The scalar Nevanlinna-Pick interpolation problem with boundary conditions (Q629533)

The paper under review is devoted to the study of the Nevanlinna-Pick interpolation problem (NP problem). The NP problem studies the existence of a rational function defined in a certain domain (the open unit disc or the right half plane) with values in the unit disc, when finitely many derivatives at a finite set of interpolation points are given. The boundary NP problem requires the interpolation points to be on the boundary of the domain. In the present work, several inductive proofs are given for both problems without using matrix or operator theory, and under a more computational point of view. The first main result (Theorem 4.2) states that, for given \(\{\alpha_{1},\dots,\alpha_{K}\}\subseteq\mathbb{C}^{+}\) and \(\{\beta_{k}^{j}\in\mathbb{C}:0\leq j\leq p_{k}-1,1\leq k\leq K\}\) with \(|\beta_{k}^{j}|<1\), there exists a rational function \(\Phi\), which is analytic in \(\overline{\mathbb{C}^{+}}\), with \(\left\|\Phi\right\|_{\infty}<1\), and which verifies \(\Phi^{(j)}(\alpha_{k})=\beta_{k}^{j}\) for all \(j,k\) considered above, if and only if certain complex numbers constructed from the initial data belong to the unit disc. Moreover, an algorithm is given in order to obtain a solution of this NP problem (Algorithm 4.7). The second main result (Theorem 5.1) enlarges the information when a boundary problem is added. It establishes that if one is asked for a solution of the first problem such that \(\Phi^{(l)}(iw_{t})=\delta_{t}^{l}\) for \(0\leq l\leq m_{t}-1\) and \(1\leq t\leq T\) for a given finite set of real numbers \(w_{t}\) and complex numbers \(\delta_{t}^{l}\) with \(|\delta_{t}^{0}|<1\), then such a function \(\Phi\) exists if and only if the problem is solvable without considering the boundary conditions. Also an algorithm is given to find a solution (Algorithm 5.4). In Section 6, the authors settle necessary and sufficient conditions for the existence and uniqueness of interpolating functions of minimal norm in the previous problems (Theorem 6.1 and Theorem 6.2). An algorithm is provided, to find, when possible, a solution with a norm as close as wanted to the minimum (Algorithm 6.3). Finally, some numerical examples are given.