Modified Least-Square based Model Reduction using Time Moments and Markov Parameters

This article presents a modified least-square (MLS) method for model order reduction (MOR) of the higher order continuous-time systems by matching the time moments and Markov parameters. The MLS method is based on the selection of number of time moments and Markov parameters which are to be matched that could preserve the characteristics of high order model (HOM) into the reduced-order model (ROM). The key contribution of this article is the formulation of modified matrix expression to obtain the coefficients of numerator and denominator of ROM. It is shown that a combination of time moments and Markov parameters is required to be matched in order to preserve the transient and steady-state responses. The effectiveness of the proposed approach is shown by numerical examples.