0% found this document useful (0 votes)
89 views6 pages

Real Analysis HW 9 Solutions: N N N N N

This document contains solutions to problems from Real Analysis homework 9. The first problem shows that if a sequence of functions converges pointwise to a function f, then the total variation of f is less than or equal to the lim inf of the total variations of the functions in the sequence. The second problem examines the bounded variation and absolute continuity of a function defined using powers and trigonometric functions, showing it has bounded variation if the powers satisfy one inequality and does not if they satisfy another. The third problem examines properties of absolute continuity for various functions, including one defined using a square root that is absolutely continuous but not Lipschitz.
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)
89 views6 pages

Real Analysis HW 9 Solutions: N N N N N

This document contains solutions to problems from Real Analysis homework 9. The first problem shows that if a sequence of functions converges pointwise to a function f, then the total variation of f is less than or equal to the lim inf of the total variations of the functions in the sequence. The second problem examines the bounded variation and absolute continuity of a function defined using powers and trigonometric functions, showing it has bounded variation if the powers satisfy one inequality and does not if they satisfy another. The third problem examines properties of absolute continuity for various functions, including one defined using a square root that is absolutely continuous but not Lipschitz.
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/ 6

Real Analysis HW 9 Solutions

Problem 33: Let {fn } be a sequence of functions on [a, b] converging pointwise to f . Then
T V (f ) ≤ lim inf n T V (fn ).
Solution: Let P be any partition of [a, b], then since V (fn , P ) only depends on fn at a finite
number of points,

V (f, P ) = lim V (fn , P ).


n

Furthermore V (fn , P ) ≤ T V (fn ) and therefore

V (f, P ) = lim V (fn , P ) ≤ lim inf T V (fn ).


n n

Taking the sup over all partitions P gives

T V (f ) ≤ lim inf T V (fn ).


n

Problem 35: For α and β positive numbers, define the function f on [0, 1] by
(
xα sin (1/xβ ) for 0 < x ≤ 1
f (x) =
0 for x = 0.

Show that if α > β, then f is of bounded variation on [0, 1], by showing that f 0 is integrable
over [0, 1]. Then show that if α ≤ β, then f is not of bounded variation on [0, 1].
Solution: If α > β, then

f 0 (x) = αxα−1 sin (1/xβ ) − βxα−β−1 cos (1/xβ ).

Since f is C 1 and bounded on (0, 1) we can use the fundamental theorem of calculus for
Riemann integrals to conclude that for any partition P
n
X Z 1
V (f, P ) = |f (xk ) − f (xk−1 )| ≤ |f 0 |
k=0 0

and therefore Z 1 Z 1
0
T V (f ) ≤ |f | ≤ αxα−1 + βxα−β−1 < ∞,
0 0

1
since α > 0 and α − β > 0.
If α ≤ β, choose a partition Pn = {0, an , an−1 , . . . , a1 }, where
 nπ −1/β
an =
2
then we see that
n  −α/β
X kπ
V (f, Pn ) = .
k=1
2
Note that this series diverges as n → ∞ since α/β ≤ 1. Therefore

lim V (f, Pn ) ≤ T V (f ) = ∞.
n→∞


Problem 37: Let f be a continuous function on [0, 1] that is absolutely continuous on [, 1]
for each 0 <  < 1.
(i) Show that f may not be absolutely continuous on [0, 1].

(ii) Show that f is absolutely continuous on [0, 1] if it is increasing.



(iii) Show that the function f on [0, 1], defined by f (x) = x for 0 ≤ x ≤ 1, is absolutely
continuous, but not Lipschitz, on [0, 1].
Solution:
(i) Consider (
x sin (1/x) if 0 < x ≤ 1
f (x) =
0 if x = 0
Note that on [, 1]

0
1 1
|f (x)| = sin(1/x) − cos (1/x) ≤ 1 +

x 

Therefore f is Lipschitz on [, 1] and hence absolutely continuous on [, 1]. However,
we know from Problem 35 that f is not BV on [0, 1] and therefore not abolsolutely
continuous on [0, 1].

(ii) Suppose f is increasing, let η > 0 and choose  so that

f () − f (0) < η/2.

Since f is absolutely continuous on [, 1] choose δ > 0 in response to η/2 in the absolute
continuity condition on [,P1]. Now suppose that {(ak , bk )}Nk=1 is a collection of disjoint
N
open intervals such that k=1 |bk − ak | < δ. We may assume that  is not contained

2
in any of the intervals since we may always split any such interval (ak0 , bk0 ) into two
consecutive intervals (ak0 , ) ∪ (, bk0 ) such that, by the fact that f is increasing,

|f (bk0 ) − f ()| + |f () − f (ak0 )| = |f (bk0 ) − f (ak0 )|.

It follows that we may divide the set of intervals into n− intervals to the left of ,
− n− + + n+
{(a−k , bk )}k=1 and n+ intervals to the right of , {(ak , bk )}k=1 . We observe that

n+
X

|b+
k − ak | < δ
k=1

and so by uniform integrability on [, 1],


n+
X
|f (b+ +
k ) − f (ak )| < η/2.
k=1

Also, since f is increasing,


n−
X
|f (b− −
k ) − f (ak )| ≤ f () − f (0) < η/2.
k=1

Therefore n
X
|f (bk ) − f (ak )| < η.
k=1

(iii) Clearly √ x is not Lipschitz on [0, 1] since its derivative is unbounded as x → 0.
However, x is increasing and is Lipschitz on [, 1] for any  > 0. Therefore by (ii), f
is absolutely continuous on [0, 1].


Problem 39: Use the preceding problem to show that if f is continuous and increasing on
[a, b], then f is absolutely continuous on [a, b] if and only if for each , there is a δ > 0 such
that for a measurable subset E of [a, b],

m∗ (f (E)) <  if m(E) < δ.

Solution: Suppose that f is absolutely continuous, and let δ > 0 be chosen in response
to  > 0 in the absolute continuity condition as generalized in Problem 38. Supppose that
E ⊆ [a, b] is measurable and m(E) < δ/2. We can find a countable cover of disjoint open
intervals {(ak , bk )}∞
k=1 of E so that


!
[
m Ik < δ/2 + m(E) < δ
k=1

3
and therefore ∞
X
|f (bk ) − f (ak )| < .
k=1

Since f is increasing and continuous {f (Ik )}∞ k=1 is an open cover of f (E) and m(f (Ik )) =
f (bk ) − f (ak ). By the definition of outer measure we conclude that

X

m (f (E)) ≤ |f (bk ) − f (ak )| < .
k=1

For the converse, let δ > 0 be chosen to satisfy the converse condition with S  > 0 and let
{(ak , bk )}nk=1 be a finite collection of disjoint open interval such that E = nk=1 (ak , bk ) and
m(E) < δ. Since f is increasing we see that {f ((ak , bk ))}nk=1 are disjoint and by continuity
m(f ((ak , bk ))) = f (bk ) − f (ak ). By the convese condition
n n
!
X [
|f (bk ) − f (ak )| = m (ak , bk ) < .
k=1 k=1


Problem 43: Define the functions f and g on [−1, 1] by f (x) = x1/3 for −1 ≤ x ≤ 1 and
(
x2 cos (π/2x) 6 0, x ∈ [−1, 1]
if x =
g(x) =
0 if x = 0.

(i) Show that both f and g are absolutely continuous on [−1, 1]

(ii) For the partition Pn = {−1, 0, 1/2n, 1/[2n − 1], . . . , 1/3, 1/2, 1} of [−1, 1], examine
V (f ◦ g, Pn ).

(iii) Show that f ◦ g fails to be of bounded variation, and hence also fails to be absolutely
continuous, on [−1, 1].

Solution:
(i) To show the absolute continuity of f , we note that it is monotone on [−1, 1], and so it
will be absolutely continuous if
Z 1
f 0 = f (1) − f (−1).
−1

Since f 0 (x) = 13 x−2/3 is Riemann integrable on [−1, −] ∪ [, 1] for any  > 0, and
f (1) = 1, and f (−1) = −1, then by monotone convergence

1 1 2/3
Z
x dx = lim (1)1/3 + 1/3 − (−1/3 ) − (−1)1/3 = 2,
 
3 −1 →0

4
and therefore f must be absolutely continuous.
To show that g is absolutely continuous, we simply show that it is the indefinite integral
of some function over [a, b]. Define g(x) = 2x cos (π/2x) + 21 sin (π/2x), then we claim
that Z x
g(s)ds = x2 cos (π/2x) = f (x)
−1

Once again using the fact that g(x) is integrable on [−1, −] ∪ [, 1] for any  > 0, we
can say that if x < 0, then we already have
Z x
g(s)ds = x2 cos (π/2x) = f (x).
−1

However if x ≥ 0, then
Z x
g(s)ds = lim x2 cos (π/2x) − 2 cos (−π/2) + 2 cos (π/2) − 0
 
−1 →0

= x2 cos (π/2x) = f (x).

Therefore f (x) is absolutely continuous.


(ii) Let Pn = {−1, 0, 1/2n, 1/[2n−1], . . . , 1/3, 1/2, 1} be a partition of [−1, 1]. We examine
V (f ◦ g, Pn ),

V (f ◦ g, Pn ) = V (x2/3 cos1/3 (π/2x), Pn )


2n−1
X cos1/3 (nπ − kπ/2) cos1/3 (nπ − (k − 1)π/2)
=
2/3
− 2/3

k=1
(2n − k) (2n − k + 1)

Using the fact that


(
0 k odd
cos (nπ − kπ/2) = cos (nπ) cos (kπ/2) =
(−1)n+k/2 k even

We can consider the sum over only even indicies in k, each of which is counted twice
in the sum
n−1
(−1)(n+k)/3

X
V (f ◦ g, Pn ) = 2
(2(n − k))2/3
k=0
n−1
X 1
=2
k=0
(2(n − k))2/3
n
1/3
X 1
=2 2/3
k=1
k

5
(iii) The sum in the previous part for diverges as n → ∞ Therefore, f ◦ g has unbounded
variation, and cannot be absolutely continuous.


Problem 45: Let f be absolutely continuous on R and g be absolutely continuous and
strictly monotone on [a, b]. Show that the composition f ◦ g is absolutely continuous on
[a, b].
Solution: Let  > 0, {(ak , bk )} be a collection of disjoint intervals of [a, b], Since g is strictly
increasing, {(g(ak ), g(bk )} forms another disjoint colleciton of intervals, and by the absolute
continuity of f , there exists a δ1 > 0 such that whenever
n
X n
X
|f (g(bk )) − f (g(ak ))| <  whenever |g(bk ) − g(ak )| < δ1 .
k=1 k=1

However, by the absolute continuity of g, there exists a δ2 > 0 such that


n
X n
X
|g(bk ) − g(ak )| < δ1 whenever [bk − ak ] < δ2 .
k=1 k=1

You might also like