Scribd
Upload
6 views2 pages

Advanced Measure Theory Problems

The document presents problems and solutions related to advanced measure theory, specifically focusing on µ*-measurable sets and the Radon-Nikodym derivative. It discusses the collection of µ*-measurable sets in relation to subsets of R and demonstrates that ν is absolutely continuous with respect to µ, providing the Radon-Nikodym derivative. Additionally, it highlights a contradiction with the Lebesgue-Radon-Nikodym Theorem regarding the relationship between Lebesgue measure and counting measure.

Uploaded by

Dhruv Chauhanrajput
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views2 pages

Advanced Measure Theory Problems

The document presents problems and solutions related to advanced measure theory, specifically focusing on µ*-measurable sets and the Radon-Nikodym derivative. It discusses the collection of µ*-measurable sets in relation to subsets of R and demonstrates that ν is absolutely continuous with respect to µ, providing the Radon-Nikodym derivative. Additionally, it highlights a contradiction with the Lebesgue-Radon-Nikodym Theorem regarding the relationship between Lebesgue measure and counting measure.

Uploaded by

Dhruv Chauhanrajput
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

MMATH18-303(II): ADVANCED MEASURE THEORY

NIKITA AGARWAL

Problems for Tutorial on September 1, 2025


(1) Let X = R, S = {∅, [1, 2]}, µ(∅) = 0, µ([1, 2]) = 1.
Then
µ∗ (∅) = 0, µ∗ (E) = ∞,
for every nonempty subset E of R which is not contained in [1, 2], and µ∗ (E) = 1,
for all subsets E of R which are contained in [1, 2] (This was discussed in the class).

Find M, the collection of all µ∗ -measurable sets.

Solution: We need to find subsets E of X for which


µ∗ (A) ≥ µ∗ (A ∩ E) + µ∗ (A ∩ E c ),
for all subsets A of X.
Clearly if µ∗ (A) = ∞ or if A = ∅, the above inequality is always satisfied.

Let E be a µ∗ -measurable set. Consider A = [1, 2]. Then µ∗ (A) = 1. Hence


the RHS should be 1. Therefore exactly one of the terms on the RHS is 1 and the
other is 0.

Case 1: µ∗ ([1, 2] ∩ E) = 1 and µ∗ ([1, 2] ∩ E c ) = 0.


This is true if and only if [1, 2] ∩ E ̸= ∅ and [1, 2] ∩ E c = ∅.
This is true if and only if [1, 2] ⊂ E.

Case 2: µ∗ ([1, 2] ∩ E) = 0 and µ∗ ([1, 2] ∩ E c ) = 1.


This is true if and only if [1, 2] ∩ E = ∅ and [1, 2] ∩ E c ̸= ∅.
This is true if and only if [1, 2] ⊂ E c .

Hence
M ⊂ {E ⊂ X : [1, 2] is contained in either E or E c }.
Conversely, suppose E is such that [1, 2] is contained in E. Let A be a nonempty
subset of [1, 2]. Then A ∩ E = A and A ∩ E c = ∅. Hence the inequality is satisfied
and E is µ∗ -measurable.
Finally suppose E is such that [1, 2] is contained in E c . Let A be a nonempty
subset of [1, 2]. Then A ∩ E = ∅ and A ∩ E c = A. Hence the inequality is satisfied
and E is µ∗ -measurable.
Thus
M ⊃ {E ⊂ X : [1, 2] is contained in either E or E c }.
Therefore
M = {E ⊂ X : [1, 2] is contained in either E or E c }.
Date: September 1, 2025.
1
2 NIKITA AGARWAL

(2) Let (X, B, µ) be a finite measure space. Let E1 , . . . , Ek be a collection of measur-


able sets and let a1 , . . . , ak be a collection of real numbers. For E ∈ B, define
k
X
ν(E) = ai µ(E ∩ Ei ).
i=1
Show that ν ≪ µ. Find the Radon-Nikodym derivative of ν with respect to µ,
dν
denoted as .
dµ
Solution: Observe that µ is a finite positive measure and thus ν is a finite
signed measure.
Further if µ(E) = 0 for some E ∈ B, then µ(E ∩ Ek ) = 0, for all k. Hence
ν(E) = 0. Thus ν ≪ µ.
Since µ is a finite measure and ν is a finite signed measure, the Lebesguue-Radon-
dν
Nikodym theorem is applicable, and the derivative exists.
dµ
Now observe that
k
X k
X Z
ν(E) = ai µ(E ∩ Ei ) = ai χE∩Ei dµ
i=1 i=1 X

k Z Z k
!
X X
= ai χEi dµ = ai χEi dµ.
i=1 E E i=1
Hence !
k
X
dν = ai χEi dµ.
i=1
Thus
k
dν X
= ai χEi , µ − a.e.
dµ i=1
(3) Let X = [0, 1], M be the collection of all Lebesgue measurable subsets of X, and
let ν be the Lebesgue measure and µ be the counting measure on M. Show that
ν ≪ µ, but there is no function f for which dν = f dµ. Why does this contradict
the Lebesgue-Radon-Nikodym Theorem?

You might also like