2
$\begingroup$

Let f be a strictly monotonic positive valued continuous function defined on $[a,b]$ such that $f(a) < a$ and $f(b)>b$ where $b>a>0$ then prove that there exist some $c\in (a,b)$ such that $f’(c)>1$

My approach:

I know that if $f$ is strictly monotonous and continuous then $f$ is differentiable almost everywhere. As graph of $f$ crosses $y=x$, at some $c\in (a,b)$, $f’(c)>1.$ I am not fully convinced with my approach. Is this correct or I need differentiable condition in the problem

$\endgroup$
1
  • $\begingroup$ No. If $c$ is a point where the graph crosses the diagonal there's not reason to think that $f$ is differentiable at $c$, and if it is diifferentible at $c$ I don't see why it should follow that $f'(c)>1$. (It's not quite clear to me whether the claim is actually true...) $\endgroup$ – David C. Ullrich Sep 13 '18 at 16:38
3
$\begingroup$

If you mean to assert that there exists $c$ such that $f$ is differentiable at $c$ and $f'(c)>1$ then what you're trying to prove is false. (It's probaby true under some weaker notion of "$f'(c)>1$".)

Say $K\subset [0,1]$ is the Cantor set and $g:[0,1]\to[0,1]$ is the "Cantor-Lebesgue function". Then $g(0)=0$, $g(1)=1$, $g$ is continuous and non-decreasing, $g'(x)=0$ for every $x\in[0,1]\setminus K$, and if $x\in K$ it's not hard to see that $g$ is not differentiable at $x$ (because if $x\in K$ then for every $n\ge1$ we have $x\in[j/3^n,(j+1)/3^n]$ with $g((j+1)/3^n)-g(j/3^n)=1/2^n$).

So if you define $$f(t)=g(t)-\frac13+\frac23t$$then $f$ is continuous and strictly increasing, $f(0)<0$, $f(1)>1$, but $f'=2/3$ at every point where $f$ is differentiable.

$\endgroup$
3
  • $\begingroup$ Excellent !! Thanks a lot $\endgroup$ – Makar Sep 13 '18 at 17:53
  • 2
    $\begingroup$ (+1) It's actually the lack of absolute continuity that makes the claim false. If $f$ were absolutely continuous, then the statement can easily seen to be true: indeed, if one assumes $f'(x) \leq 1$ almost everywhere, then (in view of absolute continuity of $f$) we have $$ f(x) = f(a) + \int\limits_a^x f'(t) dt \leq f(a) + x - a < x, $$ where the last inequality is due to $f(a) < a$. But then setting $x=b$ we get $f(b) < b$ which is a contradiction. $\endgroup$ – Hayk Sep 13 '18 at 19:28
  • $\begingroup$ @Hayk Yes of course - in fact I'd typed up exactly that argument when I realized that for some reason I'd been assuming $f$ was Lipshcitz... $\endgroup$ – David C. Ullrich Sep 13 '18 at 20:27

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.