The inverse commutant lifting problem. I: Coordinate-free formalism (Q543467): Difference between revisions

The authors consider the ``normalized'' commutant lifting problem (L.P.): Given two unitary operators \(\mathcal U'\) and \(\mathcal U''\) on Hilbert spaces \(\mathcal K'\) and \(\mathcal K''\), respectively, along with subspaces \(\mathcal K'_{+}\subset \mathcal K'\) and \(\mathcal K''_{+}\subset \mathcal K''\) that are assumed to be \(*\)-cyclic for \(\mathcal U'\) and \(\mathcal U''\), respectively, and such that \(\mathcal U' K'_{+}\subset \mathcal K'_{+}\), \({\mathcal U''}^{*} K''_{-}\subset \mathcal K''_{-} \), and given a contractive operator \(X: \mathcal K'_{+} \longrightarrow \mathcal K''_{-}\) which satisfies the interwining condition \(X \mathcal U'\mid_{\mathcal K'_{+}}\)= \( P_{\mathcal K ''_{-}}\mathcal U''\) \( X\), characterize all contractive interwiners \(Y\) of \((\mathcal U', \mathcal K')\) and\((\mathcal U'', \mathcal K'')\) which lift \(X\) in the sense that \(P_{\mathcal K''_{-}}Y\mid_{\mathcal K'_{+}} = X\). There are known results on linear-fractional parametrization for the set of all solutions; the associated inverse problem asks for a characterization of which Redheffer linear-fractional coefficient-matrices arise in this way for some L.P. As a starting point for their approach, the authors use coupling methods identifying solutions of the L.P. with minimal unitary extensions of an isometry constructed in a natural way from the problem data. The lift \(Y\) is uniquely defined by its moments \(w_{Y}(n)=i_{*}Y^{n}i\), where \(i_{*}\) and \(i\) are certain isometric embedding operators, and the calculation of such moments requires the computation of powers of the so-called universal unitary colligation, determined by the problem data, and the coefficients of the free-parameter unitary colligation. The Introduction of the paper under review and its Sections 2--6 contain an overview of correspondent results and basic ideas (see \textit{V. M. Adamjan}, \textit{D. Z. Arov} and \textit{M. G. Krein} [Funct. Anal. Appl. 2, 269--281 (1968); translation from Funkts. Anal. Prilozh. 2, No. 4, 1--17 (1968; Zbl 0179.46701)]; \textit{R. Arocena} [Ukr. Math. J. 46, No. 3, 151--158 (1994) and Ukr. Mat. Zh. 46, No. 3, 147--154 (1994; Zbl 0835.47008)]; \textit{D. Z. Arov} and \textit{L. Z. Grossman} [Sov. Math., Dokl. 27, 518--522 (1983); translation from Dokl. Akad. Nauk SSSR 270, 17--20 (1983; Zbl 0543.47010)]; \textit{S. S. Boiko}, \textit{V. K. Dubovoy} and \textit{A. Ya. Kheifets} [Oper. Theory, Adv. Appl. 123, 89--138 (2001; Zbl 1008.47025)]; \textit{A. Ya. Kheifets} [Oper. Theory, Adv. Appl. 134, 288--317 (2002; Zbl 1040.47010)] and the literature therein). Sections 7 and 8 use general principles to obtain a parametrization for the set of moments \({w_{Y}(n)}_{n\in\mathbb{Z}}\) and a coordinate-free version of the inverse theorem (Theorem 8.4), to establish a formula for the Redheffer-coefficient matrix in terms of the universal unitary extension.