You are on page 1of 39

Math 7311 Test Bank Questions

Fall 2016

MATH 7311 Test Bank Questions

II.

Fall 2016

Lebesgue Measure, -algebras, Borel Sets, and Measurable


Functions.

Exercise II.1. Let A be a algebra of subsets of X, and let f : X X be arbitrary. Prove that the set
{f 1 [S] | S A}
is a algebra.

Proof. Define f 1 (A) := {f 1 [S] | S A}. Let A =

Ai such that each Ai A. Thus, A A. Since

i=1

preimages of functions preserve unions, we have


f

(A) = f

[


Ai

f 1 (Ai ) f 1 (A).

For complements, let Y \B A for all B A, since A is a algebra. Since preimages of functions preserve
intersections, we have
f

(Y \A) = f

 [
 \

\

1
Y \ Ai = f
(Y \Ai ) =
f 1 (Y \Ai ) f 1 (A)
1

Therefore, f 1 (A) is a algebra.

Exercise II.2. Suppose that g : X R is -measurable and that f : R R is Borel measurable, meaning
that f 1 (a, ) is a Borel set for each real number a. Prove that f g is measurable.
Proof. Let (X, A, ) and (R, L, l) be measure spaces. Since f is Borel measurable, and Borel sets are elements
of A, we have that
f 1 (, a) L
for all a R. This implies that f is l-measurable. Note that since g is -measurable, g 1 maps measurable
sets in R to measurable sets in X. So, we have
(f g)1 (, a) = (g 1 f 1 )(, a) = g 1 f 1 (, a)) A,
since f 1 (, a) is a measurable set in R. Therefore, we conclude that f g is -measurable.

MATH 7311 Test Bank Questions

Fall 2016

Exercise II.3. Prove that every Lebesgue measurable subset of R is sandwiched between an F -set and a
G -set of the same Lebesgue measure.
Proof.

MATH 7311 Test Bank Questions

Fall 2016

Exercise II.9. Let E R be a measurable set with the property that


l(E I)

l(I)
,
2

for every open interval I. Prove that l(E) = 0.


Proof. Since E is measurable, for all

> 0, there exists an open cover O =

In such that l(O\E) < 2 . We want to show that l(O) < . So, we have

In of disjoint open intervals

nN

l(O) = l(O E) + l(O E c ) = l(E) + l(O E c ) l(O) l(E) <

.
2

So, for O, we have


l(O) =

X
nN

l(In )

2l(E In ) = 2l(E O) = 2l(E).

nN

Thus, we have that l(O) l(E) l(E), which implies that l(E) l(O) l(E) < 2 . Finally, we conclude
that

l(O) = l(E) + l(O E c ) < + = .
2 2
Therefore, since l(O) < for all > 0, we have that l(E) = 0.

MATH 7311 Test Bank Questions

Fall 2016

Exercise II.11. Let A (0, 1) be a measurable set and l(A) = 0. Show that

l{x2 : x A} = 0 and l{ x : x A} = 0.

Proof. First note that f : (0, 1) (0, 1) with f (x) = x2 and g : (0, 1) (0,1) withg(x) = x are
homeomorphisms. So, for any open set O, we have that O2 = {x | x O} and O = { x | x O} are
open sets, and thus measurable. For any > 0, there exists an open set O (0, 1) such that
[
O=
(ai , bi )
i4

for disjoint (ai , bi ), l(O\A) < , and 4 is countable. Since l(A) = 0, we have that l(O) < . Also,
(ai , bi )2 = {x2 | x (ai , bi )} = {x2 | a2i < x2i < b2i } = (a2i , b2i )
and
p

p
(ai , bi ) = { x | x (ai , bi )} = ( ai , bi ).

Thus,
X
X
X
X
l (A2 ) l(O2 ) = l( (a2i , b2i ))
(b2i a2i ) =
(bi + ai )(bi ai ) 2
(bi ai ) < 2.
Since is arbitrary, as approaches 0, so does l (A2 ). So, A2 is measurable and l(A2 ) = 0.
Next, we have
X p
Xp

( bi ai ).
l ( x) l( ( ai , bi ))
p

To find the bound, choose O such that l(O) < /2. If > 1, then l ( A) < 1 < since A (0, 1).
Thus,
 [
 [

[ p 

p
p
l ( A) l
( ai , bi ) = l
( ai , bi ) + l
( ai , bi )
i

bi /2

ai /2

 X

p
l(0, /2] + l
( ai , bi )
ai /2

1
+
(bi ai )

2
bi + ai
ai /2

l(O)
+
2 2 2

< +
2 2
=

Therefore, l( A) = 0.

MATH 7311 Test Bank Questions

Fall 2016

Exercise II.22. Suppose A and B are measurable subsets of Rn , each one of strictly positive but finite
measure. Prove that there exists a vector c Rn such that
l((A + c) B) > 0.
(Hints: Consider the outer measure of A and B. Or, alternatively, consider the convolution 1A 1B .)
Proof. Note that the translation invariance of Lebesgue measure (Rn ) says that for any c Rn and measurable set A L(Rn ), we have that l(A) = l(A + c). Thus, 0 < l(A + c) < . So, since B is also measurable,
we have
l(A + c) = l((A + c) B) + l((A + c) B c ).
By monotonicity of l, we have that
l((A + c) B) l(A + c) and l((A + c) B c ) l(A + c).
Therefore, since l(A + c) > 0, it follows that both l((A + c) B c ) > 0 and l((A + c) B) > 0.

Exercise II.24. Suppose fn : X R is a measurable function for each n N, where (X, A, ) is a measure
space. Prove that the set
n
o

S = x lim fn (x) exists
n

is a measurable set.
Proof. We can rewrite F in terms of n and . Note that f is measurable. For > 0,
Fn () = {x X : |fm (x) f (x)| m n N}
\
=
{x X : |fm (x) f (x)| < }
mnN

Each Fn is measurable since fn f (X) is a measurable function, so Fn A for all n N. Also,


F1 () F2 () Fn ()
is an increasing nest. So, we have that
F =

[
n=1

is measurable, since fn converges to f pointwise on X.

Fn ()

MATH 7311 Test Bank Questions

Fall 2016

Exercise II.27. Prove or give a conterexample: E F is -measurable in (X, A, )


E and F are measurable with respect to and respectively.

(Y, B, ) iff both

Proof. Let F be any nonmeasurable subset of Y . Then, the measure ( )({x} F ) = 0 (E = {x}) for
each singleton set {x} in X, being a subset of a null set. Hence
O
{x} F A(X)
B(X).
Thus, a counterexample has been provided.

MATH 7311 Test Bank Questions

III.

Fall 2016

The Integral: Convergence Theorems, Product Measures and


Fubinis Theorem.

Exercise III.1. Let 1[n,n] () denote the characteristic function of the interval [n, n] (n N). Consider
the sequence of functions
 
x
, (x R).
fn (x) := [n,n] (x) sin
n
(a) Determine f (x) = lim fn (x) and show that the sequence (fn )nN converges uniformly on compact
n
subsets of R. Does the sequence converge uniformly on R?
(b) Show that
Z

f (x) dx = lim

fn (x) dx.

Are the assumptions of Lebesgues dominated convergence theorem satisfied?


Proof.
x
(a) f (x) = lim 1[n,n] (x) sin( x
n ) = lim sin( n ) = 0. Let A be a compact subspace of R such that
n

A [M, M ] for some M R+ . So, we have





 
 


x M
x
x


kfn (x) 0ksup = sup 1[n,n] (x) sin
sup sin
sup =
0.


n
n
n
xA
x[M,M ]
x[M,M ] n
So, fn (x) 0 uniformly on A. Next, consider x = n2 R. Thus,

 
 


x
sin

= 1 6= 0.
kfn (x) 0ksup = sup 1[n,n] (x) sin


n
2
xR
So, fn does not converge uniformly to 0 on R.
(b) We have
Z


1[n,n] sin

x
n

dx =

and


sin

x
n

n
dx = cos

0 dx = 0.

So,

f (x) dx = lim

fn (x) dx.

x
n

 n
n n

= =0


n

MATH 7311 Test Bank Questions

Fall 2016

Exercise III.5. Show that if


f (x) =

1
1

if x is rational,
if x is irrrational,

then f is not Riemann integrable on the interval [0, 1]. Is f Lebesgue integrable?
Proof. For the Riemann integral, finding the supremum of lower sum involves taking a sum over finitely
many closed intervals which partition [0, 1] of the infimum of f over the respective closed interval. So, the
lower Riemann integral sup s() = 1 because for each closed subinterval of R, we can always find a rational

number and an irrational number So, the infimum for each closed subinterval is 1 . Similarly, the infimum
of the upper sum is the same except we take the supremum of f in each closed subinterval. So the upper
Riemann integral inf s() = 1 because for each closed subinterval of [0, 1], the supremum of f is 1. So, since

the upper and lower Riemann integrals are not equal, f is not Riemann integrable.
Let A be the set of rational numbers in [0, 1], and let B be the set of irrational numbers in [0, 1]. Since
Q is countable, A is countable (and measurable) and thus m(A) = 0. So, we have that
m([0, 1]) = 1 = m([0, 1] Q) + m([0, 1]\Q) = m(A) + m(B) = 0 + m(B).
Thus, m(B) = 1 and B is measurable. So,
Z
Z
Z
f d =
1 d +
1 d = 1 (A) 1 (B) = 1.
[0,1]

Exercise III.7. Let f (x) be a real-valued measurable function on a finite measure space (X, A, ). Show
that
Z
cos2n (f (x)) d = {x | f (x) Z},
lim
n

where Z is the set of integers.


Proof. Let gn (x) = cos2n (f (x)). Since cos2n (x) and f (x) are measurable functions, gn (x) is measurable
as well by composition gn = cos2n (x) f (x). So, {gn } is a sequence of nonnegative measurable functions.
Note that
0 cos2n (f (x)) 1 and 1 L(X)
since X is a finite measure space. So,
cos2 (f (x)) = 1 f (x) = n (n Z) f (x) = n
and
0 cos2 (f (x)) < 1 f (x) 6= n (n Z) f (x) 6= n.
So, for x {x | f (x) Z}, the limit lim gn (x) = 1 and for x {x | f (x) 6 Z}, the limit lim gn (x) = 0.
n

So, lim gn (x) = 1{x|f (x)Z} (x) + 0, which is a measurable function. By Lebesgues Dominated Convergence
n
Theorem,
Z
Z
Z
lim
gn (x) =
1{x|f (x)Z} (x) d =
1 d = {x|f (x) Z}.
n

It is worth noting that f

{x|f (x)Z}

X
1

(Z) = {x|f (x) Z}, which is measurable since Z is measurable.

MATH 7311 Test Bank Questions

Fall 2016

Exercise III.8.
(a) Show that f (x) = xr is a Lebesgue integrable function on [0, 1] if 0 r < 1.
(b) If 0 r < 1, let
Z

an =
0

dx
.
n1 + xr

Compute lim an . Be sure to quote the appropriate integration theorems to justify your calculations.
n

Proof.

Exercise III.9. Let fn : R R be defined by



fn (x) =

1
n

if |x| n,
if |x| > n.

(a) Show that fn converges to 0 uniformly on R, and that


Z
fn (x) dx = 2
lim
n

while

Z 
R


lim fn (x) dx 6= 2.

(b) Explain why this example does not contradict the Lebesgue dominated convergence theorem.
Proof.

10

MATH 7311 Test Bank Questions

Fall 2016

Exercise III.12. Prove that

(sin nt)f (t) dt = 0

lim

for every Lebesgue integrable function f on R. (Hint: Do the problem first for step functions.)
Proof.

Z
cos(xy)g(y) dy for x R where g is an integrable function on R. Show

Exercise III.14. Define f (x) =


that f is continuous.

Proof. Note that for each x0 R, cos(xy) is continuous at x0 , for any y R. In other words,
Z for each > 0,
there exists a > 0 with |x x0 | < such that | cos(xy) cos(x0 y)| < /M , where M =
g(y)dy, which
exists since g is integrable over R. So, for |x x0 | < , we have
Z
Z
Z


= [cos(xy) cos(x0 y)]g(y)dy
cos(xy)g(y)dy
cos(x
y)g(y)dy
0


R
R
ZR
<
/M g(y)dy
R
Z
= /M
g(y)dy

= .
Therefore, for every x0 R, f is continuous.

Z
Exercise III.15. Let f be a differentiable function on [1, 1]. Prove that lim

Proof.

11

<|x|1

1
f (x) dx exists.
x

MATH 7311 Test Bank Questions

Fall 2016

Exercise III.18. Compute the following limit and justify your calculations:
Z 
x
x n
1+
dx.
lim
sin
n 0
n
n
Proof.

Exercise III.29. Let f be a nonnegative integrable function on [0, 1]. Suppose that for every n N
Z

(f (x))n dx =

f (x) dx.
0

Show that f is almost everywhere the characteristic function for some measurable set.
Proof. Since the above equality holds for any n N, we have
Z 1
Z 1
Z 1
Z
n
n1
n+1
(f (x)) dx =
(f (x))
dx =
(f (x))
dx =
0

f (x) dx.

So,
Z

(f (x))n1 (f (x) 1) dx =

and
Z

(f (x))n (f (x) 1) dx

((f (x))n (f (x))n1 )(f (x) 1) dx =

(f (x))n1 (f (x) 1)2 dx = 0.

Since g(x) = (f (x))n1 (f (x) 1)2 L+ , g(x) = 0 almost everywhere. So, f (x) 1 = 0 f (x) = 1
or f (x)n1 = 0 f (x) = 0 almost everywhere for all x [0, 1]. Since the only two values in the range
of the characteristic are 0 and 1, we can partition [0, 1] into sets A0 , A1 with A0 = {x : f (x) = 0} and
A1 = {x : f (x) = 1}. f (x) is a simple function such that f (x) = 0 1A0 + 1 1A1 . Since f (A0 ) = 0 , 0 = 0.
Also, f (A1 ) = 1 = 1. So, we have that
f (x) = 1A1 .
Z 1
Since
f (x) dx = (A1 ), we have that A1 is a measurable set.
0

12

MATH 7311 Test Bank Questions

Fall 2016

Exercise III.30. Provide an example of a sequence {fn } of measurable functions on [0, 1] such that fn f
Z 1
Z 1
a.e., and fn 0, yet lim inf
fn dl 6=
f dl.
0

Proof. For every n N, define

n if x (0, 1/n)
0 otherwise

fn (x) =
Z
Note that fn 0, the zero function. But,
Z 1
Z 1
lim inf
fn dl 6=
f dl.
0

Z
fn dl = 1 for all n N while

0 dl = 0. Thus,
0

Exercise III.31. Determine

Z
lim

sin
1

n3
dx.
n 1 + n2 x3

x

Proof.

Exercise III.35. Let p(x1 , x2 ) be a polynomial in two variables, but not the zero polynomial. Prove that
the set of points x R2 with p(x) = 0 has 2-dimensional Lebesgue measure (l l)(p1 (0)) = 0.
Proof.

13

MATH 7311 Test Bank Questions

Fall 2016

Exercise III.47. Let (X, A, ) be a measure space for which (X) < . Show
Lq (X, A, ) Lp (X, A, )
whenever 1 p q .
Proof.

Exercise III.48. Let (X, A, ) be a probability space, meaning that the nonnegative measure (X) = 1.
Suppose that f Lp (X, A, ).
(a) Use Jensens inequality to prove that
Z

! p1
|f |p d

provided that 1 p < q < .


(b) Prove that kf kp kf kq for all 1 p q .
Proof.

14

! q1
|f |q d

MATH 7311 Test Bank Questions

IV.

Fall 2016

Bounded Variation, Absolutely Continuous Functions and Measures, Convex Functions, the Fundamental Theorem of Calculus, and the Radon-Nikodym Derivative.

Exercise IV.3. Show that if f is absolutely continuous on [a, b], then the total variation Vax f where
a x b is given by
Z x
x
Va f =
|f 0 (t)| dl(t).
a
0

Vax f .

x+h
(Hint: Let
(x) =
show F (x) |f (x)| a.e. by using
Z F
ZFirst
Z c Vx f |f (x + h) f (x)| and then
c
c
conclude
|f 0 (x)| dx
F 0 (x) dx VAc f . Then show Vac f
|f 0 (x)| dx.)
a

Proof.

15

MATH 7311 Test Bank Questions

Fall 2016

Exercise IV.8. Show that the product of two absolutely continuous functions on a closed finite interval
[a, b] is absolutely continuous.
Proof. Let f and g be absolutely continuous functions on [a, b]. Then, for each > 0, there exist f and g
such that for each n N, and for each choice of
a x1 < y1 x2 < y2 xn < yn b,
with

n
P

(yi xi ) < f and

i=1

n
P

(yi xi ) < g , we have

i=1
n
X

|f (yi ) f (xi )| <

i=1

n
X

and
.
|g(yi ) g(xi )| <
2N
2M
i=1

So,
n
X

|f (yi )g(y1 ) f (xi )g(xi )| =

i=1

n
X
i=1
n
X

|f (yi )g(yi ) f (yi )g(xi ) + f (yi )g(xi ) f (xi )g(xi )|


|f (yi )|g(yi ) g(xi )| + |g(xi )||f (yi ) f (xi )|.

i=1

Let |f (yi )| M , |g(xi )| N and := min{f , g }. So,

n
P

(yi xi ) < and

i=1
n
X

|f (yi )|g(yi ) g(xi )| + |g(xi )||f (yi ) f (xi )|

i=1

n
X

M |g(yi ) g(xi ) +

i=1

<M
Therefore, f g is absolutely continuous.

16

+N
= .
2M
2N

n
X
i=1

N |f (yi ) f (xi )

MATH 7311 Test Bank Questions

Fall 2016

f (x) dx. Show that

Exercise IV.10. Let f be a nonnegative integrable function on [0, 1] and let I =


0

1 + I2

1 + f 2 (x) dx 1 + I.

Proof. Note that a function


is convex iff its second derivative is positive. So, we define the function

: R+ R by (x) = 1 + x2 .

Exercise IV.11. Suppose f is a nonnegative integrable function on [0, 1]. Prove that
s
Z 1
Z 1p
f (t) dt
f (t) dt.
0

Proof.

17

MATH 7311 Test Bank Questions

Fall 2016

Spring 2014 Exam


Problem 1. Let (X, A, ) be a measure space. Suppose An A, An An+1 for n = 1, 2, . . . , and

S
A=
An . Show
n=1

(A) = lim (An ).


n

Proof. We need to construct a set of disjoint sets from the collection {Ai }. Since {Ai } is increasing, let

n1
n1
S
S
S
Bi = Ai+1 /Ai . So the Bi s are disjoint, and
Bi = A. Also, note that
Bi =
Ai+1 /Ai = An .
i=1

i=1

i=1

Therfore, by the countable subadditivity of , we have


[
 X
 n1

n1
[ 
X
Bi =
Bi = lim (An ).
(A) =
(Bi ) = lim
(Bi ) = lim
i=1

i=1

i=1

i=1

Problem 2. Let (X, A) be a measure space. A function f : X R is measurable if f 1 (, a) A for


all real numbers a R. Show that f is measurable iff f 1 (B) A for all Borel sets B R.
Proof.
()
Let f be measurable and B R a Borel set. Let C be the collection of sets E R such that f 1 (E) is
measurable. Suppose E C. Then f 1 (E c ) = (f 1 (E))c , which is measurable, so E c C. Suppose {Ei}
is a sequence of sets in C. Then f 1 (Ei) = f 1 (Ei), which is measurable, so Ei C. Thus C is a
-algebra. Now for any a, b R with a < b, {x : f (x) > a} and {x : f (x) < b} are measurable. i.e. (a, )
and (, b) are in C. Thus (a, b) C and C is a -algebra containing all the open intervals so it contains all
the Borel sets. Hence f 1 (B) is measurable.
()

18

MATH 7311 Test Bank Questions

Fall 2016

Problem 3. Let E R be a measurable set. Suppose 0 < c < 1 and


l(E I) c l(I)
for all intervals I. Show l(E) = 0.
Proof. Since 0 < c < 1, for ease of notation, we will write c = a1 for a > 1 R. Since E is measurable, for
S
In for disjoint open intervals In such that l(O\E) < a . Thus,
all a > 0, there exits an open cover O =
nN

we have
l(O) = l(O E) + l(O E c ) = l(E) + l(O\E).
So, l(O) l(E) < a . Thus, for O, we have
X
X
l(O) =
l(In )
a l(E In ) = a l(E O) = a l(E).
nN

nN

Thus, we have that l(O) l(E) (a 1)l(E), which implies that l(E) (a 1)l(E) l(O) l(E) <
Finally, we conclude that
l(O) = l(E) + l(O\E) (a 1)l(E) + l(O\E) <

a.

(a 1)
+ = .
a
a

Therefore, since l(O) < for all > 0, we conclude that l(E) = 0.

Problem 4. Suppose and are two finite (nonnegative) measures on a measurable space (X, A). Show
there is a nonnegative measurable function f on X such that for all E A,
Z
Z
(1 f ) d =
f d.
E

Proof. Observe that + . So, there exists a nonnegative function f L1 (X, A, + ) such that for
each E A,
Z
Z
Z
(E) =
f d( + ) =
f d +
f d.
E

Thus, we have
Z
(E) =

Z
f d +

Z
(E)
ZE

Z
E

Z
f d =

f d
ZE

1 d
f d =
f d
E
Z
ZE
(1 f ) d =
f d
E

19

f d
E

MATH 7311 Test Bank Questions

Z
Problem 5. Define f (x) =

Fall 2016

cos(xy)g(y) dy for x R, where g L1 (R). Show that f is continuous on R.

Proof. Exercise III.14.

Problem 6. Let gn = n[0,n3 ] .


(a) Show that
1

f (x)gn (x) dx = 0,

lim

for all f L2 [0, 1].


(b) Find a function f L1 [0, 1] for which
Z
n

f (x)gn (x) dx 6= 0

lim

Proof.
(a) For each n N, we have that
1

n3

Z
f (x)n[0,n3 ] dx =

nf (x) dx.
0

So,
Z
lim

n3

Z
nf (x) dx =

nf (x) dx.
0

Z
Since an integral over a singleton set is zero for any function, we have that

f (x)gn (x) dx = 0.
0

(b)

20

MATH 7311 Test Bank Questions

Fall 2016

Problem 7. Let (X, A, ) be a measure space. Suppose fn is a sequence of nonnegative integrable functions
on X that converges a.e. to an integrable function f , and
Z
Z
lim
fn d =
f d.
n

Let B A. Show

lim

fn d =

f d.

(Hint: Use Fatous Lemma.)


Proof. Since fn f a.e., we have a sequence of nonnegative measurable functions fn 1B such that fn 1B
f 1B a.e. By Fatous Lemma,
Z
Z
f d lim inf
fn d.
B

Now, since f 1B is integrable since f 1B f . Also, fn is integrable for sufficiently large n, which implies that
fn 1B is as well. By Fatous Lemma, we have that
Z
Z
(f f 1B ) d lim inf
(fn fn 1B ) d.
X

That is,
Z

Z
f d

Z
fn lim sup

Z
fn

Thus, lim inf

Z
f d lim inf

Z
f , and we have lim

Z
f lim sup

fn =
X

fn .
B

Z
fn d =

f d.
B

Problem 8. Let (X, A, ) be a finite measure space. Suppose fn Lp (X) for n = 1, 2, . . . , fn f a.e.,
f Lp (x), and kfn kp kf kp . Show that fn f in Lp (x). (Hint: Use Egoroffs Theorem.)
Proof.

21

MATH 7311 Test Bank Questions

Fall 2016

Fall 2013 Exam


Problem 1. Let (X, A) be a measurable space and let : A [0, ] be a finitely additive measure.
(a) Show that if (X) < , then is countably additive iff for every decreasing sequence {Aj }jN in A,
we have
!

\
Aj = lim (Aj ).

j=1

(b) Give an example of a measure space (X, A, ), ( additive) with (X) = , and a sequence of
decreasing measurable sets {Aj }jN in A such that (a) fails.
Proof.
(a) ()
Assume that is countable additive, and let {Aj }jN in A be a decreasing sequence, that is, Aj Aj+1
for all j N. We need to create a disjoint sequence from {Aj }. So, let Bj = A1 \Aj . Thus,
!
!
!

[
[
\

Bj =
A1 \Aj = A1 \
Aj
j=1

j=1

j=1

= (A1 )

!
Aj

j=1

= (A1 ) lim
n

n
\

!
Aj

j=1

= (A1 ) lim (An )


n

Therefore, eliminating (A1 ) from both sides of the equation, we get

!
Aj

= lim (An ).
n

j=1

()
Assume that for every decreasing sequence {Aj }jN in A, we have
!

Aj = lim (Aj ).
j

j=1

Since

Aj A, we have

j=1

(A1 ) = A1

j=1

\
c !

+ A1
Aj
=

j=1

j=1

!
Aj

22

!
Aj +

[
j=1

!
A1 Acj

\
j=1

!
Aj +

[
j=1

!
A1 \Aj .

MATH 7311 Test Bank Questions

Fall 2016

Thus,

!
Aj

= (A1 )

j=1

[
j=1

!
A1 \Aj

= lim (An )
n

(b)

Problem 2. Let (X, A, ) be a measure space such that (X) > 0. Let f : X R be measurable and
suppose that f is finite -a.e. Show that there exists Y A such that (Y ) > 0 and f is bounded on Y .
Proof.

23

MATH 7311 Test Bank Questions

Fall 2016

Problem 3. Let (X, A, ) be a finite measure space. Let {fn }nN be a sequence of measurable functions
fn : X R that converges almost everywhere to a measurable function f : X R. Show that the sequence
{fn }nN converges to f in measure.
Proof. Since fn converges to f pointwise almost everywhere in X, we can apply Egoroffs Theorem. So, for
each > 0, there exists a set B A such that (B) < and such that fn f uniformly on X\B. In other
words, for each > 0, there exists N N such that |fn (x) f (x)| < for all n N and all x X\B. This
means that for all x B, |fn (x) f (x)| for all n N . So, we have that
B = {x | |fn (x) f (x)| }.
Since (B) < , we can conclude that fn converges in measure to f .

Problem 4. Determine if the following limits exist. If the limit exists, find the limit and justify your answer:
Z n
x n
dx.
(a) lim
1
n 0
n
Z +
sin(x/n)
(b) lim
dx.
n R
1 + x2
Z
(c) lim
f (x)gn (x) d(x) where f : R R is a continuous
n

Proof.

24

MATH 7311 Test Bank Questions

Fall 2016

Problem 5. For n = 1, 2, . . . define



fn (x) =

1
n

if |x| n,
if |x| > n.

Find

Z
lim fn = f and lim

fn (x) d(x).
R

Explain why your result does not contradict the Lebesgue dominated convergence theorem.
Proof. Exercise III.9

Problem 6. Let (X, A, ) be a measure space. Assume that fn L1 (X, ) for n N and that
Z
X
n=1

|fn (x)| d(x) < .

Show that
(a) The series

fn (x) converges a.e. to a measurable function f on X.

n=1

(b) f L1 (X, ) and


Z
f (x) d(x) =
X

Z
X
n=1

Proof.

25

fn (x) d(x).

MATH 7311 Test Bank Questions

Fall 2016

Problem 7. Let f, g L1 (R, 1 ).


(a) Show that H(x, y) = f (x)g(y x) is integrable on R2 with respect to the Lebesgue measure 2 .
Z
(b) Show that x 7 H(x, y) is integrable with respect to 1 , that G(y) =
H(x, y) d1 (x) is integrable
R

on R with respect to 1 , and that


kGk1 kf k1 kgk1 .
Proof. Problem 9 Fall 2009.

Problem 8. For < a < b < let I = [a, b] and let be a signed measure on I (with respect to the
Borel -algebra). Let f (x) = ([a, x]). Then f is of bounded variation.
Proof. Since is a signed measure, there exists a real number Mx for each x [a, b] such that
|([a, x])| = |f (x)| Mx .
Let 4 be an arbitrary partition
4 = {a = t0 tn = x}
of [a, x] by finitely many points. Then, we have
v(4) =

n
X

|f (ti ) f (ti1 )|

i=1

n
X

Mti + Mti1 R.

i=1

Thus, f is of bounded variation.

Problem 9. Let Q = [0, 1]n = {(x1 , . . . , xn ) Rn | (j = 1, . . . , n) 0 xj 1}. Let f : Q R be a


continuous function. Show that the Lebesgue measure of the set {(x, f (x)) | x Q} Rn+1 is zero.
Proof.

26

MATH 7311 Test Bank Questions

Fall 2016

Spring 2011 Exam


Problem 1. Let (X, A, ) be a measure space and f : X R a measurable function (in the sense that
f 1 (, a) A for all a R.) Prove f 1 (B) is measurable for every Borel set B R.
Proof.

Problem 2. Let (X, A, ) be a measure space and suppose fn : X R is a measurable function for each
n N. Prove that the set
n
o
x | lim fn (x) exists
n

is a measurable set.
Proof. Exercise II.24.

Problem 3. Suppose A and B are Lebesgue measurable subsets of R, each one of strictly positive but finite
measure. Prove that there exists a number c R such that
l((A + x) B) > 0,
where l denotes Lebesgue measure. (Hint: Consider the outer measure of A and B. Or, alternatively,
consider the convolution 1A 1B and use Fubinis Theorem.)
Proof. Exercise II.22.

27

MATH 7311 Test Bank Questions

Fall 2016

Problem 4. Prove that

(sin nt)f (t) dt = 0

lim

for all Lebesgue integrable functions f on R. (Hint: The compactly supported step functions are dense in
L1 (R).
Proof.


Problem 5. Let f (x) =

sin x
x

if x > 0
Prove that lim
if x = 0.
b

Proof.

28

Z
0

f (x) dx exists but that f 6 L1 [0, ).

MATH 7311 Test Bank Questions

Fall 2016

Z
cos(xy)g(y) dy for x R and g and integrable function on R. Show that f

Problem 6. Define f (x) =


R

is continuous.
Proof. Exercise III.14.

Problem 7. Show that if f is absolutely continuous on [a, b], then the total variation Vax f , where x [a, b]
is given by
Z x
Vax f =
|f 0 (t)| dt.
a
0

(Hint: LetZ F (x) =


show that F (x) |f 0 (x)| a.e. by using Vxx+h |f (x + h) f (x)| and then
ZFirst
x
x
|f 0 (t)| dt
F 0 (t) dt Vax f . Then show the reverse inequality.)
conclude
Vax f.

Proof.

Problem 8. Let X be the set of Lebesgue measurable functions on [0, 1]. For f, g X, define
Z
(f, g) =
0

|f (x) g(x)|
dx.
1 + |f (x) g(x)|

Let fn X, n N. Show that lim (fn , f ) = 0 iff fn f in measure.


n

Proof.

29

MATH 7311 Test Bank Questions

Fall 2016

Fall 2009 Exam


Problem 1. Let (X, A, ) be a measure space for which (X) < .
(a) Suppose fn is a sequence of measurable functions such that fn f everywhere. Prove that fn f in
measure.
(b) Give an example on [0, 1] of a sequence of Lebesgue measurable functions fn converging in measure,
yet fn (x) diverges for all x. Justify that your answer is correct.
Proof.
(a) Fall 2013 Problem 3.
(b) We cut the interval [0, 1] into n pieces of length for each n N. We then create a sequence of intervals
{In }
n=1 as follows:
[0, 1],
[0, 1/2], [1/2, 1]
[0, 1/3], [1/3, 2/3], [2/3, 1]
[0, 1/4], [1/4, 2/4], [2/4, 3/4], [3/4, 1]
..
.
Define fn = 1In restricted to [0, 1]. Let f . Then {fn } 6 f pointwise since for each x [0, 1],
fn (x) = 1 for an infinite number of n N. Notice that fn =
6 0 on an interval of length 1/m for some
m n. So, as n , m and
lim m({x E | |fn (x) f (x)| > }) lim

So, {fn } f in measure.

Problem 2. Exercise II.9.

30

1
= 0.
m

MATH 7311 Test Bank Questions

Fall 2016

Problem 3. Let f 0 be real-valued Lebesgue measurable function on Rd . Define (E) =

Z
f dl.
E

(a) Show that the measure l.


(b) Show that is -finite. Note that this means that there is a sequence En of measurable sets with

S
E=
En and (En ) < for all n.
n=1

Proof.
(a) Suppose that l(E) = 0. Thus, E is a null set, and the Lebesgue integral of any function over a null set
is zero. Thus, (E) = 0. This implies that l.
(b)

Problem 4. Suppose A and B are measurable subsets of R, each one of strictly positive but finite measure.
Prove that there exists a number c R such that l((A + c) B) > 0.
Proof. Exercise II.22.

31

MATH 7311 Test Bank Questions

Fall 2016

Problem 5. Let f L1 (X, A, ), a -finite measure space with X =

Xn such that (Xn ) < for each

nN

n. Let An A be such that (An ) 0.


Z
(a) Prove: lim
f d = 0.
n

An

Z
(b) Prove or disprove: lim

f d = 0.
S

Xn

nN

Proof.
(a) Since f L1 (X, A, ), we have f = f + f for f + , f , L+ . Since X is -finite, we have
Z
Z
Z
lim
f d = lim
f + d lim
f d
n A
n A
n A
n
n
n
Z
Z
h d
= lim sup
g d lim sup
n 0gf
gS1

n 0hf
hS1

An

An

Z
lim

Mg d lim

sup

n 0gf
gS1

= lim

Z
sup

n 0gf
gS1

An

sup Mg (An ) lim

sup Mh (An )

n 0hf
hS1

n 0gf
gS1

= ( sup Mg sup Mh ) lim (An )


0gf
gS1

0hf
hS1

= 0.
(b)

32

Mh d
An

(Mg , Mh R+ )

MATH 7311 Test Bank Questions

Fall 2016

Problem 6. Prove that if fn is Lebesgue integrable on [0, 1] for each n N and if


XZ
nN

then

|fn (x)| dx < ,

fn (x) is convergent almost everywhere to an integrable function f , and

nN

f (x) dx =
0

XZ

fn (x) dx.

nN

Proof. Define the function g : N [0, 1] with g(n, x) = fn (x) on the measure space (N X, A B, )
where is the counting measure. Since fn is Lebesgue integrable, its measurable. Also, all functions defined
on N are measurable with respect to the counting measure. So, g is measurable on NZ X. Thus, |g| is
X
g d =
g. So,
measurable as well. Note that in N with respect to the counting measure, we have that
N

nN

by Tonellis Theorem,
1

XZ

|fn (x)| dx =

nN

XZ

Z
|g| d =

nN

|g| =

0 nN

|fn (x)| < .

0 nN

So, by Fubinis Theorem, we know that |g| = |fn (x)| is integrable on N. So,

|fn (x)| < . So,

nN

fn is

nN

absolutely convergent and thus convergent a.e. to a function f . Since fn is integrable for all n N, so is f .
n
P
Define sn :=
fi . So, sn is a monotone increasing sequence of integrable functions that converge to f a.e.
i=1

So, by the Monotone Convergence Theorem, we have


Z

Z
f (x) dx =

Z
lim sn = lim

0 n

Z
sn = lim

n
1X
i=1

fi = lim

n Z
X
i=1

fi =

X Z
n

fn dx.

Z
cos(xy)g(y) dy for x R, where g is an integrable function on R. Show that

Problem 7. Define f (x) =


f is continuous.

Proof. Exercise III.14.

33

MATH 7311 Test Bank Questions

Fall 2016

Problem 8. Suppose that f and g are in L1 (Rn ) and let h(x, y) = f (x y)g(y). Show that h L1 (R2n ).
Use this to show that the function
Z
f g(x) =
f (x y)g(y) dy
Rn

is defined for almost all x and that f g L1 (Rn ). Then show that
kf gk1 kf k1 kgk1 .
(Hints: (i) You may assume in this problem that the function h is Lebesgue measurable on R2n . (ii) You may
use Fubinis Theoem and/or Tonellis Theorem, being very careful to show that all hypotheses are satisfied.)
Proof. Since h is measurable on R2n , Tonellis Theorem tells us that

Z Z
|f (x y)g(y)| dl(x) dl(y)
Rn

Rn

exists. So, since f, g L1 (Rn ), we have that


!
Z
Z
Z
|f (x y)g(y)| dl(x) dl(y) =
Rn

Rn

Z
|g(y)|

f (x y) dl(x)

Rn

dl(y)

Rn

Z
|g(y)|

f (x) dl(x)

Rn

Rn

! Z
|g(y) dl(y)

=
Rn

dl(y)
!

|f (x)| dl(x)

Rn

<
Therefore, h is integrable on R2n and thus h L1 (R2n ). We also have that h is integrable
on Rn Rn . So,
Z
Fubinis Theorem tells us that h(x, ) = f (x y)g(y) is integrable a.e. on Rn . Thus,
f (x y)g(y) dl(y)
Rn

is defined almost everywhere on Rn . So, since h is integrable on Rn Rn , so is |h|. Note that |h|(, y) =
|h(, y)| = |f (x)g(y)| and |h|(x, ) = |h(x, )| = |f (x y)g(y)|. So, by Fubinis Theorem, we have

Z Z



kf gk1 =
f (x y)g(y) dl(y) dl(x)


Rn Rn
!
Z
Z

|f (x y)g(y)| dl(y) dl(x)


Rn

Rn

Z
|f (x)g(y)| dl(x)

=
Rn

dl(y)

Rn

Z
=

! Z
f (x) dl(x)

Rn

Rn

= kf k1 kgk1

34

!
g(y) dl(y)

MATH 7311 Test Bank Questions

Fall 2016

Spring 2010 Exam


Problem 1. Determine the validity of each of the following statements; simply write True or False. You do
not have to justify your answers. Assume measure and integration on the real line in the sense of Lebesgue.
(a) Suppose f is a measurable function and E is a measurable set. Then f 1 (E) is a measurable set.
(b) Suppose fn converges to f a.e. on a set of finite measure. Then fn converges to f in measure.
(c) A bounded measurable function defined on a set of finite measure is integrable.
(d) If f L [0, 1] and > 0, then there is a continuous function g such that
kf gk < .
(e) If f is of bounded variation and f 0 = 0 a.e. then f is constant a.e.
Proof. (a) False; (b) False; (c) True; (d) ; (e) False

Problem 2.
(a) State Fatous Lemma for a complete measure space (X, A, ).
(b) Give an example of a sequence fn of nonnegative functions on R for which
Z
Z
f < lim inf
fn ,
R

where f is the pointwise limit of fn . (note the strict inequality)


Proof.
(a) Suppose (X, A, ) is a complete measure space and suppose that fn : X R is measurable and
nonnegative for each n N. Suppose that lim inf n fn (x) = f (x) a.e. Then,
Z
Z
f d lim inf
fn d.
n

(b) Let A be a nonempty, measurable


set with finite positive measure, and consider the sequence of

1A
if n is odd
functions {fn } such that fn =
It follows that lim inf fn = 0. Also,
1

1
if n is even.
A
Z
Z
fn d = (AC ) if n is even. Thus,

fn d = (A) if n is odd, and


X

Z
lim inf

fn d = min{(A), (Ac )}.

Therefore, since the measure of A is nonzero, finite, and positive, so is the measure of AC . Hence,
Z
Z
f d = 0 < lim inf
fn d = min{(A), (Ac )}.
X

35

MATH 7311 Test Bank Questions

Fall 2016

Problem 3.
(a) State the Lebesgue Dominated Convergence Theorem for a complete measure space (X, A, ).
(b) Give an example of a sequence fn of nonnegative integrable functions on [0, 1] that converges pointwise
to an integrable function f and for which
Z 1
Z 1
fn 6=
f.
lim
n

Proof.
(a) Let (X, A, ) be a complete measure space. Let fn : X R be a sequence of measurable functions
converging pointwise a.e. to a function f . Suppose there exists an integrable function L such that
|fn (x)| (x) for all x X and for all n N. Then
1. f is integrable and
Z
Z
fn d.
2.
f d = lim
X

(b)

36

MATH 7311 Test Bank Questions

Fall 2016

Problem 4.
(a) State the definition of convergence in measure on a measure space (X, A, ).
(b) Give an example of a sequence {fn } of bounded measurable functions on [0, 1] which converge to a
function f in measure but does not converge pointwise at any t [0, 1].
Proof.
(a) On a measure space (X, A, ), a sequence of measurable functions fn is said to converge in measure to
a measurable function f provided that for each > 0, there exits N N such that n N implies that
{x | |fn (x) f (x)| } < .
(b) We cut the interval [0, 1] into n pieces of length for each n N. We then create a sequence of intervals
{In }
n=1 as follows:
[0, 1],
[0, 1/2], [1/2, 1]
[0, 1/3], [1/3, 2/3], [2/3, 1]
[0, 1/4], [1/4, 2/4], [2/4, 3/4], [3/4, 1]
..
.
Define fn = 1In restricted to [0, 1]. Let f . Then {fn } 6 f pointwise since for each x [0, 1],
fn (x) = 1 for an infinite number of n N. Notice that fn =
6 0 on an interval of length 1/m for some
m n. So, as n , m and
1
= 0.
m m

lim m({x E | |fn (x) f (x)| > }) lim

So, {fn } f in measure.

Problem 5. If f is absolutely continuous on an interval (a, b), E (a, b), and m(E) = 0, then
m(f (E)) = 0.
Here m denotes Lebesgue measure and f (E) = {f (x) : x E}.
Proof.

37

MATH 7311 Test Bank Questions

Fall 2016

Problem 6. Suppose f is in L4 [0, 1]. Show that


Z

f (x)
1

x4

dx

is finite.
Proof.

Problem 7. Determine

Z
lim

sin
1

 
n3
x
dx.
n 1 + n2 x3

Justify your answer.


Proof.

38

MATH 7311 Test Bank Questions

Fall 2016

Problem 8. If E is a measurable subset of [0, 1] prove that there is a measurable subset A E such that
m(A) = 12 m(E). Here m denotes Lebesgue measure.
Proof. Let E be a Lebesgue measurable subset of [0, 1] and let m be Lebesgue measure. Let
f (x) = m([0, x) E) such that f is increasing, f (0) = 0, and f (1) = m(E). Claim that f is continuous. So,
for all > 0, there exists > 0 such that for x, y [0, 1], |x y| < and |f (x) f (y)| < . Let x, y [0, 1]
so that f (y) y and f (x) x. So, since
f (x) = m([0, x) E) x
and
f (y) = m([0, y) E) y,
we have
f (x) f (y) = m([x, y] E) y x.
Take = . Then |f (x) f (y)| |x y| < . Therefore, f is continuous. By the Intermediate Value
T
1
1
Theorem, there exists a c [0, 1] such that f (c) = m(E). Then A = [0, c] E. Thus m(A) = m(E).
2
2

Problem 9. Suppose (X, A, ) is a measure space and m(X) = 1. Suppose f and g are positive measurable
functions on X such that f g 1. Show that
Z
Z
f d(x)
g d(x) 1.
X

Proof.

39

You might also like