Simplification of large linear systems using a two-step iterative method |
| |
| Authors: | FF Shoji RR Mohler TC Hsia |
| |
| Institution: | Department of Electrical and Computer Engineering, Oregon State University, Corvallis, OR 97331, USA;Department of Electrical Engineering, University of California, Davis, CA 95616, USA |
| |
| Abstract: | A simple iterative technique, which is free of certain shortcomings of the previous methods, is proposed for the approximation of large linear systems by a lower- order model. Here, the measure of the goodness of the approximate model is taken to be the value of the integral-square error between the step responses of the exact and the simplified systems. The proposed technique consists of a two-step iterative scheme. In the first step, the optimum residues are obtained by the minimization of the objective function, while the poles (or eigenvalues) are kept constant. In the second step, the poles are optimized while the residues remain fixed. This procedure is continued cyclically until the objective function is satisfactorily minimized. The necessary and sufficient conditions for existence of an optimum are satisfied in each step. The residues, poles and objective functions converge monotonically. The resulting reduced-order model obtained by this method is stable if the original system is stable. The method can also be applied to systems with repeated poles and to multivariable systems. The results are superior to those obtained previously in the steady-state, the point-by-point transient response, and the value of the integral-square error. Illustrative examples are presented. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
