Gradient methods for optimal linear system reduction†

1 July 1973

journal article
research article
Published by Taylor & Francis in International Journal of Control

Vol. 18 (4) , 767-772
https://doi.org/10.1080/00207177308932554

Abstract

Given a multi-input, multi-output linear time-invariant system, equations have been derived to characterize a system of lower dimension which optimally approximates the given system with respect to the sum of squared errors of the impulse response. Here expressions are developed for the derivative of the penalty function with respect to all parameters of the approximating system, thus enabling the use of gradient methods to minimize approximation errors.

Keywords

This publication has 9 references indexed in Scilit:

Approximation of discrete linear systems
International Journal of Control, 1973
A Solution of the Bilinear Matrix Equation $AY + YB = - Q$
SIAM Journal on Applied Mathematics, 1972
Dynamic-system simplification and an application to power-system-stability studies
Proceedings of the Institution of Electrical Engineers, 1972
Optimum approximation of high-order systems by low-order models†
International Journal of Control, 1971
A new approach to variable metric algorithms
The Computer Journal, 1970
Optimum solution of model-reduction problem
Proceedings of the Institution of Electrical Engineers, 1970
Canonical form for the matrices of linear discrete-time systems
Proceedings of the Institution of Electrical Engineers, 1969
Matrix Equation $XA + BX = C$
SIAM Journal on Applied Mathematics, 1968
Canonical forms for linear multivariable systems
IEEE Transactions on Automatic Control, 1967