1
$\begingroup$

For any function $f(x)$ we denote $\bar{f}:=\frac{1}{\Omega}\int_\Omega f(x)\,dx$. Let $\Omega\subset \mathbb{R}^n$ be a bounded smooth domain and $u(x)> 0$ be a smooth function defined on $\Omega$. I want to know whether the following is true: $$\int_\Omega |\ln u(x)-\ln \bar{u}|^2dx\le C\int_\Omega |\nabla \ln u(x)|^2dx,$$ for some $C$ independent of $u$.

I think this general result must be fault, is it true when $\Omega$ is convex?

If it is true for convex $\Omega$, I have a further question: use "$F$" instead of "$\ln$" which kinds $F$ support this type inequality.

Remark: I missed a condition that $u$ have an upper bound $M$, $C$ could depend on $M$.

$\endgroup$
  • 2
    $\begingroup$ I do not think an upper bound helps. Note that both sides of the inequality are unchanged if u is replaced by Ku. So if there is a constant that works if u is bounded above by M, that same constant will work for any u that is bounded above, and then a limit argument would extend it to unbounded u as in my counterexample. $\endgroup$ – Michael Renardy Jan 16 '14 at 23:21
5
$\begingroup$

It is false for a ball. In three dimensions, consider $u=r^a$, where $a>-3$ and let $\Omega$ be the unit ball. We find $\bar u=3/(3+a)$. As a approaches -3, the right hand side of your inequality remains bounded, the left hand side does not.

$\endgroup$
  • $\begingroup$ Your answer is correct and very nice. Thank you. But I very sorry, I had make a mistake that what I really concern, there is one more condition: $u$ has a uniform upper bound $M$. Now I edit the problem. $\endgroup$ – user44565 Jan 16 '14 at 20:38
  • $\begingroup$ The domain is bounded, why the left hand side of your example is unbounded? $\endgroup$ – user44565 Jan 16 '14 at 21:41
  • $\begingroup$ Because $\tilde u\to\infty$ as $a\to -3$ and $|\ln u|^2$ remains integrable. $\endgroup$ – Michael Renardy Jan 16 '14 at 22:28
  • $\begingroup$ Oh, I see. Sorry, I alway keep $u\le M$ in my mind! Yes, you are right, but what I really concern, need $u$ has an upper bound. $\endgroup$ – user44565 Jan 16 '14 at 22:32

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.