Real analysis, ill ask in the comments.
Suppose that \(\{A_n\}_n^\infty \subset \mathcal{M}\), where \(\mathcal{M}\) is the \(\sigma-\)algebra of Lebegue measurable sets. Also suppose \(....A_3\subset A_2\subset A_1\). Prove that \(\lambda\large(\cap_{n=1}^\infty A_n)\) = \(\lim_{n\rightarrow \infty}(A_n)\)
this is what I know, and I think I should use. given the same conditions, but with \(A_1\subset A_2\subset A_3...\), then \[\lambda\large(\cup_{n=1}^{\infty}A_n)=\lim_{n\rightarrow\infty}\lambda(A_n)\]
that should say prove that \(\lambda\large(\cap_{n=1}^\infty A_n)\) = \(\lim_{n\rightarrow \infty}\lambda(A_n)\)
I feel that DeMorgan's law is in there some place, but I'm not sure hot to apply it.
The statement is not true in general. Take this example \[ A_n = [ n, +\infty[\\ then \\ A_{n+1} \subset A_n\\ A=\cap_n A_n =\emptyset\\ \lambda (A)=0 < \lim_{n\to} \lambda\left ( A_n \right)=\infty \]
It is true however if you suppose that one of your decreasing sets has finite measure.
\[ \lim_{n\to \infty}\lambda\left ( A_n \right)=\infty \] How do you get this?
So \(\lambda([\infty,\infty]) = \infty\)?
Yes
the measure of each of the sets is infinite
Wouldn't \((\infty,\infty)=\emptyset\)?
For the case of the finite case. We can suppose that all of them have finite measure. You can read page 34 on http://www.ams.org/bookstore/pspdf/stml-48-prev.pdf Where all the sets are inside a finite interval.
Actually for each n on my example \[ \lambda(A_n) = +\infty \] Hence their limit is \[ +\infty \]
May I ask how exactly you get \[ \lambda(A_n) = \infty \]Even in the case where \(n\) is finite? What are you using?
How much do you think \[ \lambda (A_n) \] is?
Well, my best guess is \[ \lim_{m\to \infty}m-n \]based on the resource you gave me.
For each integer k, we have \[ [ n, n+k] \subset A_n\\ k=\lambda([n,n+k]) \le \lambda(A_n)\\ \] Would this argument satisfy you?
What is your background in Math @wio?
I'm not unsatisfied, I just want to see a more clear definition.
If a quantity is bigger than any number k given in advance, what could this quantity be?
Have you taken a real analysis course about Lebesgue integration?
No, I just think that \(\infty-\infty\) is involved here, so it is tricky.
ok sorry it said \(\lambda(A_1)<\infty\)
sorry, its 4am
In this case, the reference I gave you above on page 34 gives you the proof.
For the case of the finite case. We can suppose that all of them have finite measure. You can read page 34 on http://www.ams.org/bookstore/pspdf/stml-48-prev.pdf Where all the sets are inside a finite interval.
ahh great, ty.
Suppose:\[ \lim_{y\to\infty}f(x,y) =\infty \]Does this mean: \[ \lim_{x\to\infty}\lim_{y\to\infty}f(x,y) =\infty \]Or does it depend on \(f\)? What if \(f(x,y) = y-x\)?
I find this question somewhat interesting, even if not completely relevant.
YW
hey @eliassaab, on the first line for the decreasing proof, where do they get the 1 - from?
I think in their proof, the measure of the whole interval is 1. To do your proof replace 1 by the measure of your first set A_1 and take the compleent with respect to it/
ahh, i see.
very strange, they are saying the compliment is 1-... and they do that all over that book, but I never see them say we are dealing with some special interval. Maybe they assume it up front at some point in the text.
its like they use it as a place holder for the compliment....
You should be using \(|U| - |A|\) I suppose?
well for me A_1^c will be R/A_1, which has infinite measure...
maybe i need to sleep, its 430:)
gn
ill read anything you guys say in the morning.
But you may be able to use \(|U|-|U|=0\) at some point.
@eliassaab do you see where it says they are using some interval with length 1? I dont see them say that anywhere so I cant see why they use 1.
when you said "I think in their proof, the measure of the whole interval is 1. To do your proof replace 1 by the measure of your first set A_1 and take the compleent with respect to it/" why would I use my A_1 instead of 1, and not R-union....
Take \[ X=A_1\\ E_n= X \backslash A_n\\ X\backslash \cap A_n= \cup E_n\\ \lambda(X\backslash \cap A_n)= \lambda(\cup E_n)\\ \lambda(X) -\lambda(\cap A_n))=\lim_n \lambda(E_n)=\lim_n( \lambda(X) -\lambda(A_n))\\ \lambda(X) -\lambda(\cap A_n))=\lambda(X) -\lim_n\lambda(A_n)\\ -\lambda(\cap A_n))= -\lim_n\lambda(A_n)\\ \lambda(\cap A_n))=\lim_n\lambda(A_n)\\ \]
ok ill think about this one, ty
To be able to cancel \(\lambda(X)\) we need it to be finite.
yw
I also used the fact that \(E_n\) is increasing for inclusion.
where did you use that? sorry this is all new to me.
The fact that \( E_n\) is increasing
also also \(\lim_{n\rightarrow\infty}(\lambda(X)-\lambda(A_n))=\lambda(X)-\lim_{n\rightarrow \infty}\lambda(A_n)\) can we say this because lambda(X) does not depend on n?
yes
and finally, \[\lambda(A-B)=\lambda(A)-\lambda(B)\] I imagine this is true for all l-measurable sets, and that its easy to prove?
nm i answered that last question
\[ X=A_1\\ E_n= X \backslash A_n\\ X\backslash \cap A_n= \cup E_n\\ \lambda(X\backslash \cap A_n)= \lambda(\cup E_n)\\ \text{Simce } E_n \\ \text { is increasing, The measure of their union is the limit of their individual measures} \\ \lambda(X) -\lambda(\cap A_n))=\lim_n \lambda(E_n)=\lim_n( \lambda(X) -\lambda(A_n)\quad \text { (since X has finite measure} )\\ \lambda(X) -\lambda(\cap A_n))=\lambda(X) -\lim_n\lambda(A_n) \\\ \text { (since X has finite measure we can cancel and get} \\ -\lambda(\cap A_n))= -\lim_n\lambda(A_n)\\ \lambda(\cap A_n))=\lim_n\lambda(A_n)\\ \]
Are you now happy with the explanation?
i think i understand everything now:)
Excellent
ty, sorry im a little slow to this. we just got introduced to measure about 6 days ago
You are fine. This is normal. When you get used to it, it would become easy.
i hope so:) the cantor set blew my mind.
Join our real-time social learning platform and learn together with your friends!