Title :
Computing minimal partial realizations via a Lanczos-type algorithm for multiple starting vectors
Author :
Freund, Roland W.
Author_Institution :
Lucent Technol., AT&T Bell Labs., Murray Hill, NJ, USA
Abstract :
We describe a Lanczos-type procedure that reduces a given realization of a finite sequence of (moment) matrices to a minimal partial realization. A key feature of this procedure is that the underlying Lanczos-type algorithm is directly applied to the matrix triplet describing the given realization, rather than to the moment matrices. It thus avoids explicit formulation of and the usually unstable computation with the moment matrices
Keywords :
iterative methods; linear systems; matrix algebra; vectors; Krylov subspace; Lanczos-type algorithm; iterative methods; linear dynamical systems; matrix algebra; minimal partial realizations; multiple starting vectors; Circuit simulation; Large-scale systems; Linear systems; Matrix decomposition; Reduced order systems; Robustness; Sparse matrices; Transfer functions; Very large scale integration;
Conference_Titel :
Decision and Control, 1997., Proceedings of the 36th IEEE Conference on
Conference_Location :
San Diego, CA
Print_ISBN :
0-7803-4187-2
DOI :
10.1109/CDC.1997.649603
