Stable Optimal Model Reduction of Linear Discrete-Time Systems

Abstract

A new algorithm is presented for finding reduced-order models of the linear time-invariant systems. Proposed recursive algorithm minimizes a weighted mean squared impulse (or step) response error between the original system and the reduced order model. This proposed optimization algorithm require solving very easy linear equations rather than the usual complicated non-linear equations. The procedure guarantees a reduced order stable model at each recursive stage. Method is easily extended to multi-input/multi-output systems. Algorithm’s performance is compared with other methods reported in the literature.