Numerical shadows: measures and densities on the numerical range (Q630556): Difference between revisions

Let \(\Omega_N\) be the unit sphere in \(\mathbb C^N\) (with respect to Euclidean norm) and let \(W(M)=\{(Mu,u)\:u\in\Omega_N\}\) be the numerical range of \(M\in\mathbb C^{N\times N}\). The numerical shadow of \(M\) at a point \(z\in\mathbb C\) is defined as the probability that \((Mu,u)=z\), where \(u\) stands for a random vector \(u\in\Omega_N\). The authors present a rather detailed study of the numerical shadow for some classes of matrices, e.g. \(2\times 2\) matrices, Hermitian and normal matrices, nilpotent Jordan cells. Direct sums of matrices are considered as well.