Controller reduction in the two-block H∞-optimal design for distributed plants

From MaRDI portal
Publication:3979760





The aim of the paper is to approximate an infinite-dimensional optimal controller by a finite-dimensional one. The optimization problem is the so-called two-blocks \(H^ \infty\) mixed sensitivity minimization problem for a class of SISO distributed plants. Namely, given a plant \(P\), a controller \(\hat C\) has to be chosen so that \[ \mu(C)=\left\|{W_ 1(1+PC)^{-1}\atop W_ 2 C(1+PC)^{-1}}\right\|_ \infty \] takes its minimum. In a a previous paper by the author et al., under suitable assumptions, the optimal control was proved to exist and to be infinite- dimensional. In the present paper conditions are given so that a \(k\)- dimensional controller \(C_ k\) can be defined with the property that \(C_ k\) tends to \(\hat C\) in the gap metric. Moreover under the same assumptions it is proved that \(\mu(C_ k)\) tends to \(\mu(\hat C)\) and the closed loop system \((C_ k,P)\) is stable for sufficiently large \(k\). An application to an unstable delay system is also given.











This page was built for publication: Controller reduction in the two-block H∞-optimal design for distributed plants

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3979760)