Laguerre-SVD reduced-order modeling

Abstract

A reduced-order modeling method based on a system description in terms of orthonormal Laguerre functions, together with a Krylov subspace decomposition technique is presented. The link with Pade approximation, the block Arnoldi process and singular value decomposition (SVD) leads to a simple and stable implementation of the algorithm. Novel features of the approach include the determination of the Laguerre parameter as a function of bandwidth and testing the accuracy of the results in terms of both amplitude and phase.