The Schur problem and block operator CMV matrices (Q484078): Difference between revisions

The representation of the multiplication operator in a separable Hilbert space with respect to a basis of orthogonal polynomials is a Hessenberg matrix. However, for an alternative basis of orthogonal Laurent polynomials, the representation is a five-diagonal CMV matrix [\textit{M. J. Cantero} et al., Linear Algebra Appl. 408, 40--65 (2005; Zbl 1072.42020)]. Both the Hessenberg and the CMV matrix have elements that can be expressed in terms of the Schur parameters of a Schur function associated with the orthogonality measure. These parameters can be computed, e.g., via the Schur algorithm that solves a (truncated) Schur interpolation problem (equivalent to a trigonometric moment problem). All solutions of the latter problem can be parameterized as a linear fractional transform (depending on four functions defined by the Schur recursion) and an arbitrary Schur function. All this is well known and generalized to operator-valued functions. In this paper, the Schur algorithm is taken as a starting point to solve the (truncated) Schur interpolation problem for operator-valued Schur functions between Hilbert spaces (these are holomorphic and contractive in the complex unit disk). The parametrization of all solutions is analyzed in terms of CMV matrices. See also [the author, Methods Funct. Anal. Topol. 15, No. 3, 201--236 (2009; Zbl 1199.47060)].