Konferenční příspěvek

Flídr M., Šimandl M., Král L.

Proceedings of the 11th IASTED International Conference on Intelligent Systems and Control

Orlando, USA

2008

978-0-88986-777-2

Ing. Miroslav Flídr, Ph.D. (FAV Západočeská univerzita, katedra kybernetiky)

adaptive control, dual control, cautious control

clanek_ISC08_MF.pdf [file]

prezentace_ISC08_MF.pdf [file]

The problem of optimal stochastic control of discrete stochastic state space system with uncertain parameters is considered. This problem is analytically unsolvable even for simple systems. A suboptimal dual control approach is presented which makes it possible to solve the multistage optimisation problem analytically. The analytical solvability is guaranteed by employing the partial certainty equivalence principle where the certainty equivalence assumption is enforced only on part of the augmented state. The dual properties of the resulting control are then ensured by a modification of the criterion which evaluates quality of estimates using weighted prediction errors. The proposed dual controller is compared to other non-dual controllers in a numerical example.