English

Global Sobolev inequalities and Degenerate P-Laplacian equations

Analysis of PDEs 2018-01-30 v1

Abstract

We prove that a local, weak Sobolev inequality implies a global Sobolev estimate using existence and regularity results for a family of pp-Laplacian equations. Given ΩRn\Omega\subset\mathbb{R}^n, let ρ\rho be a quasi-metric on Ω\Omega, and let QQ be an n×nn\times n semi-definite matrix function defined on Ω\Omega. For an open set ΘΩ\Theta\Subset\Omega, we give sufficient conditions to show that if the local weak Sobolev inequality % (\fintBfpσdx)1pσC[r(B)\fintBQfpdx+\fintBfpdx]1p \Big(\fint_B |f|^{p\sigma}dx\Big)^\frac{1}{p\sigma} \leq C\Big[ r(B)\fint_B |\sqrt{Q}\nabla f|^pdx + \fint_B |f|^pdx\Big]^\frac{1}{p} holds for some σ>1\sigma>1, all balls BΘB\subset \Theta, and functions fLip0(Θ)f\in Lip_0(\Theta), then the global Sobolev inequality (Θfpσdx)1pσC(ΘQf(x)pdx)1p \Big(\int_\Theta |f|^{p\sigma}dx\Big)^\frac{1}{p\sigma} \leq C\Big(\int_\Theta |\sqrt{Q}\nabla f(x)|^pdx\Big)^\frac{1}{p} also holds. Central to our proof is showing the existence and boundedness of solutions of the Dirichlet problem {\mxp,τu=φinΘu=0inΘ, \begin{cases} \mx_{p,\tau} u & = \varphi \text{in} \Theta \\ u & = 0 \text{in} \partial \Theta, \end{cases} where \mxp,τ\mx_{p,\tau} is a degenerate pp-Laplacian operator with a zero order term: \mxp,τu=div(Qup2Qu)τup2u. \mx_{p,\tau} u = \text{div}\Big(\big|\sqrt{Q} \nabla u\big|^{p-2}Q\nabla u\Big) - \tau |u|^{p-2}u.

Keywords

Cite

@article{arxiv.1801.09610,
  title  = {Global Sobolev inequalities and Degenerate P-Laplacian equations},
  author = {David Cruz-Uribe and Scott Rodney and Emily Rosta},
  journal= {arXiv preprint arXiv:1801.09610},
  year   = {2018}
}