Schur-Agler class and Carathéodory extremal functions

Anindya Biswas · 2025

Read Schur-Agler class and Carathéodory extremal functions by Anindya Biswas (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-17658.txt.

Introduction

Anindya Biswas studies special analytic functions and their extreme cases in a 2025 arXiv paper. His setting is the Schur-Agler class, a family of functions generated by a chosen set of test functions. He focuses on Carathéodory extremal functions, which are the boundary cases that shape the class. The central question is which domains such a family can describe.

Biswas shows that, under stated conditions, only two domains allow finitely many test functions to generate the full Schur class. These are the unit disc and the product of two unit discs. He then describes the extremal functions for a well known function space's multiplier algebra. He also gives operator based representations for a broad class of hyperbolic domains.

Catalog record

CREATORAnindya Biswas
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE26 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2510-17658

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Schur-Agler class and Carathéodory extremal functions Anindya Biswas Abstract . We study the role of Carathéodory extremal functions in the Schur-Agler class generated by a collection of test functions. We show that under certain conditions, 𝔻 \mathbb{D} and 𝔻 2 \mathbb{D}^{2} are the only domains where finitely many test functions can generate the Schur class. As applications, we give a description of the Carathéodory extremals in the unit ball of the multiplier algebra of the Drury-Arveson space and give operator-theoretic Herglotz representations for any Carathéodory hyperbolic domain. 2020 Mathematics Subject Classification Primary 47A48, 32F45; Secondary 32E30 1. Introduction. Suppose that \Omega is a Carathéodory hyperbolic domain in some \mathbb{C}^{m} . For the well-studied domains the open unit disc \mathbb{D} and the bidisc \mathbb{D}^{2} , any element of the Schur class (the closed unit ball of the Banach algebra of all bounded holomorphic function on \Omega , denoted by \mathscr{S}(\Omega) ) has a structure known as the realization formula , which states that given any f\in\mathscr{S}(\Omega) , there is a Hilbert space \mathscr{X} , a unitary operator U:\mathbb{C}\oplus\mathscr{X}\rightarrow\mathbb{C}\oplus\mathscr{X} and an operator z\in B(\mathscr{X}) which is identified with the identity function on \Omega such that with U=\bordermatrix{&\mathbb{C}&\mathscr{X}\cr\mathbb{C}&a&B\cr\mathscr{X}&C&D}, one can express f as f(z)=a+Bz(I_{\mathscr{X}}-Dz)^{-1}C for all z\in\Omega . In the theory of interpolation, the realization formula has certain important implications. For instance, it is the realization formula which leads to the fact that every solvable Pick interpolation problem on \mathbb{D}^{2} has a rational inner solution ( [ 4 ] , [ 5 ] ). Due to such importance, the property of having a realization formula has been investigated further and it has been found that in the case of the annulus ( [ 9 ] , [ 13 ] ) and the symmetrized bidisc ( [ 12 ] ) similar form can be established, though not as convenient as in the case of \mathbb{D} and \mathbb{D}^{2} , but still provides useful and far reaching consequences ( [ 11 ] ). In the case of \mathbb{D} and \mathbb{D}^{2} , in a certain sense, only a finite number of elementary functions can generate the whole Schur class, namely, the identity function on \mathbb{D} and the coordinate functions on \mathbb{D}^{2} , although this is not the case for the other domains. For the annulus and the symmetrized bidisc, one requires infinitely many functions to provide a realization formula, and in this article we will see that beyond the domain \mathbb{D} and \mathbb{D}^{2} , it is indeed necessary to consider such an infinite collection. Generalizing the idea that a certain collection of functions has the property of generating the whole Schur class leads us to a notion known as the Schur-Agler class , or more precisely, the \Psi -Schur-Agler class for a given collection of functions \Psi ( [ 9 ] , [ 11 ] , [ 13 ] ). We call such a collection \Psi a set of test functions and they are introduced in the following way. Definition 1.1 . A collection \Psi of \mathbb{C} -valued functions on \Omega is called a set of test functions if the following conditions hold. (1) \sup\{|\psi(z)|:\psi\in\Psi\}<1 for each z\in\Omega . (2) For each finite subset F of \Omega , the collection \{\psi|_{F}:\psi\in\Psi\} together with the constant function generates the algebra of all \mathbb{C} -valued functions on F . We can topologize \Psi by considering it as a subspace of \overline{\mathbb{D}}^{\Omega} equipped with the product topology. For every z\in\Omega , we define an element E(z) in \mathcal{C}_{b}(\Psi) , the C^{*} -algebra of all bounded continuous functions on \Psi , such that E(z)(\psi)=\psi(z) . Clearly, (1.1) \displaystyle\|E(z)\|=\text{sup}_{\psi\in\Psi}|\psi(z)|<1 for each z\in\Omega . These functions E(z) have a key role in the development of the theory of \Psi -Schur-Agler class. Given such a set \Psi , we next introduce a collection of positive semidefinite functions or kernels k:\Omega\times\Omega\rightarrow\mathbb{C} satisfying the property \displaystyle\big((1-\psi(z)\overline{\psi(w)})k(z,w)\big)_{z,w\in\Omega}\geq\mathbf{0},\,\ \text{for all}\,\,\psi\in\Psi. These functions k are called the \Psi -admissible kernels and we will denote the set of all \Psi -admissible kernels by \mathcal{K}_{\Psi} . Now we can describe the \Psi -Schur-Agler class. Definition 1.2 . Given a collection of test functions \Psi on a domain \Omega , the \Psi -Schur-Agler class is the collection of functions f:\Omega\rightarrow\mathbb{C} for which \displaystyle\big((1-f(z)\overline{f(w)})k(z,w)\big)_{z,w\in\Omega}\geq\mathbf{0},\,\ \text{for all}\,\,k\in\mathcal{K}_{\Psi}. The \Psi -Schur-Agler class on \Omega will be denoted by \mathscr{S\mspace{-5mu}A}_{\psi}(\Omega) , or \mathscr{S\mspace{-5mu}A}_{\Psi} when the underlying domain is clear from the context. One should note that, we do not require \Psi to be a subset of \mathscr{S}(\Omega) , but then \mathscr{S\mspace{-5mu}A}_{\psi}(\Omega) , which also contains \Psi , is not a subset of \mathscr{S}(\Omega) . On the other hand, if we take \Psi from \mathscr{S}(\Omega) , \mathscr{S\mspace{-5mu}A}_{\psi}(\Omega) is then contained in \mathscr{S}(\Omega) and it inherits a number of important properties from \mathscr{S}(\Omega) . For instance, since to check positivity of \big((1-f(z)\overline{f(w)})k(z,w)\big)_{z,w\in\Omega} one needs to consider only finitely many points from \Omega at a time, \mathscr{S\mspace{-5mu}A}_{\psi}(\Omega) is closed under pointwise convergence. For the development with holomorphic test functions, an interested reader may consult [ 11 ] . The main result providing the structure of \mathscr{S\mspace{-5mu}A}_{\psi}(\Omega) is the following (see [ 9 ] , [ 11 ] , [ 13 ] ). Theorem 1.3 . For a function f:\Omega\rightarrow\overline{\mathbb{D}} , the following statements are equivalent. (1) f\in\mathscr{S\mspace{-5mu}A}_{\psi}(\Omega) . (2) f has an Agler decomposition, that is, there exists a completely positive kernel \Gamma:\Omega\times\Omega\rightarrow\mathcal{C}_{b}(\Psi)^{*} such that (1.2) \displaystyle 1-f(z)\overline{f(w)}=\Gamma(z,w)(1-E(z)E(w)^{*})\,\,\,\text{for all $z,w\in\Omega$.} (3) There is a Hilbert space \mathscr{X} , a unitary U:\mathbb{C}\oplus\mathscr{X}\rightarrow\mathbb{C}\oplus\mathscr{X} and a unital * -representation \rho:\mathcal{C}_{b}(\Psi)\rightarrow B(\mathscr{X}) such that writing U as \displaystyle U=\bordermatrix{&\mathbb{C}&\mathscr{X}\cr\mathbb{C}&a&B\cr\mathscr{X}&C&D}, one can express f as f(z)=a+B\rho(E(z))(I_{\mathscr{X}}-D\rho(E(z)))^{-1}C for all z\in\Omega . (4) There is a Hilbert space \mathscr{X} , a unital * -representation \rho:\mathcal{C}_{b}(\Psi)\rightarrow B(\mathscr{X}) and a map h:\Omega\rightarrow\mathscr{X} such that (1.3) \displaystyle 1-\overline{f(z)}f(w)=\langle\rho(1-E(z)^{*}E(w))h(w),h(z)\rangle_{\mathscr{X}}\,\,\,\text{for all $z,w\in\Omega$.} By the completely positive kernel \Gamma in ( 1.2 ) we mean that for any z_{1},\ldots,z_{n}\in\Omega , a_{1},\ldots,a_{n}\in\mathcal{C}_{b}(\Psi) , the matrix \big(\Gamma(z_{i},z_{j})(a^{*}_{i}a_{j})\big)_{1\leq i,j\leq n} is positive semidefinite. It is known ( [ 8 ] , [ 9 ] , [ 10 ] ) that, for any such \Gamma , there is a Hilbert space \mathscr{X} , a unital * -representation \rho:\mathcal{C}_{b}(\Psi)\rightarrow B(\mathscr{X}) and a map g:\Omega\rightarrow\mathscr{X}(=B(\mathscr{X},\mathbb{C})) such that (1.4) \displaystyle\Gamma(z,w)(\delta)=\langle\rho(\delta)g(w)^{*},g(z)^{*}\rangle. In this article, we will always assume that our collection \Psi of test functions are elements of \mathscr{S}(\Omega) and that there is a point w_{0}\in\Omega such that \psi(w_{0})=0 for all \psi\in\Psi . As a consequence, \mathscr{S\mspace{-5mu}A}_{\Psi}\subset\mathscr{S}(\Omega) and E(w_{0})=0 . The fact that this assumption does not alter the \mathscr{S\mspace{-5mu}A}_{\Psi} can be found in [ 11 ] . Let us now describe our work in this article. In Section 2 we establish a necessary condition for a collection of test functions to generate the Schur class and show that under certain conditions \mathbb{D} and \mathbb{D}^{2} are the only domains such that finitely many test functions can generate the Schur class. We show that the collection of test functions and the admissible kernels are closely related to certain extremal functions. In the next section we give two applications of our results. The first one describes the range of Carathéodory extremals in the unit ball of the multiplier algebra of the Drury-Arveson space. As a second application, we give an operator theoretic Herglotz representation for any Carathéodory hyperbolic domain. 2. Extremal functions

Read the complete text