A new frequency weighted model reduction technique using balanced singular perturbation approximation

Abstract

In this paper a new frequency weighted model reduction algorithm is proposed which is based on a combination of Wang et al's technique [14] and balanced singular perturbation approximation. The proposed algorithm results stable reduced order models with both single sided and double-sided weightings. A numerical example is considered to validate the proposed algorithm and the comparison with other well known techniques shows the effectiveness of the proposed algorithm.