Hardy spaces of operator-valued analytic functions (Q2267701)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Hardy spaces of operator-valued analytic functions
scientific article

    Statements

    Hardy spaces of operator-valued analytic functions (English)
    0 references
    0 references
    1 March 2010
    0 references
    The author studies the disc analogues of severals results on operator-valued Hardy spaces and BMO defined on \(\mathbb R^n\) by \textit{T.\,Mei} [``Operator valued Hardy spaces'', Mem.\ Am.\ Math.\ Soc.\ 881 (2007; Zbl 1138.46038)]. Given a semifinite von Neumann algebra \(\mathcal M\) (i.e., admitting a normal semifinite faithful trace \(\tau\)), the author denotes by \(\mathcal S\) the linear span of the set of positive elements such that \(\tau(\text{supp\,} x)<\infty\), where \(\text{supp}\,x\) is defined as the least projection \(p\) so that \(px=x\) and by \(L^p(\mathcal M, \tau)\), its completion under the norm \(\|x\|_p=(\tau(|x|^p)^{1/p}\). Given now a \(\sigma\)-finite measure space \((\Omega,\mathcal F, \mu)\), the author denotes by \(S(\Omega, \mathcal S)\) the space of \(\mathcal S\)-valued simple functions \(\varphi=\sum_j a_j\chi_{A_j}\), where \(a_j\in \mathcal S\), and writes \[ \|\phi\|_{L^p_c}= \|(\sum_j a_j^*a_j\mu(F_j))^{1/2}\|_p=\|(\int_\Omega |\varphi|^2d\mu)^{1/2}\|_p \] (and \(\|\phi\|_{L^p_r}=\|\phi^*\|_{L^p_c}\)), and by \(L^p(\mathcal M, L^2_c(\Omega))\) and \(L^p(\mathcal M, L^2_r(\Omega))\), respectively, its completions with these norms. For \(0<p\leq 2\), \(L^p(\mathcal M, L^2_{cr}(\Omega))\) stands for \(L^p(\mathcal M, L^2_c(\Omega))+L^p(\mathcal M, L^2_r(\Omega))\) and for \(2<p\leq \infty\), \(L^p(\mathcal M, L^2_{cr}(\Omega))\) stands for \(L^p(\mathcal M, L^2_c(\Omega))\cap L^p(\mathcal M, L^2_r(\Omega))\) equipped with the standard norms. The Hardy space \(\mathcal H^p(\mathcal M, L^2_c(\mathbb T))\) is then defined as those functions \(f\in L^p(\mathcal M, L^2_c(\mathbb T))\) such that \(P(f)\) defines an \(L^p(\mathcal M)\)-valued analytic function. Similarly notations for the other cases. The atomic formulation is also considered. A function \(a\in L^1(\mathcal M, L^2_c(\mathbb T))\) is called an \(\mathbb M_c\)-atom if it is supported in a subarc \(I\), \(\int_I a=0\) and \(\|a\|_{L^1_c}\leq |I|^{-1/2}\) (that is the noncommutative analogue of the classical \(2\)-atom for \(H^1\)). The space of functions in \(L^1(\mathbb T, L^1(\mathcal M))\) which admit atomic decomposition is denoted by \(H_c^1(\mathbb T, \mathcal M)\), by \(\mathcal H_c^1(\mathbb T, \mathcal M)\) the subspace such that \(P(f)\) is analytic and \(\mathcal H_{cr}^1(\mathbb T, \mathcal M)=\mathcal H_c^1(\mathbb T, \mathcal M)+ \mathcal H_r^1(\mathbb T, \mathcal M)\). The space \(BMO_c(\mathbb T, \mathcal M)\) is defined by the condition \[ \sup_I \|(\frac{1}{|I|}\int_I |f-f_I|^2dm)^{1/2}\|_{\mathcal M}<\infty \] and the corresponding subspace \(BMOA_c(\mathbb T, \mathcal M)\) of elements for which \(P(f)\) is analytic. The paper contains the Carleson type description of functions in \(BMO_c\), the \(H^1-BMO\) dualities for the analytic and non analytic cases and the area integral characterization for these spaces.
    0 references
    0 references
    Hardy spaces
    0 references
    operator-valued analytic functions
    0 references
    bounded mean oscillation
    0 references
    duality
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references