Show that the convex neighborhood in a Riemannian Manifold are subset contractibles (to any of their points).

A subset $A$ of the differential manifold $M$ is contractible to the point $a\in A$ when, the aplications $id_{A}$ (identity in $A$) and $\gamma_{a}:x\in A\to a\in A$ are homotopic (with a point basis $a$).
I know that "geodesic convexity" is a natural generalization of convexity for sets and functions to Riemannian manifold, so in a set the convexity implies the homotopy. My idea was, take a riemannian metric and take exponential map in convex neighborhood, but I don't see these set are contractible to a point. Any hint!

  • $\begingroup$ To be clear, a convex subset of a Riemannian manifold is a subset where any pair of points are connected by unique minimizing geodesics which stay in the set? $\endgroup$ – Jason DeVito May 5 '16 at 17:30
  • $\begingroup$ Yes, anyway a definition convex neighborhoods: for each $p\in M$ exist a number $\alpha>0$ such that the "geodesic ball" $B_{\alpha}(p)$ is strongly convex $\endgroup$ – User166523 May 5 '16 at 22:38
  • 3
    $\begingroup$ Once you know that the exponential map $e=\exp_a : U \to A$ is a diffeomorphism, this is very simple - since $U$ is star-shaped, the map $\sigma(x,t) = tx$ is a contraction of $U$, and thus the composition $e(\sigma(e^{-1}(x),t))$ is a contraction of $A$. $\endgroup$ – Anthony Carapetis May 6 '16 at 12:41
  • $\begingroup$ @AnthonyCarapetis your answer has sense to me ( anyway I can't prove), you can explain more your answer. Thanks! $\endgroup$ – User166523 May 7 '16 at 20:52
  • $\begingroup$ @AnthonyCarapetis is just giving you an explicit contraction from $A$ to the point $a\in A$. He is just saying that the exponential map gives you a diffeo from $A$to a start-shaped set (there is a preferred point $z$ such that for any other point in $u\in U$ the line from $u$ to $z$ is contained in the set). Now this set is obviously contractible (move along this line), the map $\sigma$ is such a deformation retraction if you want. Then he is just conjugating it via $\exp$ to get a contraction for $A$ $\endgroup$ – Spotty May 12 '16 at 9:40

Let $U$ be a geodesically convex subset of $(M,g)$. Choose a point $p\in U$. For each $q\in U$ there exists a unique minimizing geodesic starting at $p$ and ending at $q$, i.e. there is a unique tangent vector $\hat{q}\in T_pM$ such that $$q = \exp_p\hat{q}.$$ It follows that there is a diffeomorphism $$\hat{U}:=\{\hat{q}\in T_pM\mid q\in U\}\cong U$$ via the exponential map (exercise: check it is indeed smooth with smooth inverse). The contraction is then induced by the map $I\times\hat{U}\to\hat{U}$ given by $$(t,\hat{q})\longmapsto t\hat{q}.$$ Exercise: Show that the set $\hat{U}$ is star-shaped with $0$ as a center. (Hint: a shorter piece of a minimizing geodesic is again a minimizing geodesic.)

| cite | improve this answer | |

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.