$f : \mathbb{R} \to \mathbb{R}$ is a differentiable function satisfying $$f\left(\frac{x+y}{3}\right)=\frac{2+f(x)+f(y)}{3}$$

if $f'(0)=2$, find the function

My Try:

we have $$f\left(\frac{x+y}{3}\right)-\frac{f(y)}{3}=\frac{2+f(x)}{3}$$ $\implies$


Now taking Limit $x \to 0$ we have

$$\lim_{x \to 0}\frac{f\left(\frac{x+y}{3}\right)-\frac{f(y)}{3}}{\frac{x}{3}}=\lim_{x \to 0}\frac{2+f(x)}{x}$$


$$f'\left(\frac{y}{3}\right)=\lim_{x \to 0}\frac{2+f(x)}{x}$$

Now since LHS to be finite , we need $0/0$ form in RHS, hence $f(0)=-2$

Now by L'Hopital's Rule we get


Integrating we get


Putting $y=0$ we get $c=-6$



So $$f(y)=2y-2$$

Hence $$f(x)=2x-2$$

But this function is not satisfying given functional equation.

What went wrong?

  • 1
    $\begingroup$ The limit which you claimed to be $f'(y/3)$ is not, except perhaps if you implicitly assume the linearity of the map $f$, which would be a contradiction with its defining equation. By the way, setting $x=y=0$ in the defining equation, one gets $f(0) = 2$. In fact, the affine map $a(x) = 2x+2$ solves the equation for $f$ and the derivative constraint. $\endgroup$ – Jordan Payette Jan 25 '18 at 3:31

Using $x=y=0$ in the functional equation we can see that $f(0)=2$. Further note that $$f(x+h)=\frac{2+f(3x)+f(3h)}{3}, f(x) =\frac{2+f(3x)+f(0)}{3}$$ so that $$\frac{f(x+h) - f(x)} {h} =\frac{f(3h)-f(0)}{3h}$$ Taking limits as $h\to 0$ we get $f'(x) =f'(0)=2$ and thus $f(x) =2x+c$ and from $f(0)=2$ we get $c=2$ so that $f(x) =2x+2$. You can check that it satisfies the functional equation.


One readily verifies that the affine map $a(x) = 2x+2$ is a differentiable function which solves both the defining equation for $f$ and the derivative constraint $f'(0) = 2$.

Let's prove now the uniqueness of this solution. The defining equation can be rewritten $f\left( \frac{x+y}{3} \right) - \frac{f(x) + f(y)}{3} = \frac{2}{3}$. Differentiating with respect to $x$, we get $f'\left( \frac{x+y}{3} \right) - f'(x)=0$ regardless of the values $x$ and $y$, so that $f'(x)$ is a constant. The constraint $f'(0)=2$ thus yields $f'(x) \equiv 2$, which integrates to $f(x) = 2x+c$. Substituting this result into the defining equation for $f$ readily yields $c=2$.

Remark: Paramanand Singh's answer is better than mine in that it proves that $f$ is differentiable at every point by only using its defining equation and the assumption that it is differentiable at $0$. In that respect it is quite instructive. Be aware however that such a proof is in general rather difficult, hence the frequent assumption that $f$ is differentiable ; one doesn't need then to rely on the definition of a derivative, but can rather merely differentiate the defining equation as I did above.


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.