Weak type estimates of variable kernel fractional integral and their commutators on variable exponent Morrey spaces (Q2323461)

The paper is devoted to the boundedness of the fractional integral operators with variable kernels on the variable exponent weak Morrey spaces. Let \(\Omega\) be a function on \(\mathbb{R}^n\times \mathbb{R}^n\) such that \(\Omega(x,z')\in L^\infty(\mathbb{R}^n)\times L^s(\mathbb{S}^{n-1})\) with some \(s\in(1,\infty]\), \(\Omega(x,\mu z)=\Omega(x,z)\) and \(\int_{\mathbb{S}^{n-1}} \Omega(x, z')\,dz'=0\) for any \((x,z)\in \mathbb{R}^n \times \mathbb{S}^{n-1}\) and \(\mu>0\), where \(z'=z/|z|\). For \(\alpha\in(0,n)\), the fractional integral operator \(T_{\Omega,\alpha}\) related to \(\Omega\) is given by \[T_{\Omega,\alpha}(f)(x)=\int_{\mathbb{R}^n} \frac{\Omega(x,x-y)}{|x-y|^{n-\alpha}}f(y)\,dy,\] and \(T^m_{\Omega,\alpha,b}\) means the \(m\)-order commutator generated by \(T_{\Omega,\alpha}\) and \(b\). Let \(L^{p(\cdot)}(\mathbb{R}^n)\) and \(WL^{p(\cdot)}(\mathbb{R}^n)\) denote the variable Lebesgue space and weak Lebesgue space. The variable Morrey space \(\mathcal{M}_{p(\cdot),u}(\mathbb{R}^n)\) and the variable weak Morrey space \(W\mathcal{M}_{p(\cdot),u}(\mathbb{R}^n)\) with a function \(u(x,r)\) on \(\mathbb{R}^n\times(0,\infty)\) consist, respectively, of all functions \(f\) such that \[\|f\|_{\mathcal{M}_{p(\cdot),u}(\mathbb{R}^n)}=\sup_{z\in\mathbb{R}^n,r>0}\frac1{u(z,r)}\|\chi_{B(z,r)} \|_{L^{p(\cdot)}(\mathbb{R}^n)}\] and \[\|f\|_{W\mathcal{M}_{p(\cdot),u}(\mathbb{R}^n)}=\sup_{z\in\mathbb{R}^n,r>0}\frac1{u(z,r)}\|\chi_{B(z,r)} \|_{WL^{p(\cdot)}(\mathbb{R}^n)}\] are finite. In this paper, it is proved that, if \(p(\cdot)\) satisfies certain local regularity and global decay conditions, then \(T_{\Omega,\alpha}\) is bounded from \(\mathcal{M}_{p(\cdot),u}(\mathbb{R}^n)\) to \(W\mathcal{M}_{q(\cdot),u}(\mathbb{R}^n)\) when \(1\le p_-\le p_+<n/\alpha\), and from \(\mathcal{M}_{p(\cdot),u}(\mathbb{R}^n)\) to \(\mathcal{M}_{q(\cdot),u}(\mathbb{R}^n)\) when \(1<p_-\le p_+<n/\alpha\), where \(1/p(\cdot)-1/q(\cdot)=\alpha/n\). Moreover, for a function \(b\) in the Lipschitz space \(L_\beta(\mathbb{R}^n)\), the commutator \(T^m_{\Omega,\alpha,b}\) has the same boundedness of \(T_{\Omega,\alpha}\) as above. The corresponding boundedness of \(T^m_{\Omega,\alpha,b}\) when \(b\in\mathrm{BMO}(\mathbb{R}^n)\) is also obtained.