Help showing compactness of the support of a function in the Sobolev Space $W^{1,p}$

162 Views Asked by At

In Functional Analysis, Sobolev Spaces and Partial Differential Equations, by Haim Brezis, in the proof of Theorem 8.12, it is needed to show that the support of a function is compact.

The function is $u_n(x):= (1/n)G(nu(x))$, where $G \in C^1(\mathbb{R})$ and verifies $G(t) \leq |t|$ and: $$ G(t)=\left\{\begin{matrix} 0 & if \ |t| \leq 1\\ t & if \ |t| \geq 2 \end{matrix}\right.$$

And $u$ belongs to the Sobolev Space $W^{1,p}(I)$, where $I$ is a (possibly unbounded) interval in $\mathbb{R}$. Also, $u=0$ on $\partial(I)$. Recall that if $I$ is unbounded, then it verifies $$\lim_{x\in I, |x| \to \infty}{u(x)}=0 \quad \quad (1)$$.

What the proof says is:

$supp(u_n) \subset \{x \in I \ | \ |u(x)| \geq 1/n \}$, so, using the fact that $u=0$ on $\partial(I)$ and (1) then $supp(u_n)$ is compact.

And this is what I don't understand. Where is the marked in bold used? As far as I know, checking if a set is compact in $\mathbb{R}$ can be done by checking if it is closed (Which it is because the definition of the support) and bounded, which it is because it is a subset of a bounded set.

And the last fact is due to (1) if $I$ is unbounded and there is no assumption needed if $I$ is bounded. What am I missing?