5
$\begingroup$

Is there a compact Riemanian manifold $M$ not diffeomorphic to sphere or real or complex or quaternion projective space which admit a diffeomorphism $f$ with the property that $$\forall x \in M, \quad d(f(x), x)=diam(M)$$ where $d$ is the metric arising from the Riemannian metric and diameter of $M$ with repect to this metric is denoted by $diam(M)$

$\endgroup$
2
  • 4
    $\begingroup$ It's probably not what you were looking for, but the flat torus (which is $S^1 \times S^1$) will also have this property as well. In general, products of spheres and projective spaces will have diffeomorphisms satisfying this equation. $\endgroup$
    – Gabe K
    Sep 11 '18 at 22:16
  • $\begingroup$ @GabeK Yes that is right thank you. But what is an example of a compact manifold which does not satisfy the fixed point property does not admit such diffeomorphism? $\endgroup$ Sep 11 '18 at 22:25
10
$\begingroup$

The article by X. Liu and Sh. Deng "The antipodal sets of compact symmetric spaces" gives many examples, e.g. $\mathrm{SU}(2n)$, $\mathrm{Spin}(5)$, $\mathrm{Spin}(7)$,.... All those spaces have unique antipodal points which implies that the antipodal map is a diffeomorphism, see below. Here the antipodal points at $p$ are defined as $A(p)=\{ x\in M ~|~d(x,p)=\mathrm{diam}M\}$.

If a homogeneous space, equipped with a left-invariant metric has unique antipodal points then by compactness one gets a homeomorphism $f$ with the desired properties. By left-invariant of the metric the map $t\mapsto d(\exp(tX)x_0,\exp(tX)y_0)$ is constant so that $f(\exp(tX)x) = \exp(tX)f(x)$ which shows $f$ is differentiable. Here $\exp(X)$ denotes the isometry generated by the Killing field $X$.

$\endgroup$
1
  • $\begingroup$ Thank you very much for your very interesting and perfect answer. $\endgroup$ Sep 12 '18 at 21:18

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.