3
$\begingroup$

I have this two spaces $$C_{\theta}=\{u\in C(\overline{\Omega}), \sup (|x|^{\theta} |u(x)|)<\infty\}$$ with the norm $\displaystyle\|u \|_{\theta}=\sup_{\Omega}(|x|^{\theta} |u(x)|)$

and $$L_{1}^{p^*}=\{u ~~\text{measurable};~~\int_{\Omega} (|x|\cdot|u(x)|)^{p^*} dx<\infty\}$$ with the norm $\|u\|_{L^{p^*}_{1}}^{p^*}=\int_{\Omega} (|x|\cdot|u(x)|)^{p^*}dx$

Where $\Omega\subset\mathbb{R}^N$ is bounded with $0\in \Omega$ and $\theta>\frac{N}{p}>1,$ $p^*=\frac{pN}{N-p}$ and $N>p.$

I need a counter example to say that $C_{\theta}$ is not continuously embeded in $L^{p^*}_{1}$, so i'm searching a function which is in $C_{\theta}$ but not in $L^{p^*}_1$

Or a sequence $u_n$ which converge to $u$ un $C_{\theta}$ but not in $L^{p^*}_{1}$

Please help me thank you

$\endgroup$
5
  • $\begingroup$ What wuld a counter-example look like here? Sounds like you just need a proof. $\endgroup$ May 19, 2015 at 20:12
  • $\begingroup$ @ThomasAndrews if you have a direct prove i prefer , give me any idea $\endgroup$
    – Vrouvrou
    May 19, 2015 at 20:17
  • $\begingroup$ I don't have any proof. But "There does not exist an X" isn't usually be proven with a counterexample. $\endgroup$ May 19, 2015 at 20:18
  • $\begingroup$ but how i must do ? i need to be sure that there is no continuous embedding @ThomasAndrews $\endgroup$
    – Vrouvrou
    May 19, 2015 at 20:22
  • $\begingroup$ @ThomasAndrews can you tel me if the example given in the answer of this question can the counter example math.stackexchange.com/questions/1267849/… $\endgroup$
    – Vrouvrou
    May 19, 2015 at 20:44

1 Answer 1

2
+50
$\begingroup$

Let's try a canonical test function, $u(x)=|x|^\alpha$ and see if some value of $\alpha$ works. Suppose $\Omega=B(0,1)$. Then $$ \sup_{x\in\Omega}|x|^\theta|u(x)|=1\tag{1} $$ as long as $\alpha+\theta\ge0$. $$ \int_\Omega|xu(x)|^{p^*}\,\mathrm{d}x =\omega_{N-1}\int_0^1r^{(1+\alpha)p^*}r^{N-1}\,\mathrm{d}r\tag{2} $$ which is finite when $0\lt(1+\alpha)p^*+N=N\left[\frac{(1+\alpha)p}{N-p}+1\right]$. This is equivalent to $$ \frac{(1+\alpha)p}{N-p}\gt-1\iff\alpha\gt-\frac Np\tag{3} $$ For a counterexample, we would need $$ -\theta\le\alpha\le-\frac Np\tag{4} $$ Since $\frac Np\lt\theta$ by hypothesis, we can find an $\alpha$ strictly between $-\theta$ and $-\frac Np$. Thus, for $\alpha$ satisfying $(4)$, $u=|x|^\alpha$ and $\Omega=B(0,1)$ satisfy $(1)$, yet fail to have $(2)$ converge.


Since the question also wants $u\in C(\overline{\Omega})$, $C_\theta$ is not closed. We have to limit our functions to be in $C(\overline{\Omega})$. So, define $$ u_n(x)=\left\{\begin{array}{} |x|^\alpha&\text{if }|x|\ge\frac1n\\ \frac1{n^\alpha}&\text{if }|x|\lt\frac1n \end{array}\right. $$ So each $u_n$ is in $C(\overline{\Omega})$ and satisfies $(1)$. However, $u_n\to u$ monotonically as $n\to\infty$; thus, by Monotone Convergence, we have that $\|u_n\|_{\large L_{p^\ast}^1}\to\infty$.

$\endgroup$
31
  • $\begingroup$ thank you thank you thank you <3 <3 <3 $\endgroup$
    – Vrouvrou
    May 19, 2015 at 21:15
  • $\begingroup$ i can suppose that $diam(\Omega)=\rho$ and then say that $\sup_{x\in\Omega}|x|^\theta|u(x)|=\rho$ $\endgroup$
    – Vrouvrou
    May 19, 2015 at 21:18
  • $\begingroup$ There is one thing, when $\alpha <0$ $u\not in C$ which is a condition on $C_{\theta}$ so there is a problem $\endgroup$
    – Vrouvrou
    May 19, 2015 at 21:33
  • $\begingroup$ As I mentioned in $(1)$, as long as $\alpha+\theta\ge0$, we have that $u\in C_\theta$. So there is no problem as long as $\alpha\ge-\theta$, which is part of $(4)$. $\endgroup$
    – robjohn
    May 19, 2015 at 21:43
  • $\begingroup$ but when $\alpha$ is negative , $|x|^{\alpha}\notin C$ $\endgroup$
    – Vrouvrou
    May 19, 2015 at 21:47

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .