Matricial ranges of quadratic operators (Q1974434)

For a bounded linear operator \(T\) on a complex Hilbert space \(H\) the numerical range is \(W(T)= \{\langle Tx,x\rangle: x\in H,\|x\|= 1\}\). An operator \(T\) is called quadratic if it satisfies \(T^2+ \lambda_1T+ \lambda_2I= 0\), for some scalars \(\lambda_1\), \(\lambda_2\). The article contains the following main results: Theorem 1. Let \(T\) be quadratic operator with \(\sigma(T)= \{a,b\}\). (1) If \(a= b\) then \(W(T)\) is either the singleton \(\{a\}\) or the circular disc with center \(a\) and radius \({1\over 2}\|T- aI\|\). (2) If \(a\neq b\) then \(W(T)\) is either the closed line segment connecting \(a\) and \(b\) or the elliptic disc with foci \(a\), \(b\). Theorem 2. If \(T\) is a quadratic operator, then for every \(n\geq 1\), \(W^n(T)\) consists of all \(n\times n\) matrices \(A\) with \(W(A)\subseteq \overline{W(T)}\).