Title :
A method for accelerating the first-order stochastic approximation algorithms
Author :
Unton, Fryderyk Z.
Author_Institution :
Polish Academy of Sciences, Warsaw, Poland
fDate :
4/1/1981 12:00:00 AM
Abstract :
The main problem considered is the method of choosing step coefficients in first-order stochastic approximation algorithms for system identification. The proposed method increases the efficency of the Saridis and Stein algorithm [4].
Keywords :
Parameter identification; Stochastic approximation; Acceleration; Approximation algorithms; Control systems; Covariance matrix; Equations; Instruments; Parameter estimation; Stochastic processes; Stochastic systems; System testing;
Journal_Title :
Automatic Control, IEEE Transactions on
DOI :
10.1109/TAC.1981.1102647