Sol 5

Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

MATH 417 Assignment #5

1. Let X be an uncountable set, and let S := {E ⊆ X : E or E c is countable}.


(a) Show that S is a σ-algebra.
(b) Let µ be the function from S to [0, ∞) defined by µ(E) = 0 if E is countable and
µ(E) = 1 if E c is countable. Show that µ is a measure on S.
Proof . In light of the definition of S, X and ∅ belong to S. Moreover, if E ∈ S, then
E c ∈ S. By the definition of µ, we have µ(∅) = 0. Suppose Ek ∈ S for k ∈ IN. We
wish to show that E := ∪∞ k=1 Ek lies in S. There are two possibilities: either all the
sets Ek (k ∈ IN) are countable, or there exists some k0 ∈ IN such that Ekc0 is countable.
P∞
In the former case, E is countable; hence E ∈ S and µ(E) = 0 = k=1 µ(Ek ). In the
latter case, E c ⊆ Ekc0 , so E c is countable. Hence, E ∈ S. Suppose in addition that
the sets Ek (k ∈ IN) are mutually disjoint. Then Ek ⊆ Ekc0 for k 6= k0 . Therefore,
P∞
µ(E) = 1, µ(Ek0 ) = 1, and µ(Ek ) = 0 for k 6= k0 . Consequently, µ(E) = k=1 µ(Ek ).
This shows that S is a σ-algebra and µ is a measure on S.

2. Let (X, S, µ) be a measure space with S being an algebra.


(a) If E, F ∈ S and µ(E∆F ) = 0, then µ(E) = µ(F ).
(b) Prove that µ(E∆G) ≤ µ(E∆F ) + µ(F ∆G) for all E, F, G ∈ S.
Proof . (a) The set E∆F is the union of two disjoint sets E \ (E ∩ F ) and F \ (E ∩ F ).
Since µ(E∆F ) = 0, we have µ(E \ (E ∩ F )) = 0 and µ(F \ (E ∩ F )) = 0. It follows
that

µ(E) = µ(E \ (E ∩ F )) + µ(E ∩ F ) = µ(F \ (E ∩ F )) + µ(E ∩ F ) = µ(F ).

(b) We have

E \ G ⊆ (E \ F ) ∪ (F \ G) and G \ E ⊆ (G \ F ) ∪ (F \ E).

It follows that

µ(E \ G) ≤ µ(E \ F ) + µ(F \ G) and µ(G \ E) ≤ µ(G \ F ) + µ(F \ E).

Consequently,

µ(E∆G) = µ(E \ G) + µ(G \ E)


≤ µ(E \ F ) + µ(F \ G) + µ(G \ F ) + µ(F \ E)
= µ(E∆F ) + µ(F ∆G).

1
3. For a sequence (An )n=1,2,... of subsets of a set X, define

lim inf An := ∪∞ ∞
n=1 ∩k=n Ak and lim sup An := ∩∞ ∞
n=1 ∪k=n Ak .
n→∞ n→∞

Now let (X, S, µ) be a measure space with S being a σ-algebra and let (En )n=1,2,...
be a sequence of sets in S. Show the following:

(a) µ lim inf n→∞ En ≤ lim inf n→∞ µ(En ).
(b) µ lim supn→∞ En ≥ lim supn→∞ µ(En ), provided µ(∪∞

n=1 En ) < ∞.

Proof . (a) Let F := lim inf n→∞ En and Fn := ∩∞ ∞


k=n Ek for n ∈ IN. Then F = ∪n=1 Fn
and Fn ⊆ Fn+1 for n ∈ IN. It follows that

µ(F ) = lim µ(Fn ).


n→∞

But Fn ⊆ En for each n ∈ IN. Hence, µ(Fn ) ≤ µ(En ) for n ∈ IN. Consequently,

µ lim inf En = µ(F ) = lim µ(Fn ) ≤ lim inf µ(En ).
n→∞ n→∞ n→∞

(b) Let G := lim supn→∞ En and Gn := ∪∞ ∞


k=n Ek for n ∈ IN. Then G = ∩n=1 Gn
and Gn ⊇ Gn+1 for n ∈ IN. It follows that µ(Gn ) ≥ µ(En ). By our assumption,
µ(G1 ) = µ(∪∞
n=1 En ) < ∞. Therefore,


µ lim sup En = µ(G) = lim µ(Gn ) ≥ lim inf µ(En ).
n→∞ n→∞ n→∞

4. Let µ∗ be an outer measure on a set X. Suppose that (Ek )k=1,2,... is a sequence of


disjoint µ∗ -measurable sets. Show that for every A ⊆ X,

 X
µ∗ A ∩ (∪∞ E
k=1 k ) = µ∗ (A ∩ Ek ).
k=1

Proof . By the σ-subadditivity of µ∗ we have



 X
µ∗ A ∩ (∪∞ E
k=1 k ) ≤ µ∗ (A ∩ Ek ).
k=1

Since the sets Ek (k ∈ IN) are µ∗ -measurable, it was proved in the class that
n
X

(∪nk=1 Ek ) µ∗ (A ∩ Ek )

µ A∩ =
k=1

2
holds for every n ∈ IN. But ∪nk=1 Ek ⊆ ∪∞
k=1 Ek . Consequently,

n
X
µ∗ (A ∩ Ek ) ≤ µ∗ A ∩ (∪∞

k=1 Ek ) .
k=1

Letting n → ∞, we obtain

X
µ∗ (A ∩ Ek ) ≤ µ∗ A ∩ (∪∞

k=1 Ek ) .
k=1

5. For n = 1, 2, . . ., let
 
2k − 1 1 2k − 1 1
Ink := − n, + n , k = 1, . . . , 2n−1 .
2n 4 2n 4

Let
n−1
∞ 2[
[
U := Ink and K := [0, 1] \ U.
n=1 k=1

(a) Prove that there is no open interval J such that J ⊆ K.


(b) Prove that λ(K) ≥ 1/8, where λ denotes the Lebesgue measure.
Proof . (a) let J = (a, b), where a, b ∈ [0, 1] and a < b. By the Archimedean property
of the real numbers, there exists a positive integer n such that 2n (b − a) > 4. Let k0
be the least element of the set {k ∈ IN : (2k − 1)/2n ≥ a}. Then it is easily seen that
a < (2k0 + 1)/2n < b. But (2k0 + 1)/2n ∈ U . Hence, (a, b) ∩ U 6= ∅. This shows that
there is no open interval J such that J ⊆ K.
n−1
(b) For n ∈ IN, let En := ∪2k=1 Ink . Then

n−1
2X 2 1
µ(En ) = µ(Ink ) = 2n−1 n
= n.
4 2
k=1

Note that E1 ∪ E2 = (3/16, 13/16). Hence, λ(E1 ∪ E2 ) = 10/16 = 5/8. Moreover, we


have U = ∪∞n=1 En . It follows that

∞ ∞
X 5 X 1 5 1 7
λ(U ) ≤ λ(E1 ∪ E2 ) + λ(En ) = + n
= + = .
n=3
8 n=3 2 8 4 8

Consequently, λ(K) = λ([0, 1]) − λ(U ) ≥ 1/8.

You might also like