An inverse eigenvalue problem and a matrix approximation problem for symmetric skew-Hamiltonian matrices (Q2463441)

Two problems are discussed. First an inverse eigenvalue problem for symmetric skew-Hamiltonian matrices. Given a real \(k \times m\) matrix \(X\), the matrix of eigenvectors and a diagonal matrix \(\Lambda\), the set of eigenvalues, find all symmetric skew-Hamiltonian matrices \(N\), such that \(NX = X\Lambda\), i.e. those matrices having this eigenvectors and eigenvalues. In Theorem 1 a solvability condition is given, thus describing all solutions. The second problem is to find for a given \(k \times k\) matrix \(C\) a symmetric skew-Hamiltonian matrix \(N\) which is nearest to \(C\) in the Frobenius matrix norm. Theorem 2 describes the unique solution. In addition a perturbation analysis of the second problem is given in Theorem 3. A set of numerical examples is described.