The Algebraic Riccati Equation (ARE) cannot be formulated for the conservative/lossless and allpass cases, though the notion of `storage function' is well-defined for these cases too. New properties have been formulated recently about the storage function matrix for this case, which gave rise to new computational procedures. This paper targets improvement of this algorithm by avoiding some key computation intensive steps in minimal polynomial basis computation. We use the Zassenhaus method for basis computation for the sum and intersection of two subspaces. In addition to the conventional Zassenhaus method, for improved numerical accuracy, we propose LU and QR factorization methods with pivoting and compare the results.
|Proceedings of the 22nd International Symposium on Mathematical Theory of Networks and Systems
|ISBN van elektronische versie
|Published - 2016
|22nd International Symposium on the Mathematical Theory of Networks and Systems - Minneapolis, MN, United States
Duur: 12-jul.-2016 → 15-jul.-2016
|22nd International Symposium on the Mathematical Theory of Networks and Systems
|12/07/2016 → 15/07/2016