2
$\begingroup$

Let $(X,d)$ be a compact metric space and let $f : X → X$ be a function with the property that $d( f (x), f (y)) < d(x, y)$ whenever $x \neq y$. Show that $f$ has a fixed point, that is, there exists $x_0$ such that $f (x_0) = x_0$. (Hint: consider the function $g(x) = d(x, f (x))$ and argue that it attains its minimum and the minimum is 0)

Thank you all.

  • 0
    When you followed the given hint, at which point did you have a problem?2017-01-18
  • 0
    $X$ compact, then $X$ is bounded. So I know $x\in X, f(x)\in X, \exists k$ s.t $d(x,f(x))=k$ is min. But I do not know how to say the min is 0 based on $d(f(x),f(y))2017-01-18
  • 0
    Okay. So suppose you had $k > 0$. Then in particular $x \neq f(x)$. Which of the given properties might you want to apply in this situation?2017-01-18
  • 1
    $d(f(x),f(f(x)))2017-01-18
  • 0
    Hole in one. Bingo. Got it.2017-01-18
  • 0
    Wait, I did not get it. Why?2017-01-18
  • 0
    That's a contradiction to $$d(x,f(x)) = \min \{ d(y,f(y)) : y \in X\}.$$2017-01-18
  • 0
    You've found a smaller value than the minimum. PS: It is not that $X$ is bounded that guarantees the existence of a minimum as you seem to imply.2017-01-18
  • 0
    @AloizioMacedo Then how can I guarantee the existence of min?2017-01-18
  • 1
    That follows from the compactness of $X$ (and the continuity of $g$).2017-01-18
  • 0
    @DanielFischer Like $g: X\rightarrow K$ is continuous and $X$ is compact, then $K$ is compact, so we have the min?2017-01-18
  • 0
    Well, if $g(X) = K$, then $K$ is compact. If $g(X) \subsetneqq K$, then $K$ could be anything.2017-01-18
  • 0
    Ok, I think we know $g(X)= K$. One more question, how can I know $g$ is continuous? We do not know if $f$ is continuous.2017-01-18
  • 0
    We have $d(f(x),f(y)) \leqslant d(x,y)$ for all $x,y\in X$. That implies that $f$ is continuous.2017-01-18
  • 0
    @DanielFischer Hey Daniel. I still feel confused about how to prove $g$ is continuous. Even if now I know $f$ is continuous, how can I get that $g$ is continuous?2017-01-23
  • 0
    One part of it is that $d \colon X\times X \to \mathbb{R}$ is continuous, the other part is that $x \mapsto (x,f(x))$ is continuous. But if you don't know about the product topology yet, note that $$\lvert g(y) - g(x)\rvert = \lvert d(y,f(y)) - d(x,f(x))\rvert \leqslant d(y,x) + d(f(y),f(x)),$$ and that is $\leqslant 2 d(y,x)$.2017-01-23

1 Answers 1

1

You can take $x_{n+1} = f(x_n)$. Then since $X $ is compact there exist a convergent subsequence $x_{n_l} = y_l$ whose limit $x_0$ belongs to $X$. Now since $x_{n_l} = y_l$ converges to $x_0$ therefore $d(f(x),f(y))< d(x,y) $ $ \implies f(y_l) \rightarrow x_0$. hence $f(x_0) = x_0$