2
$\begingroup$

If g is a positive, twice differentiable function that is decreasing and has limit zero at infinity, does g have to be convex? I am sure, from drawing a graph of a function which starts off as being concave and then becomes convex from a point on, that g does not have to be convex, but can someone show me an example of an actual functional form that satisfies this property?

We know that since g has limit at infinity, g cannot be concave, but I am sure that there is a functional example of a function g:[0,∞)↦(0,∞) which is increasing, has limit zero at infinity, and is not everywhere convex, I just can't come up with it. Any ideas?

Thank you!

$\endgroup$
  • $\begingroup$ $g(x) = e^{-x^2}$ on $x \geq 0$. $\endgroup$ – cardinal Jul 7 '12 at 23:43
4
$\begingroup$

Since the functions mentioned so far are eventually convex, here is one more: $$ f(x)=e^{-x}(3+2\sin x) $$ The first derivative $$f\,'(x)=e^{-x}(2\cos x-2\sin x-3)$$ is always negative because $\cos x-\sin x\le \sqrt{2}$ for all $x$. But the second derivative $$f\,''(x)=e^{-x}(3-4\cos x)$$ changes sign infinitely many times.

$\endgroup$
2
$\begingroup$

Try $$f(x)=\frac{\pi}2-\tan^{-1}x\;.$$

$\endgroup$
  • $\begingroup$ This function is a great source of counterexamples. Rescaled, it gives you an example of a (non-strict) contraction mapping with no fixed points. $\endgroup$ – Alex Becker Jul 7 '12 at 23:47
0
$\begingroup$

Let $f(x) = \begin{cases} e^{1\over x}, & x < 0 \\ 0, & \text{otherwise} \end{cases}$.

$\endgroup$

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.