Suboptimal mean controllers for bounded and dynamic stochastic distributions

Yongji Wang, Hong Wang

    Research output: Contribution to journalArticlepeer-review

    Abstract

    Based on the recently developed algorithms for the modelling and control of bounded dynamic stochastic systems (H. Wang, J. Zhang, Bounded stochastic distributions control for pseudo ARMAX stochastic systems, IEEE Transactions on Automatic control, 486-490), this paper presents the design of a subotpimal nonlinear mean controller for bounded dynamic stochastic systems with guaranteed stability. The B-spline functional expansion based square root model is used to represent the output probability density function of the system. This is then followed by the design of a mean controller of the output distribution of the system using nonlinear output tracking concept. A nonlinear quadratic optimization is performed using the well known Hamilton-Jacobi-Bellman equation. This leads to a controller which consists of a static unit, a state feedback part and an equivalent output feedback loop. In order to achieve high precision for the output tracking, the output feedback gain is determined by a learning process, where the Lyapunov stability analysis is performed to show the asymptotic stability of the closed loop system under some conditions. A simulation example is included to demonstrate the use of the algorithm and encouraging results have been obtained. © 2002 Elsevier Science Ltd. All rights reserved.
    Original languageEnglish
    Pages (from-to)445-452
    Number of pages7
    JournalJournal of Process Control
    Volume12
    Issue number3
    DOIs
    Publication statusPublished - Apr 2002

    Keywords

    • Asymptotic stability
    • B-spline neural networks
    • Bounded dynamic stochastic systems
    • Learning process
    • Mean control
    • Optimal nonlinear control

    Fingerprint

    Dive into the research topics of 'Suboptimal mean controllers for bounded and dynamic stochastic distributions'. Together they form a unique fingerprint.

    Cite this