Higher-rank numerical ranges and Kippenhahn polynomials (Q1940100): Difference between revisions

For an \(n\times n\) matrix \(A\), the rank \(k\) numerical range of \(A\), which is coincident with the numerical range for \(k=1\), is defined as \[ \Lambda_k(A)=\{\lambda\in\mathbb{C}:\;X^*AX=\lambda I_k \text{\;for some\;} n\times k \text{\;matrix\;} X \text{\;with\;} X^*X=I_k\}, \] and the Kippenhahn polynomial of \(A\) is \(p_A(x,y,z)=\det(x\;Re A+y\;Im A+zI_n)\). The author proves that for two \(n\times n\) matrices \(A\) and \(B\), the Kippenhahn polynomials of \(A\) and \(B\) are the same if and only if \(\Lambda_k(A)=\Lambda_k(B)\) for \(1\leq k\leq \lfloor \frac{n}{2}\rfloor +1\). Combining this result with Li-Sze characterization of rank \(k\) numerical range [\textit{C.-K. Li} and \textit{N.-S. Sze}, Proc.\ Am.\ Math.\ Soc. 136, No. 9, 3013--3023 (2008; Zbl 1152.15027)], one can find that these conditions are equivalent to the situation that the eigenvalue of \(Re(e^{-i\theta}A)\) and \(Re(e^{-i\theta}B)\) coincide for all real \(\theta\).