A unitary Hessenberg \(QR\)-based algorithm via semiseparable matrices (Q2566271)

A novel algorithm for computing the eigenvalues of a unitary Hessenberg matrix is given. The proposed algorithm is \(QR\) based. The algorithm uses a linear fractional transformation to convert the initial eigenvalue problem for the unitary Hessenberg matrix into a modified eigenvalue problem for the Hermitian diagonal-plus-semiseperable matrix. Exploiting the structure of this associated eigenvalue problem enables the author to get a quadratic time using a linear memory space. Extensive numerical experiments showing the effectiveness of the proposed algorithm are given.