We know that $f\in C^{\infty}(0,\pi)$ and $f(0)=f(\pi)=0$. But $supp\ f=[0,\pi]$. From the definition, $H_{0}^{1}(0,\pi)$ is the clusear of $C_{0}^{\infty}(0,\pi)$ in $H^{1}(0,\pi)$. Since, $f$ is not defined outside $(0,\pi)$ there is not any point outside the support on which $f$ vanish. So how we can say $f\in H_{0}^{1}(0,\pi)$?


For $\epsilon>0$ small define $\phi_\epsilon(x)=\max(\sin x-\sin\epsilon,0)$, $0<x<\pi$. The support of $\phi_\epsilon$ is $[\epsilon,\pi-\epsilon]$. Its (weak) derivative is $$ \phi_\epsilon'(x)=\begin{cases} 0 & 0<x<\epsilon\\ \cos x & \epsilon \le x\le\pi-\epsilon\\ 0 & \pi-\epsilon<x<\pi \end{cases} $$ It is easy to see that $\phi_\epsilon\in H_0^1(0\,\pi)$ and $$ \lim_{\epsilon\to0}\|\sin x-\phi_\epsilon(x)\|_{H^1}=0. $$

  • $\begingroup$ Thank you very much @Julián Aguirre. But for each $\epsilon$, $\phi'_{\epsilon}$ is not continuous at $x=\epsilon$. I mean that $\phi_{\epsilon}$ is not in $C_{0}^{\infty}(0,\pi)$ or even in $C_{0}^{1}(0,\pi)$. $\endgroup$ – Albert May 8 at 12:05
  • $\begingroup$ But it is in $L^2$. $\endgroup$ – Julián Aguirre May 8 at 13:45
  • $\begingroup$ Yes, of course. But what happens for the definition? We have to find $C^{\infty}$ molifiers. $\endgroup$ – Albert May 8 at 13:56
  • $\begingroup$ Well, there are other equivalent definitions. $\endgroup$ – Julián Aguirre May 8 at 17:34
  • $\begingroup$ $H_0^1$ is a closed subspace of $H^1$. If a sequence in $H_0^1$ converges in the $H^1$ norm, then its limit is in $H_0^1$. All is needed is to show that each $\phi_\epsilon$ is in fact in $H^1_0$. This can be done by convolution with a compactly supported approximation of the identity. $\endgroup$ – Julián Aguirre May 9 at 11:31

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.