I am not sure my question is research type, but I am sure I can find here an answer.

So we have the following theorem in the book of Lawrence Evans in PDE, 2nd edition pages 294-295:

Theorem 4 (Characterization of $W^{1,\infty}$). Let $U$ be open and bounded, with $\partial U$ of class $C^1$. Then $u: U\to \mathbb{R}$ is Lipschitz continuous iff $u\in W^{1,\infty}(U)$.

Now, I want to adapt this theorem to the case that $U=M$ is a compact manifold like the $n$dimensional torus, i.e its boundary isn't necessarily $C^1$.

How does that change the proof in Evans' book?

I think it only changes step 3 in the proof, other than that the same argument follows also for the torus, am I wrong here?


  • $\begingroup$ What is the problem here? If there is no boundary, things only get easier! $\endgroup$ – Michael Renardy Jun 5 '16 at 18:05
  • $\begingroup$ @MichaelRenardy so there's no need to use here step 3, the rest of the argument follows as usual? $\endgroup$ – Alan Jun 5 '16 at 18:10

You are right. This theorem and the extension theorem (the step you are referring to in the proof) holds for domains with Lipschitz boundary. You can find a proof of this in Evans and Gariepy's book on measure theory.

  • $\begingroup$ So in order to adapt the proof to my case with the torus i only need to abandon step 3 in Evans' PDE, correct? $\endgroup$ – Alan Jun 5 '16 at 17:05
  • $\begingroup$ Thanks, I checked the book you mentioned and it's indeed there. $\endgroup$ – Alan Jun 5 '16 at 20:34

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.