2
$\begingroup$

Let $f\colon[0, 1] \to \mathbb R$ be an $m$-strongly convex function and $\mu$ be a probability measure on $[0,1].$ For any $t<1$, the goal is to find a lower bound on $\int_{0}^t f^2(x) d\mu(x)$ in terms of $t$, $m$, and $\mu$ (and nothing else). We currently have the following bound $$\int_{0}^t f^2(x) d\mu(x) \ge \frac{ m^2 t^4}{36} \mu[0,t].$$ We do not know if our bound is tight. Moreover, our proof is really long and messy. A clean/simple proof of such an elementary result would be helpful.

$\endgroup$

1 Answer 1

3
$\begingroup$

The left-hand side, $\int_{0}^t f^2(x)\, dx$, of your inequality does not contain $\mu$. Since you wanted "to find lowerbound on $\int_{0}^t f^2(x) d\mu(x)$", it appears that your desired inequality is $$\int_0^t f^2(x)\,\mu(dx)\ge cm^2t^4 \mu[0,t] \tag{1}$$ for some real $c>0$.

However, (1) is obviously false in general. E.g., let $t=1/2$, $m=1$, $f(x)\equiv x^2/2$, and let $\mu$ be the Dirac measure supported on $\{0\}$. Then $f$ is $m$-strongly convex and $\int_0^t f^2(x)\,\mu(dx)=0$, so that (1) fails to hold for any $c>0$.

$\endgroup$
2
  • $\begingroup$ Sorry about that @Iosif. An assumption I forgot to mention, we assume that $\mu$ has a density with respect to the Lebesgue measure, and the density is bounded on $[0,1]$ (but not bounded away from zero) Thank you for your answer and time, however. $\endgroup$
    – Statguy
    Commented Jun 27, 2020 at 15:07
  • $\begingroup$ @Statguy : Your question was fully answered. If you actually wanted to ask something else, you can do that in a separate post. $\endgroup$ Commented Jun 28, 2020 at 2:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.