Title :
Gradient computation in prediction error identification of linear discrete-time systems
Author :
van Zee, G.A. ; Bosgra, O.H.
Author_Institution :
Wageningen University Agriculture, Wageningen, The Netherlands
fDate :
6/1/1982 12:00:00 AM
Abstract :
A dynamic programming formulation is used to derive efficient expressions for the gradient of the identification criterion with respect to the parameters of a linear time-invariant system. These expressions reduce the computational costs of the parameter optimization step in prediction error identification.
Keywords :
Gradient methods; Parameter identification, linear systems; Prediction methods; Agriculture; Computational efficiency; Dynamic programming; Eigenvalues and eigenfunctions; Mechanical engineering; Optimization methods; Performance analysis; Physics; Predictive models; White noise;
Journal_Title :
Automatic Control, IEEE Transactions on
DOI :
10.1109/TAC.1982.1102969
