A characterization of the mixed discriminant (Q2790933)

Let \({\mathcal M}^n\) denote the set of real symmetric positive semidefinite \(n\times n\) matrices. The mixed discriminant \(D:({\mathcal M}^n)^n\to\mathbb{R}\) is the unique symmetric function satisfying NEWLINE\[NEWLINE \det{(\lambda_1A_1+\dots+\lambda_mA_m)}= \sum_{i_1,\dots,i_n=1}^m\lambda_{i_1}\cdots\lambda_{i_n}D(A_{i_1}\cdots A_{i_n}) NEWLINE\]NEWLINE for all \(m\in\mathbb{N}\), \(A_1,\dots,A_m\in{\mathcal{M}}^n\), \(\lambda_1,\dots,\lambda_m\geq 0\). Let \(F:({\mathcal M}^n)^n\to\mathbb{R}\) be a nonnegative function. The authors prove that if \(F\) is additive in each variable and if it is zero when two variables are proportional matrices of rank one, then there exists a constant \(a\geq 0\) such that NEWLINE\[NEWLINE F(A_1,\dots,A_n)=aD(A_1,\dots,A_n) NEWLINE\]NEWLINE for all \(A_1,\dots,A_n\in{\mathcal M}^n\). \textit{V. D. Milman} and \textit{R. Schneider} [Adv. Geom. 11, No. 4, 669--689 (2011; Zbl 1242.52011)] presented an analogous characterization for the mixed volume of centrally symmetric convex bodies in \(\mathbb{R}^n\).