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.

Let $f:[0,1]\to [0,1]$ be a non-decreasing concave function, such that $f(0)=0,f(1)=1$. Prove or disprove that : $$ \int_{0}^{1}(f(x)f^{-1}(x))^2\,\mathrm{d}x\ge \frac{1}{12}$$ A friend posed this to me. He hopes to have solved it, but he is not sure. Can someone help? Thanks.

share|improve this question
1  
Its $f^{-1}(x)$ not $(f(x))^{-1}$, –  shadow10 Jul 17 at 16:29
1  
You're completely right. I am wrong, so I'm deleting my previous comment to avoid mistakes. –  Adolfo Jul 17 at 16:32
1  
What have your friend done ? This Information would simplify the whole process. –  calculus Jul 17 at 17:18
    
So we have to show $\displaystyle \int_0^1 [(2f(x)f^{-1}(x))^2 - x^2]\,dx \ge 0$ ... –  r9m Jul 17 at 17:25

3 Answers 3

Let $f(x)=x^{\frac{1}{k}}$ for some $k\geq 1$. Then $f^{-1}(x)=x^k$, so we have

$$\int_{0}^{1}(x^{\frac{1}{k}}x^k)^2\,\mathrm{d}x = \int_{0}^{1}(x^{k+\frac{1}{k}})^2\,\mathrm{d}x = \frac{x^{2k+\frac{2}{k}+1}}{2k+\frac{2}{k}+1}\Big|_0^1 = \frac{1}{2k+\frac{2}{k}+1}$$

which can be made arbitrarily small.

share|improve this answer
    
This should be stated in the other way, $f(x)=x^{1/k}$ with $k\geq 1$, since we want that $f(x)$ is a concave function. –  Jack D'Aurizio Jul 17 at 20:46
    
I thought of this but didn't work it correctly out in my head. –  nayrb Jul 17 at 23:24
    
@JackD'Aurizio Thank you; I've fixed it. –  Théophile Jul 18 at 18:23

The inequality cannot hold since a non-decreasing concave function $f$ satisfing $f(0)=0$ and $f(1)=1$ can be almost everywhere arbitrarily close to one, with its inverse function being almost everywhere arbitrarily close to zero on $[0,1]$. Hence the given integral can be less than $\varepsilon$ for any $\varepsilon>0$.

Examples of such a phenomenon are given by $$ f(x) = \frac{(n+1)x}{n+x}\quad\mbox{for }n\geq 7, $$ $$ f(x) = x^r\quad\mbox{for }r\in(0,1/6),$$ $$ f(x) = x+\min\left(nx,\frac{n}{n+1}(1-x)\right)\quad\mbox{for }n\geq 4.$$

share|improve this answer

I am not so sure this works. Take $f(x)=nx$ for $x \in [0;\frac{1}{n+1}]$ and $f(x)=\frac{n}{n+1} + (x-\frac{1}{n+1})\frac{1-\frac{n}{n+1}}{1-\frac{1}{n+1}} = \frac{n}{n+1} + \frac{1}{n}(x-\frac{1}{n+1})$ on $]\frac{1}{n+1},1]$.

Then $f^{-1}(y)= \frac{y}{n}$ if $y \in [0,\frac{n}{n+1}]$ and $f^{-1}(y)= n(y-\frac{n}{n+1})+\frac{1}{n+1}$ if $y \in [\frac{n}{n+1},1]$

Then, noticing that$\forall x$ $f(x)\leq 1$ and $f^{-1}(x)\leq 1$, we can rewrite: $\int_0^1 (f f^{-1} )^2\leq \int_0^1 f^{-1}(x )^2dx \leq \int_0^{\frac{n}{n+1}}\frac{x^2}{n^2}dx + \int_{\frac{n}{n+1}}^1 1dx \leq \frac{1}{n ^2} + \frac{1}{n+1}$ which converges towards 0.

share|improve this answer

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.