3
$\begingroup$

The classical periodic Poincaré inequality states that if $u\in H^1(\mathbb T^n)$ is such that $\displaystyle\int_{\mathbb T^n} u(x)\ dx=0$ then $$\|u\|_{L^2(\mathbb T^n)}^2\leq C_d \|\nabla u\|_{L^2(\mathbb T^n)}^2,$$ for some constant $C$.

Can one drop the hypothesis $\displaystyle \int_{\mathbb T^n} u(x)\ dx=0$?

$\endgroup$
1
$\begingroup$

As literally stated? No, we can't. (I'm taking the volume of the torus to be 1 for convenience.) We can write any function as $u+c$ where $\int u = 0$; and replacing $u$ in your formula with $u+c$ the left hand side becomes $\|u\|^2_{L^2} + c\int u(x)dx + c^2 = \|u\|^2_{L^2} + c^2$. Then we obviously cannot bound this by $C_d\|\nabla (u+c)\|^2_{L^2}=C_d\|\nabla u\|^2_{L^2}$ because I can make $c$ arbitrarily large. But as long as you add a factor to counteract this on the RHS like, say, $$\|u\|^2_{L^2} \leq C_d \|\nabla u\|^2_{L^2} + \left(\int u(x)dx\right)^2,$$ then it's true.

$\endgroup$

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.