Parallel Recursive Algorithms in Asymptotically Efficient Adaptive Control of Linear Stochastic Systems

From MaRDI portal





The authors first give a comprehensive review of some recent developments in stochastic adaptive control of the ARMAX system \[ A(q^{-1})y_ n=B(q^{-1})u_{n-d}+C(q^{-1})\epsilon_ n \] where \(\{y_ n\}\), \(\{u_ n\}\) and \(\{\epsilon_ n\}\) denote the output, input and disturbance sequences respectively, and \(A\), \(B\), and \(C\) are polynomials in \(q^{-1}\). By integrating and refining several basic ideas in these developments, such as using the ``regret instead of quadratic cost criterion [cfr \textit{T. L. Lai} and \textit{H. Robbins}, Ann. Statist., 7, 1196-1221 (1979; Zbl 0426.62059)], constructing self-optimizing inputs via self-tuning [\textit{K. J. Åström} and \textit{B. Wittenmark}, Automatica J. IFAC, 9, 185-199 (1973; Zbl 0249.93049)], the recursive identification algorithm and its modified methods [\textit{P. E. Caines} and \textit{S. Lafortune}, IEEE Trans. Automat. Control AC-29, 312-321 (1984; Zbl 0538.93071)], etc., a relatively complete asymptotic solution to the above adaptive control problem is then provided.











This page was built for publication: Parallel Recursive Algorithms in Asymptotically Efficient Adaptive Control of Linear Stochastic Systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3978273)