Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I need to show that there exists a unique $x_0\in \mathbb R$ such that $f(x_0)=x_0$ for the function defined as $|f(x)-f(y)| \leq L|x-y|$ such that with $0<L<1$ and $f: \mathbb R \to \mathbb R$.

Here is what I attempted to do: I will prove that $x_0$ must be unique by contradiction. Assume that there exists two distinct values, $x_1$ and $x_2$ such that $f(x_1)=x_1$ and $f(x_2)=x_2 $ and $x_1 \neq x_2$.

Observe that $$\begin{align}|f(x_1)-f(x_2)|\leq L|x_1-x_2| & \iff |x_1-x_2| \leq L|x_1-x_2| \\ &\iff \frac{|x_1-x_2|}{|x_1-x_2|}\leq L \\& \iff 1 \leq L \end{align}$$ which contradicts the definition of $L$ which says that $0<L<1$, and thus there must only exist one value $x_0$ such that $f(x_0)=x_0$.

The only thing is, this seems too easy and can't possibly be the complete solution. Should I have started by proving that $f(x_0)=x_0$ somehow or that the function is continuous?

Rearranging the given inequality gives $$\begin{align}\frac{|f(x)-f(y)|}{|x-y|} & \leq L \iff \\ \left|\frac{f(x)-f(y)}{x-y}\right| & \leq L\end{align}$$ which makes me think of the mean value theorem, and if I take the limit as $x \rightarrow x_0$ of both sides this reminds me of the definition of a function being differentiable if the limit exists and is finite.

Can someone tell me if I am approaching the problem correctly? Any hints would be greatly appreciated.

(Also I looked at a similar question where they also mentioned continuity, so now I feel that that needs to be addressed in my solution to the problem.)

share|improve this question
    
This seems to be pretty messy: first, your function is not defined as what you say, but fulfills that condition; second, the function must be continuous non-constant otherwise the claim is false; third, assuming the above, you still haven't proved that there exists such a fixed point for that point: you only proved that there exists at most one such fixed point... –  DonAntonio Apr 22 '13 at 16:19
    
Was the downvote because, as DonAntonio mentioned, my attempt at a solution was messy? –  user66807 Apr 22 '13 at 22:22
    
I don't think so, @user66807. Perhaps it is one of our serial downvoters, which downvote just because. Don't worry about that. –  DonAntonio Apr 22 '13 at 23:17

2 Answers 2

up vote 2 down vote accepted

Hint: You've shown that if a fixed point exists, it is unique. To prove existence, think of compositions and the fact that $L^n\to0$ as $n\to\infty$.

share|improve this answer
    
okay I understand that when $n \rightarrow \infty$, $L^n \rightarrow 0$, but I am not sure why we are using this fact. Here is what I came up with $$\begin{align}\frac{|f(x)-f(y)|}{|x-y|} & \leq L \iff \\ \left|\frac{f(x)-f(y)}{x-y}\right| & \leq L \iff \\ \left|\frac{f(x)-f(y)}{x-y}\right|^n & \leq L^n \end{align}$$ Taking limit of both sides gives $\lim_{n \to \infty} \left|\frac{f(x)-f(y)}{x-y}\right|^n \leq 0$ Am I on the right track? (I just found something on compositions in my textbook, I'll read it and see if I understand your hint better.) –  user66807 Apr 22 '13 at 17:19
    
@user66807: Instead, think about repeated compositions of a single point. That is, if $f^{(n)}(x)=f\circ\cdots\circ f(x)=x_n$, then the fact that $L^n\to0$ tells you this is a Cauchy sequence. In other words, for every $\varepsilon>0$, there exists $N>0$ such that for $m,n>N$, $|x_m-x_n||<\varepsilon$. Completeness of the real line tells you that a Cauchy sequence must converge, i.e., there is a point $x_0$ with $f(x_0)=x_0$. –  Clayton Apr 22 '13 at 17:33
    
It makes more sense now, but I'm still confused. I don't think I ever got the hang (or understanding) of how to prove a sequence is Cauchy.All I know is that if I prove a sequence is Cauchy then I have proved that it converges. This is how I was shown to show the Cauchy property. With $x_n=f^{(n)}(x)=f\circ\cdots\circ f(x)$ $n$ times, then $x_m=f^{(m)}(x)=f\circ\cdots\circ f(x)$ $m$ times. Assume $m\geq n$, then $|x_m-x_n||<\epsilon \iff |f^{(m-n)}(x)|<\epsilon$. Is this what you are doing as well? If so, how does $L^n\to0$ tell me this is Cauchy? Thanks a lot for your help, too. –  user66807 Apr 22 '13 at 17:58
    
Oh I think I just got it. Since $|x_m-x_n|=|f^{(m-n)}(x)|< L^n$ and since $L^n \to 0$ as $n \to \infty$ we have that $|f^{(m-n)}(x)|$ is bounded above by something that goes to $0$ and since $\epsilon > 0$ we have that $x_n$ is a Cauchy sequence. I understand that the sequence converges, but how does this tell us that $f(x_0)=x_0$? –  user66807 Apr 23 '13 at 3:45
    
@user66807: If it didn't, then we could keep iterating and that wouldn't be the limit point. –  Clayton Apr 23 '13 at 12:25

Hints about the existence of a fixed point with $\,f\,$ continuous:

1) Let $\,x\in\Bbb R\,$ be any point, and define $\,x_1=f(x)\;,\;x_n:=f(x_{n-1})\;,\;2\le n\in\Bbb N\,$

2) Show $\,\{x_n\}\,$ is a Cauchy sequence

3) Show that $\,x_0:=\lim\limits_{n\to\infty}x_n\,$ fulfills $\,f(x_0)=x_0\,$

share|improve this answer
    
Isn't the continuity a consequence of the function being Lipschitz? –  Clayton Apr 22 '13 at 17:38
    
Indeed so, yet I didn't see in the OP mentioning this (in fact, he even asks whether he has to prove continuity). That's why I remark it. –  DonAntonio Apr 22 '13 at 18:10
    
@Clayton That is because I don't know what Lipschitz is. –  user66807 Apr 22 '13 at 18:42

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.