Whitney coverings and the tent spaces \(T^{1,q}(\gamma)\) for the Gaussian measure (Q692561)

The authors introduce the notion of tent spaces associated to the Gaussian measure and prove the atomic decomposition for these spaces. This is done by introducing a technique for handling Whitney-type decompositions for the Gaussian harmonic analysis. They start by considering the admissible balls introduced by \textit{G. Mauceri} and \textit{S. Meda} [J. Funct. Anal. 252, No. 1, 278--313 (2007; Zbl 1136.46027)] in the setting of Gaussian measure and which enjoy the doubling property. They denote \(\mathcal B_\alpha\) the family of balls \(B(x,r)\) in \(\mathbb R^n\) such that \(0<r\leq \alpha\min\{1,\frac{1}{|x|}\}\) and write \(D=\{(x,t)\in \mathbb R^n\times (0,\infty): 0<t<\min\{1,\frac{1}{|x|}\}\}.\) Denoting by \(\gamma\) the Gaussian measure given by \(d\gamma(x)=(2\pi)^{-n/2}\exp(- |x|^2/2)dx\) they define, for \(\alpha\geq 1\) and \(1<q<\infty\), the Gaussian tent space \(T_\alpha^{1,q}(\gamma)\), as the completion of compactly supported continuous functions defined on \(D\) with the norm \[ \|f\|_{T^{1,q}(\gamma)}= \int_{\mathbb R^n}\left(\int\int_D \frac{\chi_{B(y,\alpha t)(x)}}{\gamma(B(y,t))} |f(y,t)|^qd\gamma(y)\frac{dt}{t}\right)^{1/q}d\gamma(x). \] A function \(a:D\to \mathbb C\) is called a \(T^{1,q}(\gamma)\) \(\alpha\)-atom if \(\text{supp}(a)\subset T_1(B)\) for some \(B\in \mathcal B_\alpha\) where \(T_1(B)\) stands for the tent with aperture \(1\) over \(B\), meaning \((y,t)\) such that \(d(y, \mathbb R^n\setminus B)\geq t\) and \(\|a\|_{L^q(D, d\gamma \frac{dt}{t})}\leq \gamma(B)^{-1/q}\). The main result establishes that, for \(\alpha>1\), all functions \(f\in T^{1,q}(\gamma)\) admit a decomposition into \(T^{1,q}(\gamma)\) \(\alpha\)-atoms, that is, there exists \((\lambda_n)\in \ell^1\) and a sequence \((a_n)\) of \(T^{1,q}(\gamma)\) \(\alpha\)-atoms such that \(f= \sum_n \lambda_n a_n\) with \(\sum_n |\lambda_n|\leq C \|f\|_{T^{1,q}(\gamma)}\). Using the atomic decomposition it is shown that \(T_{\alpha_1}^{1,q}(\gamma)=T_{\alpha_2}^{1,q}(\gamma)\) for \(1<\alpha_1<\alpha_2\).