Generalized inverse eigenvalue problem with mixed eigendata (Q448350)

The authors establish necessary and sufficient conditions for the existence of a unique solution to the following problem. Given a \(k\times k\) Jacobi matrix \(J_k\), a nonsingular \(n\times n\) diagonal matrix \(C_n\), distinct real numbers \(\lambda\) and \(\mu\), and vectors \(x_2\) and \(y_2\) in \({\mathbb R}^{n-k}\), where \(1\leq k<n\), find an \(n\times n\) Jacobi matrix \(J_n\), with leading principal minor \(J_k\), and vectors \(x_1\) and \(y_1\) in \({\mathbb R}^k\), such that \(J_nx=\lambda C_nx\) and \(J_ny=\mu C_ny\), where \(x=(x_1^T,x_2^T)^T\) and \(y=(y_1^T,y_2^T)^T\).