Stopping Criterion in Order Reduction of Large Scale Systems Using Krylov Subspace Methods

We show that order reduction by a two‐sided Krylov subspace method is equivalent to an approximate TBR in the sense of Galerkin type conditions. The Hankel Singular Values of the original system are thereby approximated by the Hankel singular values of the Krylov reduced model which are used as a stopping criterion to find a suitable order of the reduced model. (© 2004 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)