Error Estimation for Arnoldi-based Model Order Reduction of MEMS

T. Bechtold, E.B. Rudnyi and J.G. Korvink
University of Freiburg, Germany, DE

Keywords: error estimate, Arnoldi algorithm

Abstract:
In this paper we present a heuristic error estimation for the Pade-type approximation of transfer function via Arnoldi algorithm. We suggest using the solution of the Lyapunov equations for reduced-order systems as a stop-criterion in an iterative model order reduction. We further propose a convergence criterion between each two ''neighbored'' reduced models of the order r and r+1.

Nanotech 2004 Conference Technical Program Abstract