Multivariable system reduction via modal methods and padé approximation

A combination of dominant-eigenvalues and Padé approximation methods is proposed for the reduction of a high degree multivariable system. The reduced models so derived are equivalent to those derived by the method in [7]. However, it is computationally much simpler and never fails to produce reduced models. Further it is equally applicable to systems with the number of inputs different from the number of outputs.