Let $\Omega \subset \mathbb{R}^d$ be a domain of finite Lebesgue measure, not assumed to be smooth or bounded. Is it true that the embedding of, say, $W^{1,p}_0(\Omega)$ (Sobolev functions with zero boundary value) into $L^q(\Omega)$ is compact for $1/q > 1/p - 1/d$?

  • 6
    $\begingroup$ There are necessary and sufficient conditions on $\Omega$ for the compactness of such embedding. Check Mazy'a's book Sobolev Spaces, Springer Verlag, 2011, Section 5.5.2. $\endgroup$ – Liviu Nicolaescu Sep 28 '12 at 20:57
  • $\begingroup$ Since you want this only for functions with compact support, the embedding theorem follows directly from the one for functions on $\mathbb{R}^n$. $\endgroup$ – Deane Yang Mar 3 '13 at 23:04
  • $\begingroup$ @Deane, I am not sure I follow your argument here. There is no compact embedding for functions on $\mathbb{R}^n$. Could you axpand what do you mean here, I am sure I misunderstand something. $\endgroup$ – András Bátkai Mar 4 '13 at 9:01
  • $\begingroup$ @Andras There is no compact embedding on $\mathbb{R}^N$ but since the functions have zero boundary value they can be extended to a large ball where the embedding will hold, and therefore the restriction to $\Omega$ also embeds into $L^p$ compactly for $p<p^*$, as holds for smooth domains. $\endgroup$ – Daniel Spector Mar 4 '13 at 11:40
  • $\begingroup$ @Daniel, what disturbs me in this argument is only that the ball depends on the functions. $\endgroup$ – András Bátkai Mar 4 '13 at 16:56

You can find a rather general answer in the book of Adams and Fournier, Theorem 6.16. You do have the embedding you wish, under a mild measure theoretical assumption.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.