On the optimal size of system model
The problem of finding the best size for the system model is considered. By explicitly synthesizing the complexity cost of a proposed model, one is able to transform this problem into a constrained integer programming problem.
List of symbols
(s×1) model state vector
(m×1) system control vector
(q×r) model random parameter matrix.
the model size (dimension of state vector, x)
time-varying weighting coefficients
Unable to display preview. Download preview PDF.
- Sage, A.P., and Melsa, J.L. System identification. Academic Press, New York, 1972.Google Scholar
- Special issue on linear-quadratic-Gaussian problem. IEEE Trans. on Automatic Control, Vol. 16, Dec. 1971.Google Scholar
- Graves, L.G., and Wolfe, P. Recent advances in Mathematical programming, McGraw-Hill, Inc., New York, 1963.Google Scholar