New gradient estimates for solutions to quasilinear divergence form elliptic equations with general Dirichlet boundary data (Q2279551)

Let \(\Omega\subset \mathbb{R}^n\) \((n\geq 2\)) be a bounded domain and let \(p\in (1,\infty)\). Moreover, let \(A:\Omega \times \mathbb{R}^n\rightarrow\mathbb{R}^n\) be a Carathéodory function satisfying, for some \(\Lambda_1,\Lambda_2>0\), \[ |A(x,\xi)|\leq \Lambda_1 |\xi|^{p-1}\text{ and }\langle A(x,\xi)-A(x,\eta), \xi-\eta\rangle \geq \Lambda_2(|\xi|^2+|\eta|^2)^{\frac{p-2}{2}}|\xi-\eta|^2, \] for almost every \ \(x\in \Omega\), and for all \(\xi,\eta\in \mathbb{R}^n\), and let \(F\in L^p(\Omega,\mathbb{R}^n)\) and \(\sigma\in W^{1,p}(\Omega)\). The authors consider the nonlinear boundary value problem \[ \begin{cases} \operatorname{div}(A(x,\nabla u))=\operatorname{div}(|F|^{p-2}F)&\text{ in }\Omega,\\ u=\sigma &\text{ on }\partial \Omega \end{cases} \tag{P} \] and establish fractional maximal gradient estimates for weak solutions of this problem. The complement of the domain \(\Omega\) is assumed to satisfy a \(p\)-capacity uniform thickness condition, which is weaker than the usual Reifenberg flatness condition. In particular, under the above assumptions, the authors prove that, for \(\alpha\in [0,n)\), any weak solution \(u\) of problem (P) satisfies the following estimate (\(\|\cdot\|_{q,s}\) stands for the norm in the Lorentz space \(L^{q,s}(\Omega)\)) \[ \|\mathbf{M}_\alpha (|\nabla u|^p)\|_{L^{q,s}(\Omega)}\leq C\|M_\alpha(|F|^p+|\nabla \sigma|^p)\|_{L^{q,s}(\Omega)}. \] Here \[ \mathbf{M}_\alpha(g(x)):=\sup_{\rho>0}\frac{\rho^\alpha}{|B_\rho(x)|}\int_{B_\rho(x)}|g(y)|dy, \ \ \ g\in L^1_{\text{loc}}(\Omega), \] is the fractional maximal function, \(0<s\leq \infty\), \(0<q<\frac{\Theta n}{p(n-\alpha)}\), with \(\Theta=\Theta(n,p,\Lambda_1,\Lambda_2)>1\), and the constant \(C\) depends only on \(n\),\(p\),\(q\),\(s\),\(\Lambda_1\),\(\Lambda_2\), \(\operatorname{diam}(\Omega),c_0,r_0\), with \(c_0,r_0>0\) involved in the \(p\)-capacity uniform thickness condition imposed on \(\Omega\). The approach is based on the optimal good-\(\lambda\) method, usually employed in measure data problems and here applied to a functional data problem.