Equality of numerical ranges of matrix powers (Q2321350)

The numerical range \(W(A)\) of a complex \(n\times n\) matrix \(A\) is defined as \[ W(A):=\{\langle Ax,x\rangle:x\in \mathbb{C}^n, \|x\|_2=1\}, \] where \(\langle\cdot,\cdot\rangle\) is the standard scalar product on \(\mathbb{C}^n\) and \(\|\cdot\|_2\) is the induced norm. The authors prove the following result: \(W(A)=W(A^k)\) for all \(k\in \mathbb{N}\) if and only if there are quadratic matrices \(B\) and \(C\) such that \(A\) is unitarily similar to \(B\oplus C\), \(B\) is idempotent, i.e., \(B^2=B\), and \(W(C^k)\subseteq W(B)\) for all \(k\in \mathbb{N}\). Next, they define \(k_n\) to be the smallest natural number such that the following holds: if \(A\) is an \(n\times n\) matrix satisfying \(W(A)=W(A^k)\) for all \(k\in \{1,\dots,k_n\}\), then \(A\) is unitarily similar to a matrix of the form \(B\oplus C\) with an idempotent \(B\) (\(A\) has an idempotent direct summand). It is not clear whether \(k_n\) is always finite, but the authors show the following results: \begin{itemize} \item[(i)] \(k_n\geq p_n\), where \(p_n\) is the largest prime less than or equal to \(n+1\). \item[(ii)] If \(A\) is a normal \(n\times n\)-matrix such that \(W(A)=W(A^k)\) for \(k=1,\dots,p_n\), then \(A\) has an idempotent direct summand. \item[(iii)] \(k_1=2\) and \(k_2=k_3=3\). \end{itemize}