5
$\begingroup$

Is there a non vanishing real analytic vector field $X$ on $S^3$ such that $X$ has an attractor periodic orbit(An asymptotically stable periodic orbit) ? What about the smooth case?

$\endgroup$
1
  • $\begingroup$ Did you try: Use that S^3 is a Lie group. So you have plenty of non-vanishing vector fields. Find one with a periodic orbit O, then homotope the field nearby to point inwards toward the orbit O. $\endgroup$ Commented Apr 5, 2017 at 5:31

1 Answer 1

2
$\begingroup$

$S^3$ has the structure of a Lie group. Consider a left invariant vector field $X$. (Check that you can get closed orbits!) In a tubular neighborhood of a small amount of time around a closed orbit $o$, the vector field looks like a constant field on $\mathbb{R}^3\simeq D^2 \times I$. Consider the functionr $f = \beta\cdot d(\cdot,o)^2$, where $\beta$ is a smooth compactly supported bump function. The choice of unit vector $\nu$ pointing towards the central orbit $o$ from within this neighborhood is clear. Now set $X' = X + \epsilon\cdot f\cdot \nu$ for some choice of $\epsilon$.

For $X'$ we have the old orbit $o$, unperturbed. Moreover, nearby orbits will veer towards $o$. Thus now $o$ is a stable attractor. One checks that with $\epsilon$ small enough, $X'$ is still nonvanishing.

$\endgroup$
2
  • $\begingroup$ No problem. I believe this can be tidied up a bit! $\endgroup$ Commented Apr 5, 2017 at 12:09
  • $\begingroup$ So it is an smooth example yes? what about analytic case? $\endgroup$ Commented Feb 6, 2021 at 14:03

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.