Maximal spectral distance (Q2564907)

The exact upper bound for the norm distance between two normal matrices is given. The main result is the following: For any normal \(n\times n\) matrices \(A\) and \(B\) with their respective eigenvalues denoted by \(\lambda_i\) and \(\mu_i\), \(i=1, \dots, n\), it holds that \[ \begin{multlined} F(\lambda_1, \dots, \lambda_n; \mu_1, \dots, \mu_n) \triangleq \\ \triangleq \max_{U\in {\mathcal U}} |A-UBU^* |= \max \bigl\{F(\lambda_i, \lambda_j; \mu_k, \mu_l); \quad 1 \leq i,j,k,l \leq n,\;i\neq j;\;k\neq l\bigr\}, \end{multlined} \] where \({\mathcal U}\) is the group of all \(n\times n\) unitary matrices. One also obtains an inequality related to the Cartesian decomposition of a matrix and the evaluation of the Hadamard multiplier norms for matrices in a certain broad class.