On some structured inverse eigenvalue problems (Q1370337)

Two special structured inverse eigenvalue problems are investigated. The first one is the Jacobi inverse eigenvalue problem: given some constraints on two sets of reals, find a Jacobi matrix that admits as spectrum and principal subspectrum the two given sets. The polynomial algorithm is based on a special Euclid-Sturm algorithm. The second one is the tridiagonal inverse eigenvalue problem: it is an extension to the nonsymmetric case with the nonsymmetric Lanczos algorithm. The connection between the two algorithms exhibits a similarity transform from the classical Frobenius companion matrix to the tridiagonal matrix. The nonsymmetric Lanczos algorithm may lead to slow convergence, even for a symmetric matrix, since an outer eigenvalue of the tridiagonal matrix of order \(n-1\) can be arbitrarily far from the spectrum of the original matrix.