On Koosis' approach to the proof of the Carleson interpolation theorem (Q1036951)

Koosis deduced the Carleson interpolation theorem from the Pick theorem. The author develops the Koosis approach to the Carleson theorem and related topics. Let \(\zeta=\{\zeta_j\}_{j=1}^N\) and \(w=\{w_j\}_{j=1}^N\) be two finite sets of complex numbers with mutually distinct \(\zeta_j\), \(\Im\zeta_j>0\), for all \(j\). The Pick problem is to find a function \(f\) which is analytic in the upper half-plane \(\mathbb C_+\) such that \(f(\zeta_j)=w_j\) for all \(j\) and \(\sup_{\mathbb C_+}|f|\leq1\). The Pick theorem asserts that this problem can be solved iff \(D_wU_{\zeta}\overline D_w\leq U_{\zeta}\) for the certain matrix \(U_{\zeta}\) and the diagonal matrix \(D_w\) with diagonal entries \(w_1,\dots,w_N\). Let \(M_{\zeta}\) denote the least number \(M\in[1,\infty]\) for which the Pick problem has a solution for all \(w\) satisfying \(|w_j|\leq M^{-1}\), \(j=1,\dots,N\). In Section 2, the author introduces a matrix \(A_{\zeta}=\Lambda U_{\zeta}\overline{\Lambda}\), where \(\Lambda\) is an invertible diagonal matrix. The author's approach is based on the properties: (1) \(A_{\zeta}^{-1}=\overline A_{\zeta}\); (2) the Pick theorem remains valid if \(U_{\zeta}\) is replaced by \(A_{\zeta}\); (3) roughly speaking, \(A_{\zeta}\) is invariant under conformal automorphisms of \(\mathbb C_+\). Theorem 2.8 states that \(M_{\zeta}=\|A_{\zeta}\|\). In Section 3, estimates of \(\|A_{\zeta}\|\) are obtained. In Section 4, the matrix method is applied to obtain sufficient interpolation conditions. In Section 5, \(A_{\zeta}\) is used in a proof of the Beurling theorem. In Section 6, it is shown that \(A_{\zeta}\) is closely connected with some problems for the Hardy class \(H^2\) in \(\mathbb C_+\). In Section 7, the behavior of \(A_{\zeta}\) under conformal automorphisms of \(\mathbb C_+\) is studied. In Section 8, the author establishes simple properties of self-adjoint matrices \(A\) satisfying \(A^{-1}=\overline A\). Matrices of the form \(A_{\zeta}\) are investigated in Section 9. In particular, a positive self-adjoint matrix \(A=\{a_{jk}\}_{j,k=1}^N\) with non-zero entries is represented as \(A=A_{\zeta}\) iff \(A^{-1}=\overline A\) and \(\text{rank}\{a_{jk}^{-1}\}_{j,k=1}^N=2\). Concluding remarks are given in Section 10.