Boundedness for parabolic singular integral with rough kernels and its commutators on Triebel-Lizorkin spaces (Q644621): Difference between revisions

Let \(S^{n-1}=\{x\in\mathbb{R}^n:|x|=1\}\) denote the unit sphere on \(\mathbb{R}^n\). Let \(\alpha_1,\cdots,\alpha_n\) be fixed real numbers and \(\alpha_n\geq\alpha_{n-1}\geq\cdots\geq\alpha_1\geq1\). For each \(x=(x_1,\cdots,x_n)\in\mathbb{R}^n\), set \(x_1=\rho^{\alpha_1}\cos\varphi_1\cdots\cos\varphi_{n-2}\cos\varphi_{n-1},\) \(x_2=\rho^{\alpha_2}\cos\varphi_1\cdots\cos\varphi_{n-2}\sin\varphi_{n-1},\) \(\cdots\), \(x_{n-1}=\rho^{\alpha_{n-1}}\cos\varphi_1\sin\varphi_2,\) and \(x_n=\rho^{\alpha_n}\sin\varphi_1\) for some \(\rho >0\). Then \(dx=\rho^{\alpha-1}J(x')d\rho d\sigma\), where \(\alpha=\sum_{i=1}^n\alpha_i,\,x'\in S^{n-1}\), \(d\sigma\) is the element of area of \(S^{n-1}\) and \(\rho^{\alpha-1}J(x')\) is the Jacobian of the above transform. For \(\lambda>0\), let \(A_\lambda={\displaystyle\mathrm{diag}}[\lambda^{\alpha_1},\cdots,\lambda^{\alpha_n}]\) be a diagonal matrix. Suppose that \(\Omega\) is a real valued and measurable function defined on \(\mathbb{R}^n\). For any \(\lambda>0\) and \(x\in\mathbb{R}^n\), if \(\Omega(A_\lambda x)=\Omega(x)\), then \(\Omega(x)\) is called homogeneous of degree zero with respect to \(A_\lambda\). \(\Omega\) is said to satisfy the vanishing condition if \(\int_{S^{n-1}}\Omega(x')J(x')\,d\sigma(x')=0\). Denote the ellipsoid with center at \(x\) and radius \(r\) by \(E(x,r)=\{y:\rho(x-y)<r\}\) and by \(|E(x,r)|\) the Lebesgue measure of \(E(x,r)\). A measurable and locally integrable function \(b:\mathbb{R}^n\rightarrow\mathbb{R}\) is said to belong to parabolic BMO if \[ \|b\|_{\text{BMO}}:=\sup_\theta\eta_b(\theta) =\sup_\theta\sup_{r\leq\theta}\frac{1}{|E(r)|}\int_{E(r)}|b(y)-b_{E(r)}|\,dy<\infty, \] where \(E(r)\) is any ellipsoid in \(\mathbb{R}^n\) of radius \(r\) and \(b_{E(r)}=|E(r)|^{-1}\int_{E(r)}b(y)\,dy\). The parabolic singular integral operator is defined by \[ Tf(x)=\mathrm{p.\,v.}\int_{\mathbb{R}^n}\frac{\Omega(y)}{\rho(y)^\alpha}f(x-y)\,dy. \] The commutator generated by \(T\) and \(b\) is defined as follows: \[ [b,T]f(x)=\mathrm{p.\,v.}\int_{\mathbb{R}^n}\frac{\Omega(y)} {\rho(y)^\alpha}[b(x)-b(y)]f(x-y)\,dy. \] Choose a function \(\phi\in C_0^\infty(\mathbb{R}^n)\) such that \(0\leq\phi\leq1\), \({\displaystyle\mathrm{supp}}\,(\phi)\subset\{x\in\mathbb{R}^n,1/2\leq|x|\leq2\}\) and \(\phi(x)>c_0>0\) when \(3/5\leq|x|\leq5/3\). Let \(\phi_j(x)=\phi(2^jx)\) for all \(j\in\mathbb{Z}\) and let \(x\in \mathbb{R}^n\) satisfy \(\sum_{j\in\mathbb{Z}}\phi_j(x)^2=1\) when \(x\neq0\). Denote by \(S_j\) the convolution operator whose symbol is \(\phi_j\). Denote by \(\mathcal{P}(\mathbb{R}^n)\), \(\mathcal{S}(\mathbb{R}^n)\) and \(\mathcal{S}'(\mathbb{R}^n)\) the spaces of, respectively, polynomials, Schwartz functions and Schwartz distributions on \(\mathbb{R}^n\). For \(s\in\mathbb{R}\), \(1<p,\,q<\infty\), the homogeneous isotropic Triebel-Lizorkin space \(\dot{F}_p^{s,q}(\mathbb{R}^n)\) is defined by \[ \dot{F}_p^{s,q}(\mathbb{R})=\left\{f\in S'(\mathbb{R}^n)/\mathcal{P}(\mathbb{R}^n): \|f\|_{\dot{F}_p^{s,q}(\mathbb{R}^n)}= \left\|\left(\sum_{j\in\mathbb{Z}}2^{-jsq}|S_jf|^q\right)^{1/q}\right\|_p <\infty\right\}. \] For \(s\in\mathbb{R}\), \(0<p<\infty,\,0<q\leq\infty\), and an expansive dilation \(A\), define the anisotropic Triebel-Lizorkin spaces \(\dot{F}_p^{s,q}=\dot{F}_p^{s,q}(\mathbb{R}^n,A)\) associated with the expansive dilation \(A\) as the collection of all tempered distributions \(f\) (modulo polynomials) such that \[ \|f\|_{\dot{F}_p^{s,q}}=\left\|\left [\sum_{j\in\mathbb{Z}}\left(|\mathrm{det}A|^{js} |f\ast\psi_j|\right)^q\right]^{1/q}\right\|_p<\infty, \] where \(\psi_j(x)=|\mathrm{det}A|^j\psi(A^jx)\) and \(\psi\in\mathcal{S}(\mathbb{R}^n)\) satisfies the following conditions that \({\displaystyle\mathrm{supp}}\,\widehat{\psi}\subset[-\pi,\pi]^n\backslash\{0\}\) and \(\sup_{j\in\mathbb{Z}}|\widehat{\psi}((A^\ast)^j)(\xi)|>0,\, \mathrm{for\,\,all}\,\xi\in\mathbb{R}^n\backslash\{0\},\) where \(A^\ast\) is the adjoint (transpose) of \(A\) and the Fourier transform of \(f\) is defined to be \(\widehat{f}(\xi)=\int_{\mathbb{R}^n}f(x)e^{-\pi i(x,\xi)}\,dx.\) In this paper, the authors prove that if \(s\in\mathbb{R}\), \(1<p,\,q<\infty\), \(\lambda>1\), \(\alpha_i\geq1\), \(A={\displaystyle\mathrm{diag}}[\lambda^{\alpha_i},\cdots,\lambda^{\alpha_n}]\) is a diagonal matrix and \(\Omega\in L\log^+L(S^{n-1})\) satisfies the vanishing condition and is homogeneous of degree zero with respect to \(A\), then \(T\) is bounded on the anisotropic Triebel-Lizorkin spaces \(\dot{F}_p^{s,q}(\mathbb{R}^n,A)\), namely, \(\|Tf\|_{\dot{F}_p^{s,q}(\mathbb{R}^n,A)} \leq C(1+\|\Omega\|_{L\log^+L(S^{n-1})})\|f\|_{\dot{F}_p^{s,q}(\mathbb{R}^n,A)}\) and, moreover, \(T\) is also bounded on homogeneous isotropic Triebel-Lizorkin spaces \(\dot{F}_p^{s,q}(\mathbb{R}^n)\), namely, \(\|Tf\|_{\dot{F}_p^{s,q}(\mathbb{R}^n)} \leq C(1+\|\Omega\|_{L\log^+L(S^{n-1})})\|f\|_{\dot{F}_p^{s,q}(\mathbb{R}^n)},\) where \(C\) is a positive constant independent of \(f\). Moreover, if \(\Omega\in L(\log L)^2(S^{n-1})\), \([b,T]\) be defined as above, then \([b,T]\) is bounded on the homogeneous isotropic Triebel-Lizorkin spaces \(\dot{F}_p^{s,q}(\mathbb{R}^n)\), namely, \(\|[b,T]f\|_{\dot{F}_p^{s,q}(\mathbb{R}^n)} \leq C(1+\|\Omega\|_{L(\log L)^2(S^{n-1})})\|f\|_{\dot{F}_p^{s,q}(\mathbb{R}^n)},\) where \(C\) is a positive constant independent of \(f\).