)-th symmetric function of the \(n\) variables \(\lambda=(\lambda_1,\ldots,\lambda_n)\in{\mathbb R}^n\) is defined by \[ S_k(\lambda):=\sum_{1\leq i_1<\ldots<i_k\leq n} \lambda_{i_1}\ldots\lambda_{i_k}. \] The following notation will be used: \(S_k(A)=S_k(\lambda(A)).\) Let \(E_k(A)\) be the sum of the principal minors of size \(k\) of \(A\in M_n({\mathbb R})\). \textit{R. A. Horn} and \textit{C. R. Johnson} [Matrix analysis. 2nd ed. Cambridge: Cambridge University Press (2013; Zbl 1267.15001)] proved that \[ S_k(A) = E_k(A) .\tag{1} \] Consider \(A=(a_{ij})_{1\leq i,j\leq n}\in M_n({\mathbb R})\) positive definite. The Hadamard determinant inequality gives \[ S_n(\mathrm{diag}(A))=a_{11}\ldots a_{nn}\geq \det A=S_n(A). \] If \(1\leq k\leq n\) is fixed, each principal submatrix of size \(k\) of \(A\) is also positive definite, hence applying the Hadamard inequality to each principal submatrices of \(A\) and using (1), one concludes that \[ S_k(\mathrm{diag}(A))\geq S_k(A). \] In this paper, the author shows that the previous inequality, holds for a larger class of symmetric matrices called \(k\)-positive matrices. Some consequences are discussed.