5
$\begingroup$

Is every convex sphere (in the sense of Alexandorff, which is the boundary of some convex body in $\mathbb{R}^3$) with Alexandorff curvature $\geq 1$, admitting a bijective map to the unit round sphere in $\mathbb{R}^3$ with Lipschitz constant $\leq 1$?

$\endgroup$

1 Answer 1

4
$\begingroup$

The answer is positive, without the assumption on curvature (of course we do have curvature $\ge 0$, for any convex). Let $o$ be an interior point of the body $B$ whose boundary $\partial B$ is your convex sphere, and let $r$ be small enough that $S(o,r) \subset \operatorname{int}(B)$. Consider the radial projection $\mathbb{R}^3\to B(o,r)$, which is $1$-Lipschitz. All we have to prove is that its restriction $\pi$ from $\partial B$ to $S(o,r)$ is invertible and has Lipschitz inverse.

Every ray issued from $o$ interects $\partial B$, so that $\pi$ is onto. If such a ray did intersect $\partial B$ in several point, then $o$ would have to be a boundary point, which is excluded, so that $\pi$ is one-to-one.

Now, that $\pi^{-1}$ is not Lipschitz is equivalent to say that we can find rays $\gamma$ issued from o and support hyperplanes $H$ at the intersection point $\gamma\cap \partial B$ making arbitrarily small angle. If this where the case, taking a sequence $(\gamma_n,H_n)_n$ with angle going to $0$, we could extract a convergent subsequence. This would give in the limit a ray $\gamma_\infty$ and a support hyperplane $H_\infty$ such that $\gamma_\infty \subset H_\infty$. Then $o\in \gamma_\infty\subset H_\infty$ in contradiction with $o$ being an interior point.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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