2
$\begingroup$

The problem is as follows:

Suppose that $f$ is continuous on $[0,1]$ and $f(0) = f(1)$. Let $n$ be any natural number. Prove that there is some number $x$ such that $f(x)=f(x + \frac{1}{n}).$

I was wondering if my proof was logically sound, especially the last bit. I tried to apply the logic of the Intermediate Value Theorem, but was curious as to whether the contradiction in the end actually gives me the result I want:

Let $g$ be a function such that $g(x) = f(x+\frac{1}{n}) - f(x)$. We want to show that there exists an $x \in [0,1]$ such that $g(x) = 0$. We prove by contradiction.

  1. Assume $\forall x \in [0,1], g(x) < 0.$ Then for $i = 0, 1, 2, ..., n$ we have that $g(\frac{i}{n}) < 0$. Therefore, $f(0) > f(\frac{1}{n}) > f(\frac{2}{n}) > ... > f(1),$ which means that $f(0) \neq f(1)$.
  2. By similar logic, it follows that $g(x) > 0$ cannot hold for all $x \in [0,1]$.

Hence, there must exist $a,b \in [0,1]$ such that $a$ and $b$ have different sign, i.e. $g(a) \leq 0 \leq g(b)$. By the Intermediate Value Theorem, there exists $c$ such that $g(c) = 0$, and we are done.

Thank you for your help!

$\endgroup$
  • $\begingroup$ The idea is completely sound, yet I cannot grasp the reasoning in (1): why if $\;g(x)<0\;$ for all $\;x\in[0,1]\;$ , we must have thet $\;f(0)>f\left(\frac1n\right)>\ldots etc.$ ? Where does this sequence of inequalities come from? $\endgroup$ – DonAntonio Mar 26 at 1:27
  • $\begingroup$ @DonAntonio For $f(1/n)-f(0)<0$, $f(2/n)-f(1/n)<0$, etc. $\endgroup$ – fantasie Mar 26 at 1:33
  • 1
    $\begingroup$ @DonAntonio . If $g<0$ then for $0\le j \le n-1$ we have $0>g(j/n)=f((j+1)/n)-f(j/n)$. $\endgroup$ – DanielWainfleet Mar 26 at 1:37
  • 2
    $\begingroup$ Very nice. An alternate is that $\sum_{j=0}^{n-1}g(j/n)=\sum_{j=0}^{n-1}f((j+1)/n)-f(j/n)=f(1)-f(0)=0$, so the values $g(j/n)$ for $0\le j \le n-1$ cannot be all $+$ nor all $-.$ This is called the Horizontal Chord Theorem. $\endgroup$ – DanielWainfleet Mar 26 at 1:52
1
$\begingroup$

I think your argument is both neat and sound.

Here is a way to avoid the whole contradiction argument, at the cost of using another well-known theorem about continuous functions. Because $f$ is continuous on the closed interval $[0,1]$, it attains its maximum, say $f(x_0)=M$. Now $$g(x_0-\tfrac1n)=M-f(x_0-\tfrac1n)\geq0.$$ And $$ g(x_0)=f(x_0+\tfrac1n)-M\leq0. $$ As $g$ is continuous, there exists $x$ between $x_0-\tfrac1n$ and $x_0$ such that $g(x)=0$.

The argument above does not work in principle if $x_0-\tfrac1n<0$, but that can be solved by defining $g$ as $f(0)$ outside of $[0,1]$.

$\endgroup$
  • $\begingroup$ What is the intuition behind this idea? Just curious as to how someone could come up with something like this. $\endgroup$ – alexion Mar 26 at 3:22
  • 2
    $\begingroup$ I thought of the curve climbing up from $f(0)$ up to the max (think of this as the first part) and the climbing down from the max to $f(1)$ (think of this as the second part). The horizontal distance between the two curves is $1$ at the bottom and $0$ at the max, so all distances are achieved in between. I struggled for a bit to make this into someting analytic, until I noticed that your $g$ is exactly what one needs. $\endgroup$ – Martin Argerami Mar 26 at 3:32

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.