Register to reply

Continuity of Measure ( I Think)

Tags: continuity, measure
Share this thread:
Feb27-13, 07:30 PM
P: 576
Hi, All:

I think the following deals with continuity of measure, but I'm not 100%:

Let I:=[0,1] , and let An be a sequence of pairwise-disjoint measurable sets
whose union is I ( is me? :) ) . Let {Bj} be a sequence of measurable subsets
of I , so that, for μ the standard Lebesgue measure:

Limj→∞ μ( An\cap Bj )=0 , for all n .

I want to show that above implies that : Limj→∞ μ(Bj)=0 .(**)

This is what I have:

We know that Ʃμ(An)=1 . So we must have some Ano in the
collection with μ(Ano)=a>0.

( I am assuming that the A_n's must all be of the form [a,b) , with A1=[0,a)

A2=[a,b) , etc. , plus a {1} thrown-in )

Now, I am trying to argue by contradiction , assuming that the limit above in (**) equals
some c+e ; e->0 , though I am not sure of how to show that the limit actually exists,
tho I am assuming for now that it does:

So, assuming limit in (**) exists and equals c+e ( e->0) , we have that there is an
integer N such that for all j>N :

c=c-e+e< μ(Bj)< c+e+e

In particular, μ(Bj)>c>0 .

Now, I can find an open set Oj, for each j , with

μ(Cj)=μ(Oj) .

I know the quantification here is tricky; I am then using that:

Oj= \/(cji ,dji)

And, since m(Bj)>c for all j>N , there is an index for the j's --

use j=1 without loss of generality -- such that m(c1,d1)>0

Now, this interval (c1,d1) must intersect some interval

An , and the intersection must be of one of the forms:

[x,y) , (x,y] , or (x,y) . In either case, the measure of the intersection is

y-x>0 , contradicting the assumption condition (**) that

Limj→∞ μ(Bj)=0.

I think I'm on the right track, but not 100%. Please critique.

Phys.Org News Partner Science news on
Mysterious source of ozone-depleting chemical baffles NASA
Water leads to chemical that gunks up biofuels production
How lizards regenerate their tails: Researchers discover genetic 'recipe'
Feb27-13, 10:22 PM
P: 3,240
Can you translate your question to Latex or PDF, cause it's hard to read through ascii. I am not 20 anymore that I have the patience to read that way.
Mar9-13, 09:13 AM
Sci Advisor
P: 1,169
Use the fact that the tail of the sum m(Ai) goes to zero, since the total sum is 1

Then use the fact that, by the limit condition, there is a K>0 with

Lim_n->0 (Bk /\An)=0 , for all k>K . Then , from the fact that m(Ai)->0 ,

Use Bk=(Bk /\ U Ai) , to conclude that m(Bk)->0.

Register to reply

Related Discussions
Condition for local absolute continuity to imply uniform continuity Calculus 0
Cauchy sequences and continuity versus uniform continuity Calculus & Beyond Homework 1
Real Analysis: Continuity and Uniform Continuity Calculus & Beyond Homework 5
Measure Theory - The completion of R^2 under a point mass measure Calculus & Beyond Homework 1
Lipschitz Continuity and measure theory Calculus 7