2
$\begingroup$

I have to prove the following statements.

Let $U$ be an open subset of $\mathbb{R}^n$ and $u:U\rightarrow\mathbb{R}$ a locally Lipschitz function. Then if $U$ is convex, $$ \|Du\|_{L^{\infty}(\Omega)}=L(u, U). \quad(\star)$$ ($L(u, U)$ is the Lipschitz constant of $u$ in $U$). If $U$ is non-convex, then $$ \|Du\|_{L^{\infty}(U)}<L(u, U). $$

My attempt. If $U$ is non-convex (and so the general case), the result follows from the Rademacher's Theorem that states that if $u$ is locally Lipschitz continuous in $U$ then $u$ is differentiable a.e.. Is it right? If $U$ is convex, then, given $x, y\in\overline{U}$ (and hence the segment joining $x$ and $y$ is all contained in $\overline{U}$), by approximating $u$ by a sequence of mollifiers, we get $$ |u(x)-u(y)|\leq\|Du\|_{L^{\infty}(U)}|x-y| $$ Can we conclude that the equality $(\star)$ holds?

Thank You

$\endgroup$
2
$\begingroup$

In general, we have $\|Du(x)\| \le L(u,U)$ for any $x \in U$ at which the derivative is defined. To see this, let $h={Du(x)^T \over \|Du(x)\|}$ and note that $L(u,U) \ge \lim_{t \to 0} {u(x+th)-u(x) \over t} = Du(x)h = \|Du(x)\|$. Hence $L(u,U) \ge \operatorname{ess sup}_U \|Du\|$

By taking $U=(0,1)\cup(2,3)$ and defining $u=1_{(2,3)}$ we see that $\operatorname{ess sup}_U \|Du\| = 0 < L(u,U) = 1$, so the strictness is possible for non convex $U$.

For convex $U$, let $x,y \in U$ and then $u(x)-u(y) = \int_0^1 Du(y+t(x-y)) dt (x-y)$ (the derivative is defined ae.) from which we get $\|u(x)-u(y)\| \le \operatorname{ess sup}_U \|Du\| \|x-y\|$ hence $L(u,U) \le \operatorname{ess sup}_U \|Du\||$, and so we have equality.

$\endgroup$

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.