Approximate parametric optimization of infinite-dimensional systems
The author considers the problem of finding \(g\in M_n\), such that \[ \|f- g\|_{L^2(0,\infty)} =\min_{y\in M_n} \|f- y\|_{L^2(0,\infty)}, \] where \(M_n\) is the \(n\)-dimensional subspace of the complex Hilbert space \(L^2(0,\infty)\) spanned by an \(n\)-tuple of normalized eigenvectors of the operator \(A^* h=h'\), \(D(A^*)= W^{1,2}[0,\infty]\), corresponding to eigenvalues \(\{\lambda_k\}^n_{k=1}\subset \Pi^-\). In a previous work, the author showed that the solution \(g= P_nf\), where \(P_n\) denotes the orthoprojector onto \(M_n\) can be expressed in terms of the Malmquist functions. In the present paper an alternative approach which is more suitable for standard mathematical software is presented. The convergence problem is discussed from the theoretical and the numerical viewpoint and the results are illustrated for the problem of synthesizing the optimal proportional controller for stabilizing a distributed plant.
This page was built for publication: Approximate parametric optimization of infinite-dimensional systems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4254141)
