2

For my Algebraic Topology class I have solved the following exercise:

Let $x_0 \in S^1$ and let $f: S^1 \to S^1$ be a continuous map with $f(x_0)=x_0$. Suppose moreover that $f_*:\Pi_1(S^1,x_0) \to \Pi_1(S^1,x_0):[g] \mapsto k[g]$ for some natural number $k>2$. (So $f_*$ is multiplication with $k$.) Show that there are certainly $k-2$ other fixed points for $f$, besides $x_0$.

Now I need to give an example of such a map $f$ with more than $k-1$ fixed points. I don't have to work it out in a strict mathematical way (a description in words is enough). Can someone help me to find such an example?

1 Answers1

2

For the second part, note that it is easy to modify $f$ so that it is identity in a small neighborhood of $x_0$, thus giving it infinitely many fixed points.

For the first part: we may assume $x_0=1$ and thus there exists a continuous function $\tau: \mathbb{R} \rightarrow \mathbb{R}$ such that $f(e^{it})=e^{i\tau(t)}$ and $\tau(0)=0$.

Since $f$ induces in the fundamental group the multiplication by $k$, $\tau(2\pi)=2k\pi$.

Now, define, for $1 \leq l < k-1$, $g_l(t)=-2l\pi+f(t)-t$. $g_l$ is continuous, negative at $0$, positive at $2\pi$, so has a root $0 < x_l < 2\pi$, and $e^{ix_l}$ is a fixed point of $f$. Besides, $l \longmapsto f(x_l)-x_l=2l\pi$ is injective, so the $0<x_l<2\pi$ are pairwise distinct, therefore so are the $e^{ix_l}$.

Aphelli
  • 34,439
  • Thank you for your answer. Actually I already solved the first part, I just need to find an example with more than $k-1$ fixed points. I'm not sure what you mean by modifying $f$. I know that, for example, $f: S^1 \to S^1: z \mapsto z^k$ is a function with $k-1$ fixed points and such that $f_*$ is multiplication with $k$. Would it be possible to modify this function to get infinitely many fixed points such that the induced homomorphism is still multiplication with $k$? – mathstudent Jan 03 '20 at 17:39
  • Yes. Just take eg $f(e^{it})=e^{ig(t)}$ for $-\pi \leq t \leq pi$, where $g(t)=c(t)t$ for $c$ continuous, $c(t)=1$ if $|t| < \pi/4$, $c(t)=k$ if $|t| > 3\pi/4$. – Aphelli Jan 03 '20 at 19:50
  • 1
    @Mindlack shouldn't it say that $\tau(2\pi) = 2k\pi$? – Mee98 Jan 18 '20 at 18:15
  • @Mee98: I edited, thank you. – Aphelli Jan 18 '20 at 22:00