A Large Scale System Model-Order Reduction Method Based on SVD-Krylov

A SVD-Krylov method for large scale MIMO(multi-input multi-output) system model-order reduction is proposed in this paper. Its aim is to combined the singular value decomposition(SVD)and Krylov methods by retaining the best feature that can be applied for solving some problems for the large scale system model-order reduction. The method matches the first r Markov parameters and minimizes the error in the remaining ones in the least-squares sense. The reduced model is asymptotically stable, matches a certain number of moments, and minimizes a weighted error in the discrete time case. The effectiveness of the proposed approaches is tested by the Iss (international space station)model that in the SLICOT library, getting the frequency-response, the error and the error bounds of different order reduced model. The result shows that the proposed method is efficiently.