1
$\begingroup$

Let F be the distribution function of random variable X, that is $F(t)=Prob[X<t]$ and know that $E[X]=0$ and $Var[X]=1$. I am trying to understand how to show the following inequality
$\int_{1/u}^{\infty} e^{ut}dF(t) \leq\sum_{k=1}^\infty e^{k+1} Prob[X\geq k/u]$

$\endgroup$
1
$\begingroup$

note that $ut \leq [ut] + 1$ therefore

$$ \int_{1/u}^{\infty} e^{ut}dF(t) \leq \int_{1/u}^{\infty} e^{[ut] + 1}dF(t) = \int_{1/u}^{\infty} \sum_{k=1}^\infty e^{k+1} 1_{\frac{k}{u}<t\leq \frac{k+1}{u}}dF(t)= \sum_{k=1}^\infty e^{k+1} \bigg(F\bigg(\frac{k+1}{u} \bigg) - F\bigg(\frac{k}{u} \bigg)\bigg) = \sum_{k=1}^\infty e^{k+1} \Bbb{P}\bigg(\frac{k}{u} <X\leq\frac{k+1}{u} \bigg) \leq \sum_{k=1}^\infty e^{k+1} \Bbb{P}\bigg(\frac{k}{u} <X \bigg) $$

Note that we didn't use that $\Bbb{E}(X) = 0$ nor that $Var(X)=1$

$\endgroup$
2
  • $\begingroup$ why is $$ \int_{1/u}^{\infty} e^{[ut] + 1}dF(t) =\sum_{k=1}^\infty e^{k+1} [F\bigg(\frac{k+1}{u} \bigg) - F\bigg(\frac{k}{u} \bigg)] $$ ? (I am assuming there should an extra parenthesis there ) Is it an approximation or an exact result? $\endgroup$
    – Slug Pue
    Sep 3 '15 at 14:20
  • $\begingroup$ I edited the question @SlugPue $\endgroup$ Sep 3 '15 at 14:33

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.