Realizations of holomorphic and slice hyperholomorphic functions: the Krein space case (Q783654)

The authors investigate some realization results for operator-valued functions, analytic in a neighbourhood of the origin, in the complex setting and in the quaternionic setting as well. In the complex setting, which will be briefly described, the authors start by fixing a Hilbert space \(\mathcal C\), associated with a reproducing kernel Hilbert space \({\mathbf H}^-_{2,r}(\mathcal C)\), where \(0<r<1\), the latter consisting of power series of the form \(f(z)=\sum_{u=1}^\infty f_uz^{-u}\) with coefficients in \(\mathcal C\), satisfying \(\sum_{u=1}^\infty R^{2u}\Vert f_u\Vert^2<\infty\), \(R=1/r\). They also consider an \({\mathbf L}(\mathcal C,\mathcal C)\)-valued function \(\Phi\), analytic in the disk centered at 0, with radius \(r_0<r\) and continuous on its closure. Other used entities are the function \(\Phi^\sharp (z)=(\Phi(\bar{z}))^*\), the reproducing Krein space \(\mathcal{L}(\Phi^\sharp)\) defined via the function \(\Phi^\sharp\), and the operator \(R_0\) given by \(R_0f(z)=z^{-1}(f(z)-f(0))\) for \(z\neq 0\), and \(=f'(0)\) for \(z=0\). The main result in the complex setting is Theorem 4.5, which asserts that the Krein space \(\mathcal{L}(\Phi^\sharp)\) is \(R_0\) invariant, and \(R_0\) is coisometric in this space. Moreover, if \(C\) is the point evaluation map at the origin, then \(C\) is a bounded linear operator from \(\mathcal{L}(\Phi^\sharp)\) into \(\mathcal C\), and \[ \Phi(z)=i\Im\Phi(0)+\frac{1}{2}C(I-zR_0^{[*]})(I+zR_0^{[*]})^{-1}C^{[*]}, \] where the upper index \({[*]}\) means ``conjugation'' in the Krein space. Unlike in other previous papers by several authors including the present ones, this representation replaces a unitary map in a Krein space, not necessarily continuous, by a continuous coisometry. Operator ranges are one of the main tools used by the authors. In the last three sections of this paper, the authors deal with the quaternionic setting. They recall some facts on slice hyperholomorphic functions and prove the quaternionic version of the realization theorem. A~main ingredient in this part of the work is the use of what the authors call a \(\star\)-operation (see their Definition 5.3), which is a product in the class of slice hyperholomorphic functions. This is, in fact, a version of a similar concept given by Definition 4.3.15 appearing in the monograph by [\textit{F. Colombo} et al., Noncommutative functional calculus. Theory and applications of slice hyperholomorphic functions. Basel: Birkhäuser (2011; Zbl 1228.47001)], which may look somewhat mysterious. In order to understand the naturalness of such a concept, one might consider an equivalent definition of the space of slice hyperholomorphic functions via a linear isomorphism given by a Cauchy type transform, defined on an algebra of stem functions (which form a special class of holomorphic functions in the complex plane), which was recently given by the reviewer in [Integral Equations Oper. Theory 92, No. 2, Paper No. 18, 19 p. (2020; Zbl 1443.47088)] in an attempt to simplify and explain some older results of the present and other authors. In particular, the \(\star\)-operation in the space of usual slice hyperholomorphic functions (or in a more general context as well), may be introduced by means of a~transfer of the natural multiplicative structure of the algebra of stem functions (or of other more general algebras) via the Cauchy transform.