End-point estimates and multi-parameter paraproducts on higher dimensional tori (Q986619): Difference between revisions

For a vector \(\vec t\in\mathbb R^{sd}\) and \(1\leq j\leq s\), let \(\rho_j(\vec t)=(t_j,t_{j+s},\ldots, t_{j+s(d-1)})\in\mathbb R^d\). And for \(1\leq j\leq d\), let \(\vec t_j=(t_{s(j-1)+1},\ldots, t_{js})\) so that \(\vec t=(\vec t_1,\ldots, \vec t_d)\). For a multiindex \(\alpha\in \mathbb N_0^{sd}\), let \(\rho_j(\vec \alpha)=(\alpha_j,\alpha_{j+s},\ldots, \alpha_{j+s(d-1)})\in\mathbb N_0^d\) and \(|\rho_j(\alpha)|= \alpha_j+\alpha_{j+s}+\cdots+\alpha_{j+s(d-1)}\). For a uniformly bounded function \(m\) on \(\mathbb R^{sd}\), which is smooth away from the origin, one says \(m\) is an \(s\)-parameter multiplier if \(|\partial ^\alpha m(\vec t)| \leq C\prod_{j=1}^s |\rho_j(\vec t)|^{-|\rho_j(\alpha)|}\) for all multiindices \(|\alpha|\leq sd(d+3)\). Given such a multiplier \(m\), one defines the associated multiplier operator \(\Lambda_m^{(s)}(f_1,\ldots, f_d):\mathbb T^s\to\mathbb C\) as \[ \Lambda_m^{(s)}(f_1,\ldots, f_d)(\vec x) =\sum_{\vec t\in \mathbb Z^{sd}}m(\vec t)\vec f_1(\vec t_1)\dots\vec f_d(\vec t_d) e^{2\pi i\vec x\cdot(\vec t_1+\cdots+\vec t_d)}. \] One of the author's main results is the following: For any \(s\)-parameter multiplier \(m\), it follows that \(\Lambda_m^{(s)}: L^{p_1}\times\cdots\times L^{p_d}\to L^p\) for \(1<p_j<\infty\) and \(1/p_1+\cdots+1/p_d=1/p\). If any or all of the \(p_j\) are equal to \(1\), this still holds with \(L^p\) replaced by weak \(L^p\) space \(L^{p,\infty}\) and \(L^{p_j}\) replaced by \(L(\log L)^{s-1}\). In particular, \(L(\log L)^{s-1}\times\cdots\times L(\log L)^{s-1}\to L^{1/d,\infty}\). The first part corresponds to the \(\mathbb R^{sd}\) case, given by \textit{C. Muscalu, J. Pipher, T. Tao} and \textit{C. Thiele} [Acta Math. 193, No.~2, 269--296 (2004; Zbl 1087.42016), and Rev. Mat. Iberoam. 22, No.~3, 963--976 (2006; Zbl 1114.42005)].