Null/pole interpolation problem in the class \(\mathcal {RI}\) of rational functions intertwining solutions of linear differential equations (Q2514167)

Define the ODE system \[ \mathcal{E}(\alpha,\beta,\gamma)\equiv \lambda\beta(t)u_\lambda(t)-\alpha(t){\partial\over\partial t}u_\lambda(t)+\gamma(t)u_\lambda(t)=0. \] For two such systems \(E=\mathcal{E}(\sigma_1,\sigma_2,\gamma)\) and \(E_*=\mathcal{E}(\sigma_{1*},\sigma_{2*},\gamma_*)\), the class \(\mathcal{RI}\) consists of rational matrix functions \(S(\lambda,\cdot)\) that transform a solution of \(E\) into a solution of \(E_*\), i.e., the null-pole structure of the first solution has to be replaced by a null-pole structure of the second solution for a particular value of \(\lambda\). This class \(\mathcal{RI}\) appears naturally in the theory of 2D systems that are invariant in one direction. These were investigated, e.g., in [the first author, Integral Equations Oper. Theory 71, No. 4, 455--490 (2011; Zbl 1259.47014)] where a theory of vessels was developed. First, the classical null-pole interpolation problem is introduced. Then the 2D systems and vessels are explained, and this is combined to eventually find a solution \(S(\lambda,\cdot)\in\mathcal{RI}\). Some attention goes to the Hermitian realization for contractions in \(\mathcal{RI}\).