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.

I'm preparing to the second mini-test in measure theory. Here is one of the problems I cannot deal with. I would appreciate any help, thank you.

Let $\mu$ be a Radon measure on $\mathbb{R}$, suppose that $A$ is a $\mu$–measurable subset of $[a,b]$ and let $h$ be a positive number. Prove that $$\frac{1}{2h}\int_a^b\mu(A\cap(x-h,x+h))\,\text{d}x\le \mu(A).$$

share|improve this question
add comment

1 Answer 1

up vote 2 down vote accepted

If you apply Fubini (or Tonelli) in the second equality below (everything is positive) you get $$ \frac1{2h}\,\int_{[a,b]}\mu(A\cap(x-h,x+h))\,dx=\frac1{2h}\,\int_{[a,b]}\int_{(x-h,x+h)}\,1_A(t)\,d\mu(t)\,dx\\ \leq\frac1{2h}\int_{(a-h,b+h)}\int_{[t-h,t+h]}\,1_A(t)\,dx\,d\mu(t)=\int_{(a-h,b+h)}\,1_A(t)\,d\mu(t)=\mu(A\cap(a-h,b+h))=\mu(A). $$

share|improve this answer
    
So do we have in fact the equality? –  Kuba Helsztyński Nov 19 '12 at 0:26
    
No, we don't. Now I see where my mistake was. When I'm doing Tonelli, the region on the right-hand-side is smaller than the one I'm using. So I'm changing that equal sign to a $\leq$. If that's not clear enough, let me know and I'll try to clarify. –  Martin Argerami Nov 19 '12 at 0:32
    
Do you mean the reason for inequality sign is that you enlarge the region from $[a,b]$ to $(a-h,b+h)$? –  Kuba Helsztyński Nov 19 '12 at 0:37
    
No. The region of integration is a parallelogram. When you swap the variables for Tonelli, its expression complicates a little. If I have to explain this to my Calculus II students, I would say that the region (in Stewart's Calculus language) is Type I but not Type II. So, for the second integrals I'm switching to a slightly larger parallelogram, hence the inequality. If you draw the region, you'll see it quickly. –  Martin Argerami Nov 19 '12 at 0:40
    
Ah, ok, now I see. We are adding two small triangles to the initial parallelogram and integrating over slightly larger one. All is clear now, thank you for your help! –  Kuba Helsztyński Nov 19 '12 at 1:22
show 1 more comment

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.