Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $\mu$ be a probability measure on $\mathbb{R}^m$ (so $\int_{\mathbb{R}^m} \mu(d x) = 1$).

Let $f_i:\mathbb{R}^m \rightarrow \mathbb{R}_{\geq 0}$ be integrable functions and let also $\limsup_{i \rightarrow \infty} f_i$ be an integrable function.

Assume that

$$ \limsup_{i \rightarrow \infty} \int_{K} f_i(x) \ \mu(d x) \leq \ c \quad \text{ for all compact sets } K \subset \mathbb{R}^m.$$

Find under which additional assumptions (for instance on the $f_i$s and/or on $\int f_i$s) we also have:

$$ \limsup_{i \rightarrow \infty} \int_{\mathbb{R}^m} f_i(x) \ \mu(d x) \ \leq \ c$$

Notes: The measure in the two integrals is the same. What changes is the domain: from $K$ (compact, arbitrarily large) to $\mathbb{R}^m$. $c \in \mathbb{R}_{\geq 0}$.

share|improve this question
    
I suspect it's true. –  Michael Hardy Mar 2 '12 at 4:23
    
@Adam I don't know the site etiquette, but it strikes me as against it to ask a question, accept an answer, then change the question based on the answer (and unaccept answer), but I'll remove my answer as it no longer seems relevant. –  ShawnD Mar 2 '12 at 20:00
    
I'm sorry about that.. I thought opening a new (closely-related) question would have been even worse.. By the way, I did appreciate your answers. –  Adam Mar 2 '12 at 20:14
    
I'll undelete because of comments and note that it is no longer an answer. –  ShawnD Mar 2 '12 at 20:15
add comment

1 Answer

This is not true in general.

Consider $\mathbb{R}$ with Lebesgue measure. Let $f_i(x)=i$ on the interval $[i,i+1]$. For any compact set $K$, for large $i$, $f_i$ will be $0$ on $K$, so your first limit is $0$. However, $\int_{\mathbb{R}}f_i(x)dx=i$, so your second limit is $\infty$.

EDIT

This was an answer to the original question, but is not relevant to current question. Perhaps the comments are useful though.

share|improve this answer
    
If $\limsup_{i\rightarrow \infty} f_i$ is integrable, does the result hold true? –  Adam Mar 2 '12 at 5:39
1  
Even if you made $f_i(x)=1$ on $[i,i+1]$ so that $\lim_{i \rightarrow \infty} f_i =0$, the result would still not hold (with say $c=1/2$). –  ShawnD Mar 2 '12 at 5:44
    
i guess $\lim_{i \rightarrow \infty} f_i=0$ in the first example, but the integrals are blowing up. –  ShawnD Mar 2 '12 at 5:58
1  
I think you could still have mass that escapes. I think this probably is true if your system is uniformly integrable (en.wikipedia.org/wiki/Uniform_integrability) –  ShawnD Mar 2 '12 at 6:25
1  
Are you able to prove it's true if your random variables are uniformly integrable? –  ShawnD Mar 3 '12 at 23:13
show 6 more comments

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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