English

The Dirichlet problem for nonlocal operators with singular kernels: convex and nonconvex domains

Analysis of PDEs 2015-11-03 v2

Abstract

We study the interior regularity of solutions to the Dirichlet problem Lu=gLu=g in Ω\Omega, u=0u=0 in RnΩ\R^n\setminus\Omega, for anisotropic operators of fractional type Lu(x)=0+dρSn1da(ω)2u(x)u(x+ρω)u(xρω)ρ1+2s. Lu(x)= \int_{0}^{+\infty}\,d\rho \int_{S^{n-1}}\,da(\omega)\, \frac{ 2u(x)-u(x+\rho\omega)-u(x-\rho\omega)}{\rho^{1+2s}}. Here, aa is any measure on~Sn1S^{n-1} (a prototype example for~LL is given by the sum of one-dimensional fractional Laplacians in fixed, given directions). When aC(Sn1)a\in C^\infty(S^{n-1}) and gg is C(Ω)C^\infty(\Omega), solutions are known to be CC^\infty inside~Ω\Omega (but not up to the boundary). However, when aa is a general measure, or even when aa is L(Sn1)L^\infty(S^{n-1}), solutions are only known to be C3sC^{3s} inside Ω\Omega. We prove here that, for general measures aa, solutions are C1+3sϵC^{1+3s-\epsilon} inside Ω\Omega for all ϵ>0\epsilon>0 whenever Ω\Omega is convex. When aL(Sn1)a\in L^{\infty}(S^{n-1}), we show that the same holds in all C1,1C^{1,1} domains. In particular, solutions always possess a classical first derivative. The assumptions on the domain are sharp, since if the domain is not convex and the spectral measure is singular, we construct an explicit counterexample for which uu is \emph{not} C3s+ϵC^{3s+\epsilon} for any ϵ>0\epsilon>0 -- even if gg and Ω\Omega are CC^\infty.

Keywords

Cite

@article{arxiv.1502.00782,
  title  = {The Dirichlet problem for nonlocal operators with singular kernels: convex and nonconvex domains},
  author = {Xavier Ros-Oton and Enrico Valdinoci},
  journal= {arXiv preprint arXiv:1502.00782},
  year   = {2015}
}
R2 v1 2026-06-22T08:20:14.984Z