Asymptotic estimates for interpolation and constrained approximation in \(H^2\) by diagonalization of Toeplitz operators (Q1811167): Difference between revisions

Let \({\mathbb D}\) be the unit disk, \({\mathbb T}\) be the unit circle in the complex plane, and let \({\mathbb T}=I\cup J\), where \(I\) and \(J\) are two disjoint circles \[ I=(e^{-ia},e^{ia}),\quad J=[e^{ia},e^{i(2\pi-a)}],\quad 0<a<\pi. \] It is well known that the Hardy space \(H^2=H^2({\mathbb D})\) consists of all functions in \(L^2({\mathbb T})\) whose Fourier coefficients of strictly negative index are equal to zero. These functions have a Poisson extension in \({\mathbb D}\) that is holomorphic. One can recover the function from its extension by taking non-tangential limits on \({\mathbb T}\) (these limits exist a.e. on \({\mathbb T}\)). So, the authors consider \(H^2\) both as a subset of \(L^2({\mathbb T})\) and as a Hilbert space of holomorphic functions on \({\mathbb D}\). The main purpose of the article is to investigate the following bounded extremal problem: For given \(f\in L^2(I)\), \(\Psi\in L^2(J)\), and \(M>0\), find a function \(g=g_\Psi\in H^2\) to minimize \(\| f-g\| _{L^2(I)}\) under the constraint \(\| \Psi-g\| _{L^2(J)}\leq M\). The authors consider two ways: approximation and interpolation. The solution of the bounded extremal problem can be expressed in terms of a real parameter \(\lambda\in(-1,+\infty)\) playing the role of a Lagrange multiplier. The authors obtain some sharp estimates of the approximation error \(e_\Psi(\lambda)=\| f-g_\Psi(\lambda)\| _{L^2 (I)}^2\) and \(M_\Psi(\lambda)=\| \Psi-g_\Psi(\lambda)\| _{L^2(J)}\). By means of new formulae for the spectral decomposition of certain Toeplitz operators, sharp estimates for Carleman and Krein-Nudel'man approximation schemes are obtained. Some estimates for approximation in the Sobolev class are derived. The problem of approximation of traces of meromorphic functions is also well investigated. In addition, some pointwise convergence results are obtained. The convergence of Goluzin-Krylov approximants is investigated. As an illustration to the results obtained, a numerical example based on experimental data from a hyperfrequency filter is provided. At the conclusion the authors present some analysis of the results obtained and pose some problems. An extended bibliography of the article may be very useful for the reader, especially who is interested in the following investigations in this area.