Non-isotropic flag singular integrals on multi-parameter Hardy spaces (Q717716)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Non-isotropic flag singular integrals on multi-parameter Hardy spaces
scientific article

    Statements

    Non-isotropic flag singular integrals on multi-parameter Hardy spaces (English)
    0 references
    5 October 2011
    0 references
    The author proves that if \(T\) is a flag singular integral operator then for any \(0<p\leq 1\) there is a constant \(C(p)\) such that \[ \|T(f)\|_{H^p_F}\leq C(p)\|f\|_{H^p_F} \, . \] It is also proved that if such if a linear operator is \(L^2\) bounded and bounded on the flag \(H^p\) space then it is bounded from the flag \(H^p\) space to \(L^p\). The space \(H^p_F(\mathbb{R}^n\times \mathbb{R}^m)\) consists of those locally \(p\)-integrable distributions such that a suitable square function \(S_d(f)\) belongs to \(L^p(\mathbb{R}^n\times \mathbb{R}^m)\). The square function \(S_d(f)\) has square \[ S_d^2 (f)(x,y)= \sum_{j,k,I,J} |\psi_{j,k}\ast f\, (x_I,x_J)|^2 \chi_I(x)\chi_J(y)\, . \] Here, \(I\subset\mathbb{R}^n\), \(J\subset\mathbb{R}^m\) are dyadic cubes with sidelengths \(\ell(I)=2^{-j-N}\), \(\ell(J)=2^{-k-N}+2^{-2j-N}\) for a large but fixed \(N\) and \(x_I,x_J\) are fixed points in \(I\) and \(J\). Also, \(\psi_{j,k}\) has the special form \[ \psi_{j,k}(x,y) =\psi^{(1)}_j\ast_2 \psi^{(2)}_k (x,y) \] where \( \psi^{(1)}_j (x,y)=2^{(n+2m)j} \psi^{(1)}(2^j x, 2^j y)\) and \( \psi^{(2)}_k (z)=2^{mk} \psi^{(2)}(2^k z)\). By \(\ast_2\) one means the convolution in the second parameter \(\mathbb{R}^m\) so \[ \psi\ast_2\phi (x,y)=\int_{\mathbb{R}^m} \psi (x,y-z)\, \phi(z)\, dz \] when \(\psi\) is defined on \(\mathbb{R}^{n+m}\) and \(\phi\) on \(\mathbb{R}^m\). The author proves the following Calderón type reproducing formula, \[ f(x,y) = \sum_j\sum_k\sum_I\sum_J |I||J| \widetilde{\psi}_{j,k} (x,y,x_I,y_J)\psi_{j,k}\ast f(x_I,y_J) \] where \(\widetilde{\psi}_{j,k} (x,y,x_I,y_J)\in\mathcal{S}_F (\mathbb{R}^n,\mathbb{R}^m)\), and the series converges in \(\mathcal{S}_F(\mathbb{R}^n\times\mathbb{R}^m)\). To define \(\mathcal{S}_F (\mathbb{R}^n,\mathbb{R}^m)\) one first considers a space \(\mathcal{S}_M(\mathbb{R}^{n+m},\mathbb{R}^m)\) consisting of those Schwartz functions on \(\mathbb{R}^{n+m} \times \mathbb{R}^{m}\) such that the Schwartz semi-norms satisfy several special Lipschitz type conditions enumerated in this work. The space \(\mathcal{S}_F =\mathcal{S}_{F,M} (\mathbb{R}^n,\mathbb{R}^m)\) then consists of those \(f\) having extensions \(f^\sharp\in \mathcal{S}_M(\mathbb{R}^{n+m},\mathbb{R}^m)\) with \(f(x,y)=\int_{\mathbb{R}^m} f^\sharp (x, y-z,z)\, dz\) with norm defined by the infimum of the \(\mathcal{S}_M\)-norms of all such extensions of \(f\). The reproducing formula allows one to quantify \(L^p\) norms in terms of the square function \(S_d\). The space \(H^p_F(\mathbb{R}^n\times \mathbb{R}^m)\) consists of those \(f\) such that \(S_d\) belongs to \(L^p\). A key fact in proving the boundedness of Calderón-Zygmund type flag singular integrals is that the values of \(\psi_{j,k}\ast f\) do not oscillate too much on \(I\) and \(J\) satisfying the prescribed length conditions. The definitions of the flag singular integrals are as in the work of \textit{A. Nagel, F. Ricci} and \textit{E. M. Stein} [J. Funct. Anal. 181, No. 1, 29--118 (2001; Zbl 0974.22007)].
    0 references
    0 references
    flag singular integral
    0 references
    Hardy space
    0 references
    max-min inequality
    0 references
    0 references
    0 references