Which of the following is false $?$

$A.$ Any continuous function from $[0,1]$ to $[0,1]$ has a fixed point.

$B.$ Any homeomorphism from $[0,1)$ to $[0,1)$ has a fixed point.

$C.$ Any bounded continuous function from $[0,\infty)$ to $[0,\infty)$ has a fixed point .

$D.$ Any continuous function from $(0,1)$ to $(0,1)$ has a fixed point.

Now , if we take $f(x)=x^2$ , then $f((),1))\subset (0,1)$ but it does not show any fixed point. So, option $D$ is our false statement .

And option $A$ is a well known result .

So, turns out that , the statements in option $B$ and $C$ are both correct. I need help to prove them.

For $B$ , my thought is that the said homeomorphism can be extended to $[0,1]$ and the fixed point theorem will apply but we will need to show that the fixed point is not the point $x=1$. Is that right $?$ But I don't know $C$ .

Thanks for any help.

  • 1
    $\begingroup$ Yes, the homeomorphism can be extended to $[0,1]$, but that requires proof. For another way, what property distinguishes $0$ from the other points in $[0,1)$? (Thus every homeomorphism $[0,1) \to [0,1)$ must have $0$ as a fixed point. There may be others, but $0$ must be one.) For C, you could think about how the intermediate value theorem relates to the situation. $\endgroup$ – Daniel Fischer Dec 18 '15 at 10:43
  • $\begingroup$ A only is true if the function is surjective to me, maybe Im wrong. $\endgroup$ – Masacroso Dec 18 '15 at 11:14
  • 2
    $\begingroup$ Dear @Masacroso,A is always true. Consider $g(x)=f(x)-x$ and use Intermediate value theorem. $\endgroup$ – Dontknowanything Dec 18 '15 at 11:25
  • 1
    $\begingroup$ For part C, you can restrict the function from $[0,M]$ to $[0,M]$ where $M$ is an upper bound of the range. $\endgroup$ – Prince Kumar Dec 7 '16 at 11:20

For B consider that $0$ is the only point $p\in [0,1)$ such that $[0,1)\backslash \{p\}$ is a connected space, so a homeomorphism $f$ must have $f(0)=0.$ Another way is that $B=\{[0,d):0<d<1\}$ is a neighborhood base at $ 0$ such that $\forall b\in B\; (\bar b\backslash b$ has just one member.) No other point in $[0,1)$ has a nbhd base with this property, so $f(0)=0$.....For C if $f(0)=0$ we are done. If $f(0)>0$ the continuous function $g(x)=f(x)-x$ is positive at $x=0$ and negative when $x>\sup \{f(y) :y\geq 0\}$ so for some $x$ we have $ g(x)=0.$


HINT. The same proof based on the intermediate value theorem that you use to prove A works for D too. You don't even need $f$ to be bounded; any less-than-linear bound $f(x)\le Cx^\alpha,\ 0< \alpha < 1$ will suffice.


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.