Dyadic representation and boundedness of nonhomogeneous Calderón-Zygmund operators with mild kernel regularity (Q2414687): Difference between revisions
From MaRDI portal
Removed claims |
Changed an Item |
||
Property / author | |||
Property / author: Ana Grau De La Herran / rank | |||
Normal rank | |||
Property / author | |||
Property / author: Tuomas P. Hytönen / rank | |||
Normal rank |
Revision as of 07:35, 12 February 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Dyadic representation and boundedness of nonhomogeneous Calderón-Zygmund operators with mild kernel regularity |
scientific article |
Statements
Dyadic representation and boundedness of nonhomogeneous Calderón-Zygmund operators with mild kernel regularity (English)
0 references
17 May 2019
0 references
In this paper the authors settle a linear dyadic representation formula and provide applications to \(T(1)\) and \(A_2\) theorems. In the following lines we make more precise this statement. Let \(\mu\) be a Borel measure of order \(n\) (not necessarily doubling), satisfying the growth condition \[ \mu(B(x,r))\leq C\cdot r^n \] for every \(x\in\mathbb{R}^d\) and every \(r>0\). Let \(T\) be a Calderón-Zygmund operator of order \(n\) on \(\mathbb{R}^d\), with respect to \(\mu\) that admits the kernel representation \[ Tf(x)=\int_{\mathbb{R}^d}K(x,y)f(y)d\mu(y), \quad x\notin \text{supp }f, \] where \(K\) satisfies the following conditions: \[ |K(x,y)|\leq\frac{C}{|x-y|^n} \] and \[ |K(x,y)-K(x^\prime,y)|+|K(y,x)-K(y,x^\prime)|\leq \frac{C}{|x-y|^n}\omega\Big(\frac{|x-x^\prime|}{|x-y|}\Big) \] whenever \(|x-x^\prime|\leq 1/2 |x-y|\). Here \(\omega\) is a modulus of continuity: an increasing and subadditive (\(\omega(a+b)\leq\omega(a)+\omega(b)\)) function with \(\omega(0)=0\). Assume that \(T\) also satisfies ``\(T(1)\)'' conditions in the local form, namely \[ \begin{cases} \Vert T1_Q\Vert_{L^2(\mu)} &\leq C\mu(Q)^{1/2},\\ \Vert T^\ast1_Q\Vert_{L^2(\mu)} &\leq C\mu(Q)^{1/2},\end{cases} \] for all cubes \(Q\subset\mathbb{R}^d\), where \(Q=\mathbb{R}^d\) is also regarded as a cube in the case that \(\mu(\mathbb{R}^d)<\infty\). Theorem: Under the above assumptions on \(T\) and \(\mu\), and for suitable test functions \(f,g\in L^2(\mu)\), the operator \(T\) admits a representation \[ \langle{Tf},{g}\rangle =\mathbb{E}\sum_{\substack{k\in\mathbb{Z} \\ k\neq 0}}\omega(2^{-|{k}|})\Big(\langle{R_k f},{g}\rangle+\langle{Q_k f},{g}\rangle\Big) +\mathbb{E}\langle{\Pi_{T1}f},{g}\rangle+\mathbb{E}\langle{f}\Pi_{T^\ast1},g\rangle, \] where: \(1)\) \(\mathbb{E}\) is the expectation over a random choice of dyadic systems on \(\mathbb{R}^d\), \(2)\) each \(\Pi_b\), \(b\in\{T1,T^\ast1\}\), is a dyadic paraproduct with \[ \|\Pi_b\|_{L^2(\mu)\to L^2(\mu)}\leq C, \] \(3)\) each \(R_k\) and \(Q_k\) is a dyadic operator with \[ \|R_k\|_{L^2(\mu)\to L^2(\mu)}\leq C,\qquad \|Q_k\|{L^2(\mu)\to L^2(\mu)}\leq C\sqrt{|k|}. \] If \(\mu\) is doubling, then each \(R_k\) and \(Q_k\) is a sum of \(O(|k|)\) dyadic shifts of complexity \(O(|k|)\). The most usual choices for \(\omega\) are the so called Dini conditions: \[\int_0^1\omega(t)\Big(1+\log\frac{1}{t}\Big)^\alpha\frac{dt}{t}<\infty.\] From the representation theorem above, the authors derive the following Corollary. Corollary: Under the above assumptions on \(T\) and \(\mu\), if the modulus of continuity satisfies the Dini condition with \(\alpha=\frac12\), then \(T\) acts boundedly on \(L^2(\mu)\). In the case of \(\mu\) being the Lebesgue measure, another consequence of the representation theorem is the following. Corollary: Let \(n=d\) and \(\mu\) be the Lebesgue measure on \(\mathbb{R}^d\). Under the above assumptions on \(T\), if the modulus of continuity satisfies the Dini condition with \(\alpha=2\), then \(T\) satisfies the \(A_2\) inequality \[ \|Tf\|_{L^2(w)}\leq C[w]_{A_2}\|f\|_{L^2(w)}. \] It is worth noting that this corollary does not recover the best possible class of Calderón-Zygmund operators which is \(\alpha=0\) due to \textit{M. T. Lacey} [Isr. J. Math. 217, 181--195 (2017; Zbl 1368.42016)]. However it still provides the result for the largest class of operators represented by a linear representation instead of a non-linear one.
0 references
representation theorems
0 references
dyadic representation formula
0 references