Model reduction by minimization of integral square error performance indices (Q918911)

A high-order system is given by the strictly proper function \[ F(s)=\frac{c_ 0+c_ 1s+...+c_{n-1}s^{n-1}}{d_ 0+d_ 1s+...+d_ ns^ n}. \] It is assumed that the poles of F() are distinct. Two second-order reduced models depending on some parameters are considered. The parameters are obtained by minimizing a performance index which measures the integral square error between the output responses of the high-order system and the reduced models.