1
$\begingroup$

I've been working on a problem that wants me to show that given a function $f$ that is continuous at the point $c$ that, $$f(c)>0 \to \exists \delta\;\ \text{such that}\;\ f(x)>0\; \forall x \in(c-\delta,c+\delta) $$

I thought about taking $\epsilon < f(c)$ so that $0<f(c)<2f(c)$, but I'm pretty sure I've come up with a case where this wouldn't work. I know I'm probably missing some simple trick. I would appreciate any hints or advice.

$\endgroup$
5
  • $\begingroup$ Why don't you tell us the case where you think this approach wouldn't work? $\endgroup$ – Michael Albanese Jul 22 '15 at 2:41
  • $\begingroup$ Well I took $f(x)=x^2$ and let $c=2$ which yielded $0<f(x)<8$ for $x \in (2-\delta, 2+\delta)$ I think I may be put off by the fact that I don't have a way of finding an exact $\delta$ that would work , other than by inspection. $\endgroup$ – boseman Jul 22 '15 at 2:47
  • $\begingroup$ You can find a $\delta$ if you really want, but that's not required. You just have to show the existence of some $\delta$ and you have done that (by using the continuity of $f$). $\endgroup$ – Michael Albanese Jul 22 '15 at 2:49
  • $\begingroup$ So this proof wouldn't be considered "constructive", but still works? $\endgroup$ – boseman Jul 22 '15 at 2:52
  • $\begingroup$ I suppose not, but there is no way it could be. You can't express the relationship between $\varepsilon$ and $\delta$ in a way which holds for every continuous function $f$. $\endgroup$ – Michael Albanese Jul 22 '15 at 3:00
1
$\begingroup$

Assume $f(c)>0$, then use $\epsilon := f(c)/2 >0$. By continuity, given any

$\epsilon>0$ , there is a $\delta$ with $$|f(c')-f(c)|< \epsilon $$ when $|c'-c|< \delta$.

Specializing here to $\epsilon =f(c)/2$, we get that $|f(c')-f(c)|< f(c)/2$ when $|c'-c|< \delta $, so that, expanding the expression in terms of $f(c')$....

$\endgroup$

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.