Skip to main content
10 events
when toggle format what by license comment
Apr 2, 2019 at 22:28 comment added Mateusz Kwaśnicki I bet the answer is "yes", although in an integral form rather than point-wise. I do not know the details, though. "Higher-order Hardy's inequality" (in Lipschitz domains) seems to be a good search phrase.
Apr 2, 2019 at 22:06 comment added user123672 @MateuszKwaśnicki I see. Thank you. Then can we say anything at all about the boundary decay of $u \in H^2_0(\Omega)$?
Apr 2, 2019 at 21:58 comment added Mateusz Kwaśnicki I mean: $\delta < 1/2$. Then $u(x) = (1 - x^2)^{2 - \delta} \mathbb{1}_{(-1,1)}(x)$ satisfies $u, u', u'' \in L^2$, and therefore $u \in H^2(\mathbb{R})$. And since $u = 0$ outside $(-1, 1)$, in fact $u \in H^2_0((-1,1))$.
Apr 2, 2019 at 21:42 comment added user123672 @MateuszKwaśnicki What do you mean by "if $\delta$ is small enough"?
Apr 2, 2019 at 21:40 history edited user123672 CC BY-SA 4.0
added 1 character in body
Apr 2, 2019 at 21:38 comment added user123672 @MateuszKwaśnicki You're right: I've edited the question. We can assume whatever smoothness is necessary (although I hope none).
Apr 2, 2019 at 21:38 comment added Mateusz Kwaśnicki As for the upper bound, $|u(x)| \le C \operatorname{dist}(x, \partial \Omega)^2$, this is not true even in dimension $1$: $(1-x^2)^{2-\delta}$ is in $H_0^2((-1, 1))$ if $\delta > 0$ is small enough.
Apr 2, 2019 at 21:37 history edited user123672 CC BY-SA 4.0
deleted 3 characters in body
Apr 2, 2019 at 21:35 comment added Mateusz Kwaśnicki Any smoothness assumptions on the boundary? Also, in any case there is no hope for the lower bound (if $\asymp$ stands for the ratio being bounded between two positive constants).
Apr 2, 2019 at 21:31 history asked user123672 CC BY-SA 4.0