Let $\Omega$ an open and bounded domain in $R^n$ . Let $u \in W^{1,p} (\Omega) \cap L^{\infty}(\Omega)$ $(2 \leq p < \infty)$ . Let $B \subset \subset \Omega$ a open ball and consider $u_1$ the restriction of $u$ on $B$. Then we have $u_1 \in W^{1,p} (B) \cap L^{\infty}(B)$. Let $T : W^{1,p} (B) \rightarrow L^{p}(\partial B) $ the trace operator. Can I say that $T(u_1) \in L^{\infty}(\partial B)$ ? Intuitively this is true .. But I dont know how to proof or counter example... Someone can give me a help? I am not good with the trace operator...

thanks in advance

  • $\begingroup$ Please have a look here. It gonna helpful. math.stackexchange.com/questions/1079409/… $\endgroup$
    – spatially
    Dec 25, 2014 at 0:01
  • $\begingroup$ Why do you ask all of your questions using the word intuitive? $\endgroup$
    – Tomás
    Dec 26, 2014 at 11:53
  • $\begingroup$ Hi Tomás . I write this when the affirmation seems to be true in simple cases (for example : smooth functions) $\endgroup$ Dec 26, 2014 at 14:51

1 Answer 1


Let $\phi_n\in C^\infty(B)$ be a sequence satisfying: $\phi_n\to u_1$ in $W^{1,p}(B)$ and $\|\phi_n\|_\infty\le M$ for all $n$, where $M$ is a positive constant (see here).

Note that $\phi_n\to u_1$ in $W^{1,p}(B)$, therefore, $(\phi_n-M)^+\to (u_1-M)^+$ in $W^{1,p}(B)$.

By one hand, $T((\phi_n-M)^+)=0$ for all $n$. On the other hand, once $(\phi_n-M)^+\to (u_1-M)^+$ in $W^{1,p}(B)$, we have that $T((\phi_n-M)^+)\to T((u_1-M)^+)$, thus $T((u_1-M)^+)=0$.

The last paragraph does implies that $$\inf\{\alpha\in\mathbb{R}:\ T((u_1-\alpha)^+)=0\}<\infty.$$

To conclude, use this post.


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.