Let's consider two non-negative concave functions $f_1(x)$ and $f_2(x)$ such that $f_1(x) \geq f_2(x)$ always, $f_1(x)= f_2(x) = 0$ at $x=0$ and both attain their maxima at $x\rightarrow \infty$, i.e., $f'_1(x)=f'_2(x)=0$ for $x\rightarrow \infty$. Moreover, $\lim_{x \rightarrow \infty} f_1(x) - f_2(x) = 1$.

Is it possible to state that $f_1(x)-f_2(x)$ also attains maxima at $x\rightarrow \infty$ without doing any calculations with the specific form of functions or the only option is to find the first derivative of $f_1(x)-f_2(x)$ and show that it's always positive??


A simple counterexample is $$ f_1(x) = \min(4x, 2) , \\ f_2(x) = \min(x, 1) $$ which satisfy the given conditions, but the maximum of $f_1(x) - f_2(x) $ is $\frac 32$ at $x = \frac 12$.

A counterexample with differentiable functions is $$ f_1(x) = \frac{4}{\pi} \arctan(4x), \\ f_2(x) = \frac{2}{\pi} \arctan(x) $$ which satisfy the given conditions, but $$ f_1'(x) - f_2'(x) = \frac{2(7 - 8x^2)}{\pi(x^2+1)(16x^2+1)} $$ shows that the maximum of the difference is attained at $x = \sqrt{7/8}$.

  • $\begingroup$ Thanks @Martin... I guess then the only option is to check the first derivative. $\endgroup$ – kphy Jul 16 '17 at 7:37

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.