Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

If $f \in C^1 ([0,T] , L^2) \cap C^0 ([0,T] , W^{1,2} )$, then how can I conclude that $$ \left \| \frac{\partial f}{\partial t} \right \|_{L^\infty([0,T] \times \Bbb R^n )} < \infty ?$$ Here $f$ is defined on $[0,T] \times \Bbb R^n$ , and the notation $f \in C^1([0,T], L^2)$ means that $\| f(t) \|_{L^2 (\Bbb R^n)}$ is continuously differentiable on $[0,T]$, $W^{s,p}$ means the usual Sobolev space.

share|improve this question
    
Handwaving: $\partial f/\partial t = \mathrm df/\mathrm dt - \partial f/\partial x\ \partial x/\partial t$, thus the first inclusion gives (ess.) finiteness for the first term on the rhs, the second gives (ess.) finiteness for the second term, thus their difference is also essentially finite. –  filmor Oct 21 '12 at 15:45
    
@filmor Thank you, but here $x$ is not depend on $t$. $f = f(t,x_1 , \cdots , x_n)$. –  Ann Oct 21 '12 at 15:49
    
I don't think this is true: Take $f(t,x)=t\eta(x)|x|^{n-\alpha}$ with $\eta\in C_c^{\infty}(\mathbb{R}^n)$ a cut-off function, then for a suitable $\alpha>0$ we have $f\in W^{1,2}$, but $\eta(x)|x|^{n-\alpha}\notin L^{\infty}$. –  Jose27 Oct 21 '12 at 17:52
add comment

1 Answer 1

up vote 0 down vote accepted

Unless I am missing something, this does not seem to be correct. Let $n=3$, and define $$ F(x) = \|x\|^{-1/3} \quad\text{ for $\|x\|\le 1$,} $$ and extend it smoothly to a function with compact support on $\mathbb{R}^3$. Then $$ \|\nabla F(x)\| = \frac13 \|x\|^{-4/3} \quad\text{ for $\|x\|\le 1$},$$ so $F \in L^2(\mathbb{R}^3)$, $\nabla F \in L^2(\mathbb{R}^3)$ and thus $F\in W^{1,2}(\mathbb{R}^3)$. Now define $$f(t,x) = t F(x). $$ Then $$ \frac{\partial f}{\partial t} (t,x) = F(x), $$ so this example seems to satisfy the even stronger assumption that $f \in C^1([0,T],W^{1,2})$, but it $\frac{\partial f}{\partial t} = F \notin L^\infty$.

share|improve this answer
    
The Sobolev embedding doesn't give that inclusion: Take $f(x)=\log\log(1+|x|^{-1})$. –  Jose27 Oct 21 '12 at 18:06
    
Jose27, thanks, I deleted the wrong comment. –  Lukas Geyer Oct 21 '12 at 18:17
    
Thank you very much. But how it becomes if we add the additional assumption: $f(t,x)$ is bounded on $[0,T] \times \Bbb R^n$ ? –  Ann Oct 21 '12 at 19:42
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.