Interpolation and harmonic majorants in big Hardy-Orlicz spaces (Q926380)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Interpolation and harmonic majorants in big Hardy-Orlicz spaces
scientific article

    Statements

    Interpolation and harmonic majorants in big Hardy-Orlicz spaces (English)
    0 references
    0 references
    27 May 2008
    0 references
    For a sequence \(\Lambda\) in the unit disk \(\mathbb D\) and a space of holomorphic functions \(X \subset H(\mathbb D)\), the interpolation problem consists in describing the trace of \(X\) on \(\Lambda\), i.e., the set of restrictions \(X| _\Lambda\) regarded as a sequence space. A sequence \(\Lambda \subset\) \(\mathbb D\) is called free interpolating for \(X\) if [\((a_n) \in X| _\Lambda\) and \((b_n)\) bounded] implies \((a_nb_n) \in X| _\Lambda\). Free interpolation in Hardy spaces is characterized by the well-known Carleson condition. The result extends to Hardy-Orlicz spaces contained in the scale of classical Hardy spaces \(H^p\), \(p > 0\). Assume that \(\varphi :\mathbb R\to [0,+\infty )\) is strongly convex, that is, \(\varphi\) is convex, nondecreasing, and satisfies \(\lim_{t \to \infty} \varphi (t)/t = \infty\) and \(\varphi (t+2) \leq M \varphi (t) + K\) \((t \geq t_0)\) for some constants \(M,K \geq 0\) and \(t_0 \in\mathbb R\). The last property is implied by the so-called \(\Delta_2\)-condition: There exist constants \(M,K \geq 0\) and \(t_0\) such that \(\varphi (2t) \leq M \varphi (t) + K\) \((t \geq t_0)\). The function \(\varphi\) is said to satisfy the \(\nabla_2\)-condition if there exist \(d > 1\) and \(t_0 > 0\) such that \(2 \varphi (t) \leq \varphi (dt) / d\) \((t \geq t_0)\). If \(\varphi\) is a strongly convex function and \(\Phi := \varphi \circ \log\), then the associated Hardy-Orlicz class is defined as \[ H_\Phi = \left\{f \in N^+: \int_{\mathbb T} \varphi (\log | f(t)| ) \, d\sigma (t) < +\infty\right\}, \] where \(\mathbb T\) is the unit circle, \(\sigma\) is the normalized length measure on \(\mathbb T\), \(f(t)\) is the non-tangential boundary value of \(f\) at \(t \in\mathbb T\), and \(N^+\) is the Smirnov class, that is, the class of holomorphic functions \(f\) on \(\mathbb D\) such that \[ \lim_{r \to 1} {1 \over 2\pi} \int_0^{2\pi} \log^+ | f(re^{i\theta})| \, d\theta = {1 \over 2\pi} \int_0^{2\pi} \log^+ | f(e^{i\theta})| \, d\theta . \] By \(L^\varphi\) it is denoted the standard Orlicz space of measurable functions \(u\) such that \(\varphi \circ | u| \in L^1(\mathbb T)\). For the Smirnov and the Nevanlinna classes, interpolating sequences have been characterized in terms of the existence of harmonic majorants (quasi-bounded in the case of Smirnov class) in a recent paper by \textit{A. Hartmann, X. Massaneda, P. Thomas} and \textit{A. Nicolau} [J. Funct. Anal. 217, 1--37 (2004; Zbl 1068.30027)]. Since the Smirnov class can be regarded as the union over all Hardy-Orlicz spaces associated with the strongly convex functions, it is natural to ask how the condition changes from the Carleson condition in \(H^p\) to harmonic majorants in \(N^+\). The aim of the paper under review is to narrow down this gap from \(N^+\) to ``big'' \(H_\Phi\) spaces. The author characterizes free interpolating sequences for a class of \(H_\Phi\) spaces that carry an algebraic structure and are strictly bigger than \(\bigcup_{p>0} H^p\). Specifically, the main result of this paper is the following. Let \(\varphi\) be a strongly convex function satisfying the \(\Delta_2\)-condition. Then the following two assertions are equivalent: {\parindent=6mm \begin{itemize}\item[(a)] \(\Lambda\) is a free interpolating sequence for \(H_\Phi\). \item[(b)] There exists a positive measurable function \(w \in L^\varphi\) such that \(\varphi_ \Lambda \leq P[w]\), where \(P[w]\) is the Poisson transform of \(w\) and \(\varphi_ \Lambda\) is defined as \(0\) if \(z \notin \Lambda\), and as \(\log | B_\lambda (\lambda )| ^{-1}\) if \(z = \lambda \in \Lambda\). Here \(B_ \lambda (z)\) is the Blaschke product \(\prod_{\mu \in \Lambda \setminus \{\lambda\}} {| \mu| \over \mu} {\mu - z \over 1 - \overline{\mu}z}\). \item[(c)] \(H_\Phi | _ \Lambda = \{(a_\lambda): \exists w \in L^\varphi\), \(w\geq0,\) with \(\log^+ | a_\lambda | \leq P[w] (\lambda )\}.\) \end{itemize}} If, moreover, \(\varphi\) satisfies the \(\nabla_2\)-condition, then the above three conditions are equivalent to the following: There exists a constant \(C>0\) such that for any sequence of non-negative numbers \((c_\lambda )\), \[ \sum_{\lambda \in \Lambda} c_\lambda \varphi_\Lambda (\lambda ) \leq C \left\|\sum_{\lambda \in \Lambda} c_\lambda P_\lambda\right\|_{(L^\varphi )^*}, \] where \(P_z(t) = (1 - | z| ^2)/| t-z| ^2\) is the usual Poisson kernel. Note that Carleson's condition can be reformulated as \(\sup_{\lambda} \varphi_\Lambda (\lambda ) < +\infty\). The existence of harmonic majorants defined by functions in the last \(L^\varphi\) is also discussed within a general situation. In fact, the main result is proved via the following theorem: Let \(\varphi\) be a strongly convex function satisfying both \(\Delta_2\) and \(\nabla_2\) conditions. If \(u\) is a non-negative Borel function on the unit disk, the following two assertions are equivalent: (a) There exists a function \(w \in L^\varphi\) such that \(u(z) \leq P[w](z)\) for all \(z \in\mathbb D\). (b) \(\sup \{\int u \, d\mu:\mu\) positive measure on \(\mathbb D\) with \(\| P[d\mu ]\| _{(L^\varphi )^*} \leq 1 \} < +\infty\). The paper is finished with a class of examples of separated Blaschke sequences which are interpolating for certain Hardy-Orlicz spaces without being interpolating for slightly smaller ones.
    0 references
    0 references
    0 references
    0 references
    0 references
    free interpolating sequences
    0 references
    Smirnov class
    0 references
    Nevanlinna class
    0 references
    Carleson condition
    0 references
    0 references
    0 references