0% found this document useful (0 votes)
97 views9 pages

Assignment 08 03

This document contains Andrew van Herick's solutions to several math problems from his Math 710 class. The problems involve showing that sums and products of simple functions are simple, that preimages of measurable functions are measurable, and proofs of Egorov's Theorem and Lusin's Theorem about convergence of functions.

Uploaded by

boris peña
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
0% found this document useful (0 votes)
97 views9 pages

Assignment 08 03

This document contains Andrew van Herick's solutions to several math problems from his Math 710 class. The problems involve showing that sums and products of simple functions are simple, that preimages of measurable functions are measurable, and proofs of Egorov's Theorem and Lusin's Theorem about convergence of functions.

Uploaded by

boris peña
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
You are on page 1/ 9

Andrew van Herick

Math 710
Dr. Alex Schuster
Nov. 2, 2005

Assignment 8
Section 3: 20, 24, 25, 30, 31

20. Show that the sum and product of two simple functions are simple. Show that

A\B = A B

A[B = A + B A B

Ac = 1 A

Proof.

Let E R and suppose f; g : E ! R are two simple functions. By Proposition 19 (p.


67, Royden) both f g and f + g are measurable. Because f and g are simple f (E) and
g (E) are …nite sets. It follows that both

fuv 2 R : u 2 f (E) ; v 2 g (E)g and


fu + v 2 R : u 2 f (E) ; v 2 g (E)g

each have cardinality which is at most jf (E)j jg (E)j ; in other words, …nite. It is easy to
see that

(f g) (E) fuv 2 R : u 2 f (E) ; v 2 g (E)g and


(f + g) (E) fu + v 2 R : u 2 f (E) ; v 2 g (E)g

Hence it follows that (f g) (E) and (f + g) (E) must be …nite as well. By de…nition
f g; f + g are simple functions.

Let A; B R:

First we’ll show that A\B = A B: Let x 2 R: If x 2 A \ B then x 2 A and x 2 B;


so A (x) = 1 and B = 1: Hence

A\B (x) = 1 = 1 1 = A (x) B (x) = A B (x) :

If x 2
= A \ B; then A\B = 0: By de Morgan’s laws x 2 Ac [ B c ; i.e. x 2 = A or x 2 = B:
Either way A (x) B (x) = 0. Hence A\B (x) = A B (x) : This covers all cases of x
and establishes that A\B (x) = A B (x) for all x 2 R: By de…nition A\B = A B:
2

Now we’ll show that Ac =1 A: Let y 2 R: If x 2 A then Ac (y) = 0 and A (y) =1


by de…nition, so
Ac (y) = 0 = 1 1=1 A (y) = (1 A ) (y) :

If x 2 Ac ; then then Ac (y) = 1 and A (y) = 0; so

Ac (y) = 1 = 1 0=1 A (y) = (1 A ) (y) :

This covers all cases of y and establishes that Ac (y) = (1 A ) (y) for all y 2 R: Hence
Ac = 1 A:

To show that A[B = A B. By de Morgan’s laws A [ B = Ac \ B c : By what we


proved above

A[B 1 = (A[B)c

= Ac \B c

= Ac Bc

= ( 1+ A ) (1 B)

= A B + B + A 1

It follows immediately that A[B = A + B A B; as desired.


Andrew van Herick 3

24. Let f be measurable and B a Borel set. Then f 1 [B] is a measurable set.

Proof.

Let f be measurable and B a Borel set. Let E := E R:f 1 [E] is measurable : We’ll
show that E contains B; the collection of Borel sets.

First we’ll show that E contains every open interval. Let I R be an open interval.
Then I = (a; b) for some (possibly extended) real numbers a; b such that a < b: Note that
I = ( 1; b) \ (a; 1) : As f is measurable f 1 (( 1; b)) = fx : f (x) < bg is measurable
by de…nition. Similarly f 1 ((a; 1)) is also measurable. By the fact that the measurable
sets form a -algebra

1 1 1 1
f (I) = f (( 1; b) \ (a; 1)) = f (( 1; b)) \ f ((a; 1))

is also measurable. Hence I 2 E: This establishes that I 2 E for every open interval I R:

Now we’ll show that E is a -algebra. Let A0 be a countable index set and fE g 2A0
be a collection of sets from E: Then by the fact that the measurable sets form a -algebra
0 1
[ [
f 1@ E A= f 1 (E )
2A0 2A0

is measurable, since f 1 (E ) 2A0 is a countable collection of measurable sets. Now let


c
E 2 E: By the same reasoning f 1 (E c ) is measurable since f 1 (E c ) = f 1 (E) : This
establishes that E is a -algebra.

Thus E is a -algebra containing every open interval. We proved in Exercise 10.4.5.


(p. 311, Wade) that every open set in R is a countable union of open intervals. It follows
immediately that every open subset of R belongs to E: This establishes that E is a -algebra
containing every open subset of R:

As B is the smallest -algebra containing all the open sets, it follows that B E: As
B 2 B; it follows immediately that B 2 E: Equivalently B is measurable.
4
Andrew van Herick 5

25. Show that if f is a measurable real valued function and g a continuous function de…ned
on ( 1; 1) ; then g f is measurable.

Proof.

Let f be a measurable real valued function and g a continuous function de…ned on


( 1; 1) : As of g f is R; an open set the domain of g f is measurable.

Let 2 R and E = fx : g f (x) > g : It is easy to see that

1
E = fx : g f (x) 2 (a; 1)g = (g f ) ((a; 1)) :

By Lemma 1 below E = f 1 g 1 (( ; 1)) : Because g : R ! R is continuous and ( ; 1)


is open, by Theorem 10.58 (p. 317, Wade) g 1 ((a; 1)) is is open in R: By de…nition
g 1 ((a; 1)) is a Borel set. Problem 24 above E = f 1 g 1 (( ; 1)) is a measurable set.

Lemma 1 Let E; F; G be sets and suppose f : E ! F and g : F ! G: It A G; then


(g f ) 1 (A) = f 1 g 1 (A) :

Proof. Let x 2 (g f ) 1 (A) : By de…nition (g f ) (x) = g (f (x)) 2 A: By the de…nition of


inverse images we get f (x) 2 g 1 (A) and x 2 f 1 g 1 (A) : Now let y 2 f 1 g 1 (A) :
Then f (y) 2 g 1 (A) : By de…nition g (f (y)) 2 A. As g (f (y)) = (g f ) (y) we have
(g f ) (y) 2 A: By de…nition y 2 (g f ) (A) : This establishes that (g f ) 1 (A) =
f 1 g 1 (A) :
6
Andrew van Herick 7

30. Prove Egoro¤’s Theorem: If hfn i is a sequence of measurable functions that converge
to a real-valued function f a.e. on a measurable set E of …nite measure, then given
> 0; there is a subset A E with mA < such that fn converges uniformly on
EnA: [Hint: Apply Proposition 24 repeatedly with "n = n1 and n = 2 n .]

Proof.

Let > 0: For each k 2 N de…ne "k = k1 ; k = 2 k . Clearly "k ; k > 0: Using
Proposition 24 (p. 73, Royden) for each k 2 N choose Ak E such that m (Ak ) < k , and
choose Nk 2 N such that for all x 2
= Ak and all n Nk ;

jfn (x) f (x)j < "k :


S
Let A := k2N Ak : Because each Ak E it follows that A E: Note that each Ak is
measurable. By Proposition 13 (p. 62, Royden)
1
X 1
X 1
X
k
m (A) = m (Ak ) < k = 2 =
k=1 k=1 k=1

This establishes that m (A) < :

Now we’ll show that fn converges uniformly on EnA: Let " > 0: Then there exists
1
M 2 N such that M < ": Choose N 2 N such that N = NM : Let n N and x 2 EnA:
T
As x 2 A and A = k2N (Ak )c ; we have x 2 (AM )c : From our construction of AM above
c c

1
jfn (x) f (x)j < "M = < ":
M
This establishes that for all " > 0 there is a N 2 N such that n N implies jfn (x) f (x)j <
" for all x 2 EnA: By de…nition fn is uniformly convergent on EnA:

This establishes the existence of a subset A E with mA < such that fn converges
uniformly on EnA:
8
Andrew van Herick 9

31. Prove Lusin’s Theorem: Let f be a measurable real-valued function on an interval


[a; b] : Then given > 0; there is a continuous function ' on [a; b] such that such that
m fx : f (x) 6= ' (x)g < : [Hint: Use Egoro¤’s Theorem, Propositions 15 and 22,
and Problem 2.40.]

Proof.

Let f be a measurable real-valued function on an interval [a; b] : Let > 0: As f is real


valued f takes on the values 1 on ;; a set of measure zero. Let k = k1 : By Proposition
22 (p. 71, Royden) for each k 2 N choose a continuous function hk on [a; b] such that
jf hk j < k : It follows almost immediately from this construction that hk (x) ! f (x)
for every x 2 [a; b] : By Egoro¤’s Theorem there is a subset A [a; b] with m (A) < 2
such that hk converges to f uniformly on [a; b] nA: By Proposition 15 (p. 63, Royden) there
exists a closed set F [a; b] nA such that m ([a; b] nAnF ) < 2 : Because all of the hk ’s are
continuous on [a; b] ; they are also continuous on F [a; b] : Hence f is continuous on F by
Exercise 7.7.1(a) (p. 191, Wade). By Problem 2.40 (p. 49, Royden) there is a continuous
function g : R ! R such that f (x) = g (x) for each x 2 F: Let ' = gj [a; b] : Then ' is
clearly continuous on [a; b] : Clearly fx 2 [a; b] : f (x) 6= ' (x)g ([a; b] nF ) [ A: Because
([a; b] nF ) [ A = ([a; b] nAnF ) [ A; it follows that

m (fx : f (x) 6= ' (x)g) m (([a; b] nF ) [ A) = m (([a; b] nAnF ) [ A)


< + =
2 2
This establishes the existence of a continuous function ' : [a; b] ! R such that m fx : f (x) 6= ' (x)g <
as desired.

You might also like