0
$\begingroup$

Let $(X_i(x))_{i\in\mathbb{N}}$ be a sequence of real random variables with $\operatorname{E}[X_i(x)]=\mu(x)\leq\infty$, for all $i\in\mathbb{N}$ and $x\in\mathbb{R}$. Let's assume that we have $$\sup_{x\in\mathbb{R}}\|X_i(x)\|_{L^1(\Omega)}\xrightarrow{n\rightarrow\infty}0.$$ Is it possible to follow that $$\sup_{x\in\mathbb{R}}|X_i(x)|\xrightarrow{p}0,~~\text{as}~~n\rightarrow\infty?$$

$\endgroup$
  • $\begingroup$ By $\|X_i(x)\|_{L^1(\Omega)}$ do you mean something like $\int |X_i(x)(\omega)| d \mu(\omega)$? And by $|X_i(x)|$ do you mean $|X_i(x)(\omega)|$? $\endgroup$ – copper.hat Feb 4 '14 at 7:36
  • $\begingroup$ @copper.hat : Yes, for $\sup_{x\in\mathbb{R}}\|X_i(x)\|_{L^1(\Omega)}$ that's what I mean. By $\sup_{x\in\mathbb{R}}|X_i(x)|\xrightarrow{p}0$ I mean: For all $\epsilon>0$ it holds $$\sup_{x\in\mathbb{R}}P(\{\omega\in\Omega:|X_i(x)(\omega)|>\epsilon\})~~\text{tends to}~~ 0,$$ as $n\rightarrow\infty$. $\endgroup$ – stroem Feb 4 '14 at 8:28
  • $\begingroup$ I fixed my answer. $\endgroup$ – copper.hat Feb 4 '14 at 8:51
1
$\begingroup$

Let $\epsilon>0$, and $A_i(x) = \{ \omega | |X_i(x)(\omega)| > \epsilon\}$.

Since $|X_i(x)| \ge \epsilon 1_{A_i(x)} $ you have $\int |X_i(x)| \ge \epsilon P(A_i(x))$, and so $P(A_i(x)) \le {1 \over \epsilon} \int |X_i(x)| \le {1 \over \epsilon} \sup_x \int |X_i(x)|$.

Consequently, you have $\sup_x P(A_i(x)) \le {1 \over \epsilon} \sup_x \int |X_i(x)|$, from which the result follows.

$\endgroup$
  • $\begingroup$ Thanks a lot for the fast and understandable answer! $\endgroup$ – stroem Feb 4 '14 at 8:55
  • $\begingroup$ Delighted to help! $\endgroup$ – copper.hat Feb 4 '14 at 8:59

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.