Multiplicative Cauchy functional equation on symmetric cones (Q2354657)

In this interesting paper, the author considers the following logarithmic Cauchy functional equation \(f(\mathbf{x}) + f( w_{I} \cdot \mathbf{y} \cdot w_{I}^T ) = f( w_\mathbf{x} \cdot \mathbf{y} \cdot w_\mathbf{x}^T ) \) for all \(\mathbf{x}\) and \(\mathbf{y}\) in the cone \(\Omega_+\) of symmetric positive definite matrices, where \(w_\mathbf{x}\) is a measurable function such that \(\mathbf{x} = w_\mathbf{x} \cdot w_\mathbf{x}^T\) for any square symmetric matrix \(\mathbf{x}\), and \(I\) is the identity matrix. To determine the solution \(f : \Omega_+ \to \mathbb{R}\) of this functional equation without any regularity assumptions on \(f\) the author uses two types of \(w_\mathbf{x}\), namely (i) when \(w_\mathbf{x}= \sqrt{\mathbf{x}}\), where \(\sqrt{\mathbf{x}}\) is the unique symmetric positive definite square root of \(\mathbf{x}\) and (ii) \(w_\mathbf{x} = t_\mathbf{x}\), where \(t_\mathbf{x}\) is the lower triangular matrix in the Cholesky decomposition of \(\mathbf{x}\). The above functional equation can be generalized to symmetric cones setting. In this general setting, the main results of this paper are: Theorem 1. Let \(f : \Omega \to \mathbb{R}\) be a function satisfying \(f( \mathbf{x} ) + f(\mathbf{y} ) = f (\mathbb{P} ( \sqrt{\mathbf{x}} )\mathbf{y} )\) for any \(\mathbf{x}\) and \(\mathbf{y}\) in the symmetric cone \(\Omega\). Then there exists a logarithmic function \(L\) such that \(f(\mathbf{x}) = L ( \text{det} (\mathbf{x}) )\) for all \(\mathbf{x} \in \Omega\). Theorem 2. Let \(f : \Omega \to \mathbb{R}\) be a function satisfying \(f( \mathbf{x} ) + f(\mathbf{y} ) = f (t_{\mathbf{y}} \mathbf{x} )\) for any \(\mathbf{x}\) and \(\mathbf{y}\) in the cone \(\Omega\) of rank \(r\), \(t_{{}_\mathbf{y}} \in {\mathcal T}\), where \({\mathcal T}\) is the triangular group with respect to the Jordan frame \((\mathbf{c}_i)_{i=1}^r\). Then there exist logarithmic functions \(L_1 , L_2,\dots , L_r\) such that \(f( \mathbf{x} ) = \sum_{k=1}^r L_k ( {\Delta_k} (\mathbf{x} ) ), \) where \(\Delta_k\) is the principal minor of order \(k\) with respect to \((\mathbf{c}_i)_{i=1}^r\).