An algorithmic approach for system decomposition and balanced realized model reduction

This paper deals with the development of an algorithm for unstable continuous/discrete-time system decomposition into completely stable and completely unstable subsystems using real Schur transformation technique. The application of the proposed algorithm is investigated for unstable LTI plant or controller reduction and the available balanced realized model reduction theories applicable for stable systems are extended for unstable, continuous/discrete-time systems. The algorithm and its applications are illustrated with MATLAB-based computer-simulated results for continuous and discrete-time systems. (C) 2004 The Franklin Institute. Published by Elsevier Ltd. All rights reserved.