On model reduction with consecutively selected rational basis

Abstract

In this paper, we study model reduction by using rational orthogonal bases by consecutively selecting poles for the basis functions. An algorithm modified from adaptive Fourier decomposition (AFD) in Hardy space is presented. In this algorithm, the one-by-one selection of poles makes the numerical algorithm easy to implement, which also keeps the steady-state values equal. The approximating results are comparable to other classical methods.