Quantitative weighted estimates for Rubio de Francia's Littlewood-Paley square function (Q2659017)

This paper considers the weighted estimates for the Rubio de Francia's Littlewood-Paley square function \(Tf\) associated with a family \(\Omega=\{\omega_k\}_{k\in\mathbb{Z}}\) of intervals in \(\mathbb{R}\) with finite overlapping: \[ Tf(x):=\left(\sum_{k\in\mathbb{Z}}|f*\check{\mathfrak1}_{\omega_k}(x)|^2\right)^{1/2}. \] The paper proves the following main result that gives a quantitative weighted bounds for the operator \(T\) on the \(w\)-weighted \(L^p\) space with \(w\in A_{p/2}\): \[ \|T\|_{L^p(w)\rightarrow L^p(w)}\lesssim[w]_{A_{p/2}}^{\max(\frac{1}{p-2},1)}, 2<p<\infty, \] where \(A_p\) is the Muckenhoupt class of weights \(w\) satisfying \([w]_{A_p}:=\sup_{Q}\langle w \rangle_Q (\langle w^{1-p'}\rangle_Q)^{p-1}<\infty\). In particular, for \(3\le p<\infty\), the exponent \(\alpha_p:=\max{(\frac{1}{p-2},1)}=1\) is sharp. Based on two existing results: Theorem 4.1 from [\textit{F. Bernicot} et al., Anal. PDE 9, No. 5, 1079--1113 (2016; Zbl 1344.42009)] and Theorem 4.3 from [\textit{D. Frey} and \textit{B. Nieraeth}, J. Geom. Anal. 29, No. 1, 247--282 (2019; Zbl 1415.42005)], the proof of the main result is reduced to the domaination estimate of the bilinear form \[ \langle Tf,\vec{g}\rangle:=\sum_{k=1}^N\int_{\mathbb{R}}f\ast\check{\mathfrak 1}_{\omega_k}(x)g_k(x)dx, \, \vec{g}=\{g_k\}_{k=1}^N. \] That is, \[ |\langle Tf, \vec{g}\rangle|\lesssim \sum_{I\in \mathcal{S}}|I|\langle f\rangle_{2,I}\langle |\vec{g}|\rangle_I \] with some sparse collection \(\mathcal{S}\) depending on \(f\) and \(\vec{g}\). The main body of the paper is then on the proof of the above domination estimate. It involves time-frequency analysis of the bilinear form via \(L^1\)-normalized wave packets associating with the tiling \(\mathbb{P}\) from the family \(\Omega\). The detailed proof involves localization estimate and decomposition of vectorial tree in \(\mathbb{P}\), sparse collection using \(p\)-Hardy-Littlewood maximal function, good tiles in \(\mathbb{P}\), and so on. The paper is concluded by some discussions and open problems related to the main results.