The Schur algorithm for generalized Schur functions. I: Coisometric realizations (Q4534545)

A complex-valued function \(f\), defined on the domain \(\Omega\oplus \Omega\), where \(\Omega\) is a subset of the complex plane, is said to have \(\kappa\) negative squares if \(\kappa\) is the smallest nonnegative integer for which the Hermitian matrix \((f(z_i, z_j))_{j=1}^{n}\) has at most \(\kappa\) negative eigenvalues for all choices of the natural number \(n\) and the points \(z_1,\dots,z_n\) in \(\Omega\). For each non-negative integer \(\kappa\), we then define the set \(S_\kappa\) to consist of all those complex-valued functions \(s\) that are meromorphic on the unit disc \({\mathbb D}\) and for which the kernel \[ S_s(z_1, z_2) := {{1-s(z_1)\overline{s(z_2)}}\over{1-z_1\overline{z_2}}} \] has \(\kappa\) negative squares. (Equivalently, \(S_\kappa\) consists of all functions \(s\) having \(\kappa\) poles in the disc and such that \(\limsup_{r\uparrow 1} {|s(re^{it})|}\leq 1\) for almost all \(t\in [0, 2\pi]\).) The union \(S:=\bigcup_{\kappa =0}^{\infty}{S_\kappa}\) is the set of generalized Schur functions. (The classical Schur functions comprise \(S_0\).) The set of those elements of \(S\) that are analytic in some neighborhood of the origin will be denoted \(S^0\). The classical Schur algorithm, applied to a function \(s_0 \in S_0\), yields a sequence of functions \(\{s_j\}\), each in \(S_0\), defined recursively by \[ s_{j+1}(z) = \biggl({{1}\over{z}}\biggr) {{s_j(z)-s_j(0)}\over{1-\overline{s_j(0)}s_j(z)}} \] for \(j\geq 0\). The generalized Schur algorithm applies to functions in the set \(S^0\) defined above. The recursion is the same as in the classical case provided that \(|s_j(0)|<1\), but variations on this pattern can also arise. Theorem 3.1 assures us that, if \(s_0\) is in \(S^0\), then all of the functions \(s_j\) that are generated by the algorithm also lie in \(S^0\). The focus of the present paper is not the Schur algorithm itself, however, but certain operators, called colligations, that induce representations of the functions \(\{s_j\}\). Specifically, for any complex function \(s\) that is analytic in some neighborhood of the origin, there is a Krein space \(({\mathcal K}, \langle \cdot,\cdot\rangle)\), a contraction \(T\) on \({\mathcal K}\), elements \(u\) and \(v\) in \({\mathcal K}\), and a complex number \(\gamma\) such that the function \(s_V\), defined by \[ s_V(z) := \gamma + z\langle (1-zT)^{-1}u, v\rangle \] coincides with \(s(z)\) on some neighborhood of 0. The operator matrix \[ V:= \begin{pmatrix} T&u\cr \langle \cdot, v\rangle &\gamma\end{pmatrix} \] is called an operator colligation on \({\mathcal K}\oplus {\mathbb C}\) while the function \(s_V\) is the operator representation of \(s\) induced by \(V\). If \(V^*\) is an isometry and \(v\) is a cyclic vector for the operator \(T^*\) (i.e., the closed linear span of \(\{ {T^*}^n v\}_{n=0}^{\infty}\) coincides with \({\mathcal K}\)), then \(V\) is said to be a coisometric colligation. If \(V\) is unitary and the closed linear span of \(\{T^m u, {T^*}^n v: m,n\geq 0\}\) coincides with \({\mathcal K}\), then \(V\) is said to be a unitary colligation. In fact, every function \(s\) that is analytic in a neighborhood of 0 has either a coisometric or a unitary operator colligation. The goal of this paper is to study, for a function \(s_0\) in \(S^0\), the coisometric colligations associated to the functions \(\{s_j\}\) generated by the generalized Schur algorithm, and to express these colligations in terms of the coisometric colligation associated to \(s_0\). This is accomplished in Theorems 5.1, 6.1, 7.1, and 8.1, each of which covers one of the variations in the recursion that defines the generalized Schur algorithm. Interestingly, Theorem 9.1 and Corollary 9.2 together show that, for \(s_0\) in \(S^0\), the sequence \(\{s_j\}\) eventually (for large enough \(j\)) consists of classical Schur functions. For the entire collection see [Zbl 0972.00051].