0

One can easily show that $W_0^{1,2}((a,b)) \subset C^0((a,b))$ for any finite interval $(a,b)$. Intuitively $W_0^{1,2}((a,b))$ should contain more functions than $C_0^0((a,b))$, but how to prove that?

I guess a function which is nowhere differentiable in $(a',b') \subset (a,b)$ should do the trick, but I have, sadly, no idea how to prove that.

Any hints?

Keba
  • 2,485
  • 13
  • 28

1 Answers1

2

Let $f$ be the cantor function on $[0,1]$. Let $g(x) = (1-x) f(x)$ on $[0,1]$. Then, $g$ is continuous and $g(0) = g(1) = 0$.

For almost every $x\in[0,1]$ we have $$ g'(x) = \underbrace{(1-x) f'(x)}_{=0} - f(x) = -f(x) < 0.$$ If $g$ is (locally) absolutely continuous, then $g$ would be strictly decreasing, which contradicts $g(0) = g(1)$. Thus, $g$ is not locally absolutely continuous and therefore not in $W^{1,2}[0,1]$.

Now, extend $g$ with 0 on $(-\epsilon, 1+\epsilon)$ for $\epsilon > 0$ if you like.

user251257
  • 9,229
  • Thanks for your fast answer. Why is every function in $W_0^{1,2}(\Omega)$ absolutely continuous? – Keba Aug 10 '15 at 19:40
  • @Keba in fact each $W^{1,1}$ function $f:I \to\R$ on an bounded interval $I$ is absolutely continuous. You can approximate $f$ by $C^\infty$ function w.r.t. the $W^{1,1}$ norm. On bounded interval $W^{1,2}$ functions are also $W^{1,1}$. – user251257 Aug 10 '15 at 19:49
  • Ah okay. Thanks again :) – Keba Aug 10 '15 at 19:51