scientific article; zbMATH DE number 125309

From MaRDI portal
Publication:4027005





[For the entire collection see Zbl 0746.00071.] Integral formulas have long been used to define inner products on spaces of functions. It is natural, therefore, to seek integral formulas for inner products that arise in the study of spaces of analytic functions but that are initially formulated otherwise. The paper under review is concerned with certain Hilbert spaces of analytic functions on the open unit ball \(B\) in \(\mathbb{C}^ n\) that are invariant under a particular group action on \(G\), the connected component of the identity in the group of all biholomorphic automorphisms of \(B\). Specifically, for every \(\lambda\) in \(\mathbb{C}\), define an action \(U_ \lambda\) of \(G\) on functions in \(B\) by \(U_ \lambda (\varphi) f = (f \circ \varphi) \cdot (J \varphi)^{\lambda/(n + 1)}\), for \(\varphi\) in \(G\), where \(J \varphi (z) = \text{det} (\varphi'(z))\). For \(\lambda > n\), \(U_ \lambda\) is a unitary action of \(G\) on \(L^ 2 (\mu_ \lambda)\) which preserves the weighted Bergman space \(L^ 2_ a (\mu_ \lambda)\). (Here \(\mu_ \lambda\) is the probability measure on \(B\) given by \(d \mu_ \lambda = {\lambda - 1 \choose n} (1 - | z |^ 2)^{\lambda/n + 1}dV\), where \(dV\) is normalized Lebesgue measure.) With \(S = \partial B\), the unit sphere in \(\mathbb{C}^ n\), and \(d \sigma\) normalized Lebesgue measure on \( S\), the Hardy space \(H^ 2(S)\) is the closure of the analytic polynomials in \(L^ 2 (S,d \sigma)\). \(H^ 2(S)\) is invariant under the action of \(U_ n\) on \(G\). More generally, for \(\lambda>0\), the function \(K_ \lambda (z,w) = (1-(z, w))^{-\lambda}\), for \(z\) and \(w\) in \(B\), is positive definite. Define an inner product \((\cdot, \cdot)_ \lambda\) on \(\text{span} \{K_ \lambda (\cdot, w) : w \in B\}\) by requiring that \((K_ \lambda (\cdot, w)\), \(K_ \lambda (\cdot,z))_ \lambda = K_ \lambda (z,w)\). Letting \(H_ \lambda\) denote the completion of this space, we find that \(U_ \lambda\) is an isometric action of \(G\) on \(H_ \lambda\). (Also, \(H_ \lambda = L^ 2_ a (\mu_ \lambda)\) for \(\lambda > n\) and \(H_ n = H^ 2(S)\).) The authors also construct the spaces \(H_ \lambda\) for \(\lambda = 0\), \(-1\), \(-2, \dots\). The space \(H_ 0\) is the generalization of the Dirichlet space. Thus, the authors wish to find integral formulas for the inner products \((\cdot, \cdot)_ \lambda\) on the spaces \(H_ \lambda\) for the cases \(\lambda > 0\) and \(\lambda = 0\), \(-1\), \(-2, \dots\). They use a method of J. Peetre, the essence of which is ``an analytic continuation (in the parameter \(\lambda)\) of the integrals \(\int_ Bf \overline gd \mu_ \lambda\) via integration by parts in the radial direction. Useful formulas are obtained in the cases \(\lambda > 0\) (Theorem 1), \(\lambda = 1,2, \dots,n\) (Theorem 1 and Corollary 2), and \(\lambda = 0\), \(-1\), \(-1, \dots\) (Theorem 2 and Corollary 1). For example, letting \(Rf(z) = {\partial \over \partial t} f(tz) |_{t = 1}\) be the radial derivative of \(f\) and \({\mathcal R}_ k = \prod^ k_{j = 1} (R + n - j)\), Theorem 1 shows that, for \(f\) and \(g\) analytic in a neighborhood of \(\overline B\) and for \(\lambda > 1\), \[ (f,g)_ \lambda = {(\lambda - 1) \over n!} \int_ B ({\mathcal R}_{n-1} f)(z) \overline {g(z)} {\bigl( 1 - | z |^ 2)^{\lambda - 2} \over | z |^{2(n - 1)}} dV. \] Theorem 3 discusses analogously defined bilinear forms \((\cdot, \cdot)_ \lambda\) for arbitrary \(\lambda\) in \(\mathbb{C}\). As the article is from the proceedings of a conference, some of the details have been omitted. However, the principal techniques are discussed and proofs are included for the main results.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4027005)