1

Given $\Omega$ a bounded regular open set in $\Bbb R^d$ we consider

$C_c^\infty(\Omega)$ the space of smooth functions compactly supported in $\Omega$. For $1<p<\infty $ Let's denote by $W^{1,p}(\Omega)$ be the standard Sobolev space. It is common to denote the closure of $C_c^\infty(\Omega)$ in $W^{1,p}(\Omega)$ by $$W_0^{1,p}(\Omega):=\overline{ C_c^{\infty}(\Omega)}^{~~~~W^{1,p}(\Omega)}$$

Note that, by the trivial zero extension the space $C_c^\infty(\Omega)$ can be seen as the space of smooth functions in $\Bbb R^d$ which are compactly supported in $\Omega$ and hence it can be seen as subspaces of $W^{1,p}(\Bbb R^d)$.

I would like to compare the spaces $$W_0^{1,p}(\Omega):=\overline{ C_c^{\infty}(\Omega)}^{~~~~W^{1,p}(\Omega)}$$ and $$\overline{ C_c^{\infty}(\Omega)}^{~~~~W^{1,p}(\Bbb R^d)}$$

More generally if $0<s<1$ is the fractional oder how can we compare the following fractional Sobolev spaces?

$$W_0^{s,p}(\Omega):=\overline{ C_c^{\infty}(\Omega)}^{~~~~W^{s,p}(\Omega)}~~~ and ~~~~\overline{ C_c^{\infty}(\Omega)}^{~~~~W^{s,p}(\Bbb R^d)}$$

Definition: $W^{s,p}(\Omega)$ is the space of class of functions $u$ in $L^p(\Omega)$ such that

$$[u]^p_{{W^{s,p}(\Omega)}}:=\iint\limits_{\Omega\Omega}\frac{|u(x)-u(y)|^p}{|x-y|^{d+sp}} dxdy<\infty. $$ Which turns out to be a Banach space endowed with the natural norm $$\|u\|^p_{W^{s,p}(\Omega)}=\|u\|^p_{L^{p}(\Omega)}+[u]^p_{{W^{s,p}(\Omega)}}$$

NB As the second question might less obvious good answer to this may deserve some bounty accordingly. But I will be okay if one just answers the first question.

Guy Fsone
  • 25,237
  • @daw In fact the first question is just an appetiser of the second question. that argument you are appealing does not apply to the fractional case – Guy Fsone Mar 14 '18 at 14:56
  • https://math.stackexchange.com/questions/3985046/a-doubt-on-the-sobolev-space-w-01-p-omega?noredirect=1&lq=1 – Guy Fsone Nov 10 '22 at 11:39

1 Answers1

1

These two spaces are the same

  1. if $0<s<1/p$ or
  2. if $\Omega$ is Lipschitz and $s-1/p$ is not an integer.

This is Corollary 1.4.4.5 in the book 'Elliptic problems in nonsmooth domains' by Grisvard. The proof uses a density result that is claimed but not proven in the book. The density result (in more general form) is Theorem 6 in

A. Fiscella, R. Servadei, and E. Valdinoci. Density properties for fractional Sobolev spaces. Ann. Acad. Sci. Fenn. Math. 40.1 (2015), pp. 235–253. doi: 10.5186/aasfm.2015.4009.

daw
  • 54,637
  • 2
  • 44
  • 85