3
$\begingroup$

Find all Riemann integrable functions $f:[0,1] \to \mathbb{R}$ such that $\forall x \in [0,1]$ we have $\int_0^x f(t)dt =(f(x))^{2015}+f(x)$.

We consider $g:\mathbb{R} \to \mathbb{R}$, $g(x)=x^{2015}+x$. Then $g$ is differentiable and invertible. We denote $F(x)=\int_0^x f(t)dt$.

The hypothesis can be written as: $g^{-1}(F(x))=f(x)$ (1).

$f$ is Riemann integrable $\Rightarrow$ $F$ is continuous, so, by (1), we deduce that $f$ is continuous. Then $F$ is differentiable, so, again by (1), we have that $f$ is differentiable.

By differentiating $F(x)=g(f(x)), \forall x\in [0,1]$ or $g^{-1}(F(x))=f(x)$ we get that $f(x)=f'(x)g'(f(x))$. Can anyone help me to finish this problem? Thank you!

$\endgroup$
2
  • $\begingroup$ Where does this problem come from? The appearance of the number 2015 is typical of contest problems. Please credit your source. $\endgroup$ Apr 24 '15 at 14:58
  • $\begingroup$ source: Romanian NMO 2015 Shortlist! $\endgroup$
    – npatrat
    Apr 28 '15 at 19:30
6
$\begingroup$

In fact, there are no non-zero differentiable solutions to this equation. More generally: if $f:[0,1] \to \Bbb R$ with $|f'(x)| \leq C |f(x)|$ for some $C > 0$ for all $x \in [0,1]$ and $f(0) = 0$, then $f(x) = 0$.


Note also: plugging into original equation, we have $$ 0 = [f(0)]^{2015} + f(0) \implies f(0) = 0 $$

Back to where you were: we have a separable differential equation. In particular, $$ f(x) = f'(x)g'(f(x)) \implies\\ \frac{g'(f(x))}{f(x)} f'(x) = 1 $$ Integrating both sides (making the substitution $y = f(x)$), we have $$ \int_a^{f(x)} \frac{g'(y)}{y}\,dy = x+C \implies\\ \frac{2015}{2014}[f(x)]^{2014} + \ln[f(x)] = x+C $$ We run into a bit of a problem here, however, since $\ln[f(0)]$ is not defined. I'm not sure what to take away from this, honestly. Perhaps this implies that $f = 0$ is the only solution.

$\endgroup$
3
  • $\begingroup$ The differential equation has other solutions, but none of them satisfy $f(0)=0$ which is implied by the problem statement. (Putting $x=0$ into the original equation gives $0 = f(x)^{2015} + f(x) = f(x)(f(x)^{2014}+1)$ and the second factor is strictly positive.) You have, in effect, already argued that $f=0$ is the unique solution to the initial value problem with $f(0)=0$. Your "separation of variables" approach is invalid for $f$ with $f(0)=0$, since it involved a division by zero early on, so it is not surprising that this particular solution is not found by that method. $\endgroup$ Apr 24 '15 at 15:01
  • $\begingroup$ For those wanting a proof of the "more generally" statement in the first paragraph: see Gronwall's inequality. $\endgroup$ Apr 24 '15 at 15:04
  • $\begingroup$ @NateEldredge thanks for clarifying. The "more generally" statement is also a problem from baby Rudin. $\endgroup$ Apr 24 '15 at 17:13

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.