3
$\begingroup$

Let $a, b \in R$ ($a<b$) and let $f: [a,b] \longrightarrow R$ a continuous function s.t. $f(a) = f(b)$. I need to show that there exist $\delta > 0 \in R$ s.t. $\forall t \in [0, \delta] \exists x \in [a,b-t]$ that $f(x+t)= f(x)$

What I've done: $f$ is continuous in the closed and bounded interval $[a,b]$, then $f$ must attain a maximum and a minimum so if the maximum is in the edges then we're done because $ f(a) = f(b)$. Otherwise it has to be within the interval. I want to take a $ \delta$ around the maximum and show that $f(x+t)= f(x)$ using the Intermediate value theorem.

Can you help me formalize it?

Thanks in advance!

$\endgroup$
  • $\begingroup$ Try defining a new function $g(x)=f(x)-f(x+t)$ $\endgroup$ – Itay4 Mar 25 '17 at 16:10
0
$\begingroup$

Let's suppose that $f$ is not constant so that it attains a global min or max in the interior (the result is trivial for constant $f$). Let us assume without loss of generality that there is a global maximum $f(x_0)=M$.

I claim that $\delta = \min\{x_0-a,b-x_0\}$. For fixed $0 \le t \le \delta$, let us define the function $$g(s) = f(x_0+s) - f(x_0-t+s),$$ for $0 \le s \le t$. Notice that $g$ is well defined since $$a \le x_0 + s \le x_0 + t \le x_0 + \delta \le b,$$ and likewise $$a \le x_0 - \delta \le x_0 - t \le x_0-t+s \le x_0 \le b.$$

When $s=0$, we have $$g(0) = f(x_0) - f(x_0 - t) \ge 0,$$ since $f(x_0) = M$ is the global maximum. Likewise, when $s=t$, we have $$g(t) = f(x_0+t) - f(x_0) \le 0.$$ If ether inequality is an equality, then we have found a point $x_t$ such that $f(x_t) = f(x_t+t)$, namely $x_t=x_0$ or $x_t=x_0-t$.

So suppose both inequalities are strict. Then by the intermediate value theorem, there must exist some $s_0$ for which $g(s_0) = 0$. Then our required point is $x_t=x_0+s_0-t$, for which $$g(s_0) = f(x_0+s_0) - f(x_0-t+s_0) = f(x_t + t) - f(x_t) = 0.$$ In either case, we have found our required $x_t$. Since we have done for this every $t\in [0,\delta]$, this completes the required proof.

$\endgroup$
  • $\begingroup$ Now I understand what I missed there $\endgroup$ – user21312 Mar 25 '17 at 19:18

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.