Assuming $f(x)$ is a function of single variable, and $f'(x)$ is monotonically increasing

Then claim: $f'(\beta) \int\limits_\beta^\alpha 1 dt \leq \int\limits_\beta^\alpha f'(t) dt \leq f'(\alpha) \int\limits_\beta^\alpha 1 dt$

How do you show this is true?


$\int\limits_\beta^\alpha f'(t) dt \leq f'(\alpha) \int\limits_\beta^\alpha 1 dt$ is equivalent to $f(\alpha) - f(\beta) \leq f'(\alpha)(\alpha - \beta)$

Which theorem guarantees that this is true?

  • 1
    $\begingroup$ That is not true unless you mean that $f'$ is monotonically increasing. A simple counter-example is $f(x) = \sqrt{x}$ on $[1,2]$. btw the mean value theorem is the theorem you are looking for (and to use it to show what you want you need $f'(x) \leq f'(\alpha)$ for all $\beta \leq x \leq \alpha$). $\endgroup$
    – Winther
    Nov 7, 2015 at 18:00

1 Answer 1


I suggest not to use the mean value theorem. Also, the derivatives are just confusing (and so is the interval $[\beta,\alpha]$), so let me reformulate the statement as:

Given $g$ monotonically increasing on $[a,b]$, show that $$ g(a)\int_a^b1\,dt\leq \int_a^b g(t)\,dt\leq g(b)\int_a^b1\,dt. $$

Let us look at the left inequality. Since $g(t)\geq g(a)$ for all $t\in[a,b]$, $g(t)-g(a)\geq 0$, and the integral of a non-negative function is non-negative (this follows by definition, since no undersums/Riemann sums can give negative contribution), i.e. $$ \int_a^b g(t)-g(a)\,dt\geq 0. $$ Next, by linearity of the integral, the above is equivalent to $$ \int_a^b g(t)\,dt-g(a)\int_a^b 1\,dt\geq 0. $$ Moving the things around, we get $$ g(a)\int_a^b 1\,dt\leq \int_a^b g(t)\,dt. $$ I leave it to you to show the second inequality in the same manners.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .