The boundedness for commutators of a class of maximal hypersingular integrals with variable kernels (Q549928)

Let \(S^{{n-1}}\) be the unit sphere in \(\mathbb{R}^{n}(n\geq 2)\) with normalized Lebesgue measure \(d\sigma\). A function \(\Omega\) defined on \(\mathbb{R}^{n}\times\mathbb{R}^{n}\) is said to be in \(L^{\infty}(\mathbb{R}^{n})\times L^{q}(S^{{n-1}})\), \(q\geq1\), if \(\Omega\) satisfies the following conditions: 1. for any \(x,z\in \mathbb{R}^{n}\) and \(\lambda>0\), \(\Omega(x,\lambda z)=\Omega(x,z)\), 2. \(\|\Omega\|_{L^{\infty}(\mathbb{R}^{n})\times L^{q}(S^{{n-1}})}\equiv \sup_{x\in\mathbb{R}^{n}}\big(\int_{S^{{n-1}}}|\Omega(x,z)|^{q}d\sigma(z)\big)^{1/q}<\infty.\) For \(\gamma\geq 0\), \(\Omega\in L^{\infty}(\mathbb{R}^{n})\times L^{1}(S^{{n-1}})\) and a function \(b\) let us consider the the hypersingular integral operator \(T_{\gamma,\Omega}\), the diadic maximal hypersingular integral operator \(T^\ast_{\gamma,\Omega}\) and the diadic maximal operator \([b,T_{\gamma,\Omega}]^\ast\) associated with the commutator \([b,T_{\gamma,\Omega}]\) that are defined as follows \[ T_{\gamma,\Omega}(f)(x)=p.v.\; \int_{\mathbb{R}^{n}}\frac{\Omega(x,x-y)} {|x-y|^{n+\gamma}}f(y)dy, \] \[ T^\ast_{\gamma,\Omega}(f)(x)=\sup\limits_{l\in\mathbb{Z}}\big|\int_{|x-y|>2^l}\frac{\Omega(x,x-y)} {|x-y|^{n+\gamma}}f(y)dy\big|, \] \[ [b,T_{\gamma,\Omega}]^\ast(f)(x)=\sup\limits_{l>0}\big|\int_{|x-y|>2^l}\frac{\Omega(x,x-y)} {|x-y|^{n+\gamma}}(b(x)-b(y))f(y)dy\big|. \] Let \(\gamma\in\mathbb{R}\) and \(1<p<\infty\). The homogeneous Sobolev space \(\dot{L}_{\gamma}^{p}(\mathbb{R}^{n})\) is defined as the space of all tempered distributions \(f\) for which the expression \((|\cdot|^{\gamma}\widehat{f})^{\vee}\) is a function in \(L^p(\mathbb{R}^{n})\), where \(\widehat{g}\) and \(g^{\vee}\) denote the Fourier transform and inverse Fourier transform of \(g\). The norm in \(\dot{L}_{\gamma}^{p}(\mathbb{R}^{n})\) is defined as follows: \(\|f\|_{\dot{L}_{\gamma}^{p}}=\|(|\cdot|^{\gamma}\widehat{f})^{\vee}\|_{L^p}.\) In the paper the following theorems are proved: Theorem 1. Let \(\gamma\geq 0\). If \(\Omega\in L^{\infty}(\mathbb{R}^{n})\times L^{q}(S^{{n-1}})\) with \(q>\max \big\{1,\frac{2(n-1)}{n+2\gamma}\big\}\), and for any spherical harmonic polynomial \(\Upsilon_m\) with degree \(\leq [\gamma]\), \[ \int_{S^{{n-1}}}\Omega(x,z)\Upsilon_m (z)d(z)=0,\;\;x\in\mathbb{R}^{n},\;\; \tag{1} \] then there is a constant \(C>0\) such that \(\|T^\ast_{\gamma,\Omega}(f)\|_{L^2}\leq C \|f\|_{\dot{L}_{\gamma}^{2}}\). Theorem 2. Let \(0<\gamma<\frac{n}{2}\) and \(b\in\dot{L}_{\gamma}^{n/\gamma}\). If \(\Omega\in L^{\infty}(\mathbb{R}^{n})\times L^{q}(S^{{n-1}})\) with \(q>\max \big\{1,\frac{2(n-1)}{n+2\gamma}\big\}\) and \(\Omega\) satisfies condition (1) from Theorem 1, then there is a constant \(C>0\) such that \(\|[b,T_{\gamma,\Omega}]^{\ast}(f)\|_{L^2}\leq C \|b\|_{\dot{L}_{\gamma}^{n/\gamma}} \|f\|_{\dot{L}_{\gamma}^{2}}\).