Abstract
This paper is concerned with the minimum variance control of single-input single-output linear discrete-time stochastic systems with unknown constant parameters. Three types of on-line identification algorithms which have one adjustable parameter (a.p.) are proposed, that is, stochastic approximation type, least squares type and mixed type of stochastic approximation and least squares algorithms. Firstly, an allowable region of a.p. values for each algorithm is derived in order to ensure the global convergence of the corresponding adaptive control system. Secondly, the a.p. value which make the algorithm become asymptotically an approximation of least squares algorithm is determined. Lastly, a numerical example in the second order system is presented to illustrate the effectiveness of the proposed identification algorithm.
