Chapter 11

Sequences and Series of Functions

Where is it proved that one obtains the derivative of an infinite series by

taking derivative of each term?

Niels Henrik Abel (1802–1829)

• We say that a sequence of functions {fn : D → R} defined on a subset D ⊆ R converges

pointwise on D if for each x ∈ D the sequence of numbers {fn (x)} converge. If {fn } converges
pointwise on D, then we define f : D → R with f (x) = lim fn (x) for each x ∈ D. We denote
n→∞
this symbolically by fn → f on D.

• We say that a sequence of functions {fn } defined on a subset D ⊆ R converges uniformly on

D to a function f such that for every ε > 0 there is a number N such that

|fn (x) − f (x)| < ε for all x ∈ D, n ≥ N.

We denote this type of convergence symbolically by fn ⇒ f on D.

∞
!
• If {fn }∞
0 is a sequence of functions defined on D, the series fn is said to converge pointwise
n=0
(respectively, uniformly) on D if and only if the sequence n=0 of partial sums,
{sn }∞ given by
n
!
sn (x) = fk (x),
k=0

converges pointwise (respectively, uniformly) on D.

223

A.G. Aksoy, M.A. Khamsi, A Problem Book in Real Analysis, Problem Books in Mathematics,
DOI 10.1007/978-1-4419-1296-1_11, © Springer Science+Business Media, LLC 2010
224 CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS

• Weierstrass M-Test: Suppose that {fn } is a sequence of functions defined on D and {Mn } is
a sequence of nonnegative numbers such that

|fn (x)| ≤ Mn ∀x ∈ D, ∀n ∈ N.
" "∞
If ∞n=0 Mn converges, then n=0 fn (x) converges uniformly on D.

• Assume that fn ⇒ f uniformly on [a, b] and that each fn is integrable. Then, f is integrable
and # b # b
lim fn = f.
n→∞ a a

As a corollary to this theorem we obtain the following result:

• Termwise Integration: If each functionfn (x) is continuous on a closed interval [a, b] and if the
∞
!
series fn (x) converges uniformly on [a, b], then we have
n=1

∞ #
! b # b!
∞
fn (x)dx = fn (x)dx.
n=1 a a n=1

• Suppose that {fn } converges to f on the interval [a, b]. Suppose also that fn′ exists and is
continuous on [a, b], and the sequence {fn′ } converges uniformly on [a, b]. Then

lim f ′ (x) = f ′ (x)

n→∞ n

for each x ∈ [a, b]. As a corollary to this theorem we obtain the following result:

• Termwise Differentiation: If each function fn (x) has the derivative fn′ (x) at any point x ∈
∞
! !∞
(a, b), if the series fn (x) converges to at least one point k ∈ (a, b), and if fn′ (x)
n=1 n=1
∞
!
converges uniformly on (a, b) to a function g(x), then fn (x) converges uniformly on (a, b)
n=1
and is differentiable at any point x ∈ (a, b), whose derivative is equal to g(x). In other words,
termwise differentiation is possible, i.e.,
∞ ∞
$! %′ !
fn (x) = fn′ (x).
n=1 n=1

• A family of functions {fn } ∈ F mapping a set A ∈ Rn into Rm is equicontinuous at a point

a ∈ A if for every ε > 0, ∃δ > 0 such that

|fn (x) − fn (a)| < ε where |x − a| < δ and fn ∈ F.

The family F is equicontinuous on a set A if it is equicontinuous at every point in A.

• Arzelà–Ascoli Theorem: Let A be a compact subset of Rn and C(A, Rm ) the space of contin-
uous functions from A into Rm . A subset B of C(A, Rm ) is compact if and only if it is closed,
bounded, and equicontinuous.
CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS 225

• Weierstrass Approximation Theorem: If f is a continuous function on a closed interval [a, b],

then there exists a sequence of polynomials that converge uniformly to f (x) on [a, b].

Problem 11.1 Prove that if a sequence {fn } of continuous functions on D converges uniformly
to f on D, then f is continuous on D.

Problem 11.2 Prove that a sequence of functions {fn } defined on D is uniformly convergent
on U ⊂ D to f : U −→ R if and only if lim an = 0, where
n→∞

an := sup{|fn (x) − f (x)| : x ∈ U }, n ∈ N.

nx
Problem 11.3 Consider the sequence {fn } defined by fn (x) = , for x ≥ 0.
1 + nx
a) Find f (x) = lim fn (x).
n→∞

b) Show that for a > 0, {fn } converges uniformly to f on [a, ∞).

c) Show that {fn } does not converge uniformly to f on [0, ∞).

1
Problem 11.4 Consider the sequence {fn } defined by fn (x) = , for x ∈ [0, 1].
1 + xn
a) Find f (x) = lim fn (x).
n→∞

b) Show that for 0 < a < 1, {fn } converges uniformly to f on [0, a].

c) Show that {fn } does not converge uniformly to f on [0, 1].

nx
Problem 11.5 Consider the sequence {fn } defined by fn (x) = , for x ∈ [0, 2].
enx
a) Show that lim fn (x) = 0 for x ∈ (0, 2].
n→∞

b) Show that the convergence is not uniform on [0, 2].

226 CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS

Problem 11.6 Determine whether the sequence {fn } converges uniformly on D.

1
a) fn (x) = D = [0, 1]
1 + (nx − 1)2
b) fn (x) = nxn (1 − x) D = [0, 1]
& '
2x
c) fn (x) = arctan D=R
x2 + n3

Problem 11.7 Prove that if {fn } is a sequence of functions defined on A, then {fn } is uniformly
convergent on D if and only if for ε > 0, there exists N ∈ N such that for any m, n > N

sup |fn (x) − fm (x)| < ε.

x∈D

This is called the uniform Cauchy criterion.

Problem 11.8 Suppose that the sequence {fn } converges uniformly to f on the set D and that
for each n ∈ N, fn is bounded on D. Prove that f is bounded on D.

Problem 11.9 Suppose a sequence of functions {fn } are defined as

x
fn (x) = 2x + x ∈ [0, 1].
n
a) Find the limit function f = lim fn .
n→∞

b) Is f continuous on [0,1]?

c) Does [ lim fn (x)]′ = lim fn′ (x) for x ∈ [0, 1]?

n→∞ n→∞
# 1 # 1
d) Does lim fn (x)dx = lim fn (x)dx?
0 n→∞ n→∞ 0

Problem 11.10 Give examples to illustrate that

a) the pointwise limit of continuous (respectively, differentiable) functions is not necessarily

continuous (respectively, differentiable),

b) the pointwise limit of integrable functions is not necessarily integrable.

CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS 227

Problem 11.11 Give examples to illustrate that

a) there exist differentiable functions fn and f such that fn → f pointwise on [0,1] but
( )′
lim fn′ (x) ̸= lim fn (x) when x = 1,
n→∞ n→∞

b) there exist continuous functions fn and f such that fn → f pointwise on [0,1] but
# 1 # 1( )
lim fn (x)dx ̸= lim fn (x) dx.
n→∞ 0 0 n→∞

Problem 11.12 Suppose that {fn } is a sequence of functions defined on D and {Mn } is a
sequence of nonnegative numbers such that

|fn (x)| ≤ Mn ∀x ∈ D, ∀n ∈ N.
∞
! ∞
!
Show that if Mn converges, then fn (x) converges uniformly on D. (This is called the
n=0 n=0
Weierstrass M-test.)

Problem 11.13 Discuss the uniform convergence of the following series:

∞
! xn
a) on R
n!
n=0

!∞
sin(nx)
b) √ on [0, 2π]
n=1
n
∞
! cos2 (nx)
c) on R
n2
n=1

∞
!
Problem 11.14 Let an xn be a power series with radius of convergence r, where 0 < r ≤ +∞.
n=0
If 0 < t < r, prove that the power series converges uniformly on [−t, t].

∞ & n '2
! x
Problem 11.15 Show that the function defined by f (x) = is continuous on R.
n!
n=0
228 CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS

Problem 11.16 Prove Dini’s Theorem: Let A ⊆ R be a closed and bounded (and thus compact)
set and {fn } be a sequence of continuous functions fn : A −→ R such that

a) fn (x) ≥ 0 for any x ∈ A,

b) fn −→ f pointwise and f is continuous,

c) fn+1 (x) ≤ fn (x) for any x ∈ A, and any n ∈ N.

Prove that {fn } converges to f uniformly on A.

Problem 11.17 Give examples to illustrate that all of the hypotheses in Dini’s Theorem
(Problem 11.16) are essential.

Problem 11.18 Let fn : [1, 2] −→ R be defined by

x
fn (x) = .
(1 + x)n
∞
!
a) Show that fn (x) converges for x ∈ [1, 2].
n=1

b) Use Dini’s Theorem to show that the convergence is uniform.

c) Does the following hold:

# * ∞
+ ∞ #
2 ! ! 2
fn (x) dx = fn (x)dx?
1 n=1 n=1 1

Problem 11.19 A sequence of functions {fn } defined on a set A is said to be equicontinuous

on A if for every ε > 0, there exists δ > 0 such that

|fn (x) − fn (y)| < ε whenever |x − y| < δ

for x, y ∈ A and n ∈ N. Prove the following:

a) Any finite set of continuous functions on [0,1] is equicontinuous.

b) If {fn } is a uniformly convergent sequence of continuous functions on [0,1], then {fn } is

equicontinuous.
CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS 229

Problem 11.20 Show that there exists a continuous function defined on R that is nowhere
differentiable by proving the following:

a) Let g(x) = |x| if x ∈ [−1, 1]. Extend g to be periodic. Sketch g and the first few terms of
the sum
∞ & 'n
! 3
f (x) = g(4n x).
4
n=1

b) Use the Weierstrass M-test to show that f is continuous.

c) Prove that f is not differentiable at any point.

Problem 11.21 Let fn be a function such that

fn : (0, 1) −→ R.

a) Prove that if fn ⇒ f and fn′ ⇒ g, then f is continuous on (0,1) and f ′ = g.

b) Describe how you would construct an example to show that uniform convergence of the
derivatives is necessary.

Problem 11.22 Let fn : [0, 1] −→ R be continuous such that {fn } are uniformly bounded
on [0, 1] and the derivatives fn′ exist and are uniformly bounded on (0, 1). Prove that fn has a
uniformly convergent subsequence.

Problem 11.23 Let B be a bounded and equicontinuous subset of the set of continuous real-
valued functions defined on [a, b]. Let T : B −→ R defined as
# b
T (f ) = f (x)dx.
a

Prove that there is a function f0 ∈ B at which the value of T is maximized.

230 CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS

Problem 11.24 Suppose a, b, c, and d are constants chosen from an interval [−K, K] and let
Φ ⊂ (C[0, π], d) be a family of functions f of the form

f (x) = a sin bx + c cos dx where 0 ≤ x ≤ π.

Metric d on C[0, π] is given by d(f, g) = max |f (x) − g(x)|.

0 ≤x≤π

a) Show that Φ is a compact subset of C[0, π].

b) Show that for any continuous function g defined on C[0, π] there exist values a, b, c, and d
in [−K, K] such that
max |g(x) − (a sin bx + c cos dx)|
0 ≤x≤π

is minimum.
For an obvious reason f ∈ Φ is called minimax approximation of g.

Problem 11.25 (Bernstein Polynomials) The nth Bernstein polynomial of a continuous

function f : [0, 1] → R defined by
n
! & '& '
k n
Bn (f )(x) = f xk (1 − x)n−k .
n k
k=0

a) Show that Bn is linear, monotone map, and Bn 1 = 1 and Bn x = x.

1
b) Show that nth Bernstein polynomial for f (x) = ex is Bn (x) = [1 + (e n − 1)x]n .

c) Show that Bn (ex ) converges uniformly to ex on [0, 1].

Problem 11.26

a) Show that for any function f ∈ C[0, 1] and any number ϵ > 0, there exists a polynomial p,
all of whose coefficients are rational numbers, such that

||p − f || < ϵ.

b) Show that C[a, b] is separable.

,b
Problem 11.27 Let f be a continuous function on [a, b] and suppose that a f (x)xn dx = 0 for
n = 0, 1, 2, . . .. Prove that f (x) = 0 on [a, b].
CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS 231

Solutions

Solution 11.1

It follows from the uniform convergence of {fn } that given ε > 0, ∃N ∈ N such that
1
|fn (x) − f (x)| < ε ∀x ∈ D, n ≥ N.
3
Let c ∈ D be fixed. By the continuity of fN at c, there is δ > 0 such that
1
|fN (x) − fN (c)| < ε whenever |x − c| < δ.
3
Thus
|f (x) − f (c)| ≤ |f (x) − fN (x)| + |fN (x) − fN (c)| + |fN (c) − f (c)| < ε.

Solution 11.2

(⇒) Suppose fn ⇒ f on U . Then given ε > 0, there exists N ∈ N such that

|fn (x) − f (x)| < ε ∀n ≥ N ∀x ∈ U.

Hence,
an := sup{|fn (x) − f (x)| : x ∈ U, n ≥ N } < ε
and therefore lim an = 0.
n→∞

(⇐) Suppose lim an = 0. Then for large n and ∀x ∈ U

n→∞

|fn (x) − f (x)| ≤ sup{|fn (x) − f (x)| : x ∈ U } < ε.

That is, fn ⇒ f on U .

Solution 11.3

a) Since fn (0) = 0 for all n ∈ N, we get f (0) = 0. And for x > 0, we have
nx 1
lim = lim = 1,
n→∞ 1 + nx n→∞ 1 + 1/nx
which yields f (x) = 1 for x > 0.
b) If x ≥ a, then - -
- nx
-
-
-= 1 1
- 1 + nx − 1 - 1 + nx ≤ 1 + na ,
and thus we have
1
lim = 0,
n→∞ 1 + na

which implies {fn } converges uniformly to f on [a, ∞).

232 CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS

1
c) Let n ≥ 1. If 0 < x < , then
n
- -
- nx
-
-
- 1 1 1
- 1 + nx − 1- = 1 + nx > 1 + 1 = 2 ,

which implies {fn } does not converge uniformly to f on [0, ∞).

Solution 11.4

a) Since fn (1) = 1/2 for all n ∈ N, we get f (1) = 1/2. And for 0 ≤ x < 1, we have
1
lim = 1,
n→∞ 1 + xn
which yields f (x) = 1 for 0 ≤ x < 1.
b) If x ∈ [0, a], then - -
- 1 - n an
- − 1-= x ≤ .
- 1 + xn - 1 + xn 1 + an
Furthermore, because 0 < a < 1,
an
lim = 0,
n→∞ 1 + an

so {fn } converges uniformly to f on [0, a].

.
n 1 1
c) Given n ∈ N , let x be such that < x < 1, then < xn < 1. Therefore
2 2
- - 1
- 1 - n 1
- − 1 -= x > 2
= ,
- 1 + xn - 1 + xn 1+1 4
which implies {fn } does not converge uniformly to f on [0, 1].

Solution 11.5

a) Since
fn+1 (n + 1)xenx (n + 1) −x
= (n+1)x
= e ,
fn nxe n
fn+1
lim = e−x < 1. Now we use Problem 3.16 to show that lim fn = 0 for x ∈ (0, 2].
n→∞ fn n→∞

b) Let us find the maximum of fn in [0, 2]. Since

nenx − nenx nx
fn′ (x) = 2 = 0 =⇒ enx (n − n2 x) = 0
(e )
nx

=⇒ n2 x = n
1
=⇒ x = ,
n
CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS 233
& '
1 1
it is easy to check that 1/n is the maximum with fn = . Since
n e
& '
1 1
lim sup{|fn (x) − 0| : x ∈ [0, 2]} = lim fn = ,
n→∞ n→∞ n e

the convergence is not uniform on [0, 2].

Solution 11.6

a) Note /
1 0 x ∈ (0, 1],
lim fn (x) = lim = 1
n→∞ n→∞ 1 + (nx − 1)2 2 x = 0.
Notice that all fn ’s are continuous while the limit function is not continuous, which implies
that convergence is not uniform.

b) First notice that

lim fn (x) = lim nxn (1 − x) = 0.
n→∞ n→∞
The previous argument will not work in this case. So let us find the maximum value of fn ,
for n ≥ 1. Let n ≥ 2. We have
( )
fn′ (x) = n2 xn−1 (1 − x) − nxn = nxn−1 n − (n + 1)x .

n
Hence fn′ (x) = 0 iff x = 0 or x = . It is easy to check that the maximum value of fn (x)
n+1
is given by & '
n
sup{|fn (x) − 0| : x ∈ [0, 1]} = fn .
n+1
However,
& ' & 'n+1
n n 1
lim fn = lim = ̸= 0.
n→∞ n+1 n→∞ n+1 e
Thus convergence is not uniform.

c) We have & '

2x
lim fn (x) = lim arctan = 0.
n→∞ n→∞ x + n3
2

In order to find the maximum value of fn , let us compute its derivative

2n3 − 2x2
fn′ (x) = .
(x2 + n3 )2 + 4x2
√
Hence fn′ (x) = 0 iff x = ± n n. It is easy to check that the maximum value of |fn (x)| is given
by & '
√ 1
sup{|fn (x) − 0| : x ∈ R} = fn (n n) = arctan √ .
n n
Since lim sup{|fn (x) − 0| : x ∈ R} = 0, {fn } converges uniformly to 0 on R.
n→∞
234 CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS

Solution 11.7

(⇒) Suppose fn ⇒ f , then given ε > 0, we can find an integer N such that
ε
m ≥ N =⇒ |fm (x) − f (x)| <
2
for all x ∈ D. If m, n ≥ N , then
ε ε
|fm (x) − fn (x)| ≤ |fn (x) − f (x)| + |f (x) − fn (x)| < + =ε
2 2
for any x ∈ D.

(⇐) If given some ε > 0, we can find an N such that for any m, n ≥ N we have

sup |fn (x) − fm (x)| < ε.

x∈D

Hence for any x ∈ D, {fn (x)} is a Cauchy sequence in R which implies fn (x) converges
to f (x). Let us prove that {fn } converges uniformly to f on D. Indeed, let ε > 0. By
the uniform Cauchy criterion, there exists N ∈ N such that for all n, m ≥ N , we have
sup |fn (x) − fm (x)| ≤ ε/2. Fix n ≥ N . Then
x∈D

ε
|fn (x) − f (x)| = lim |fn (x) − fm (x)| ≤
m→∞ 2
for any x ∈ D. Hence
ε
sup |fn (x) − f (x)| ≤ < ε.
x∈D 2
This obviously implies fn ⇒ f on D.

Solution 11.8

Since {fn } converges to f uniformly on D, there exists an N such that n ≥ N implies that
|fn (x) − f (x)| < 1 for all x ∈ D. Fix n0 ≥ N , since fn0 is bounded on D, there exists a constant k
such that |fn0 (x)| < k for all x ∈ D. Thus for all x ∈ D we have that

|f (x)| ≤ |f (x) − fn0 (x)| + |fn0 (x)| ≤ 1 + k.

Thus f is bounded on D.

Solution 11.9

a) f (x) = 2x for x ∈ [0, 1].

b) 2x is continuous for all x ∈ [0, 1].

1
c) Note that lim fn′ (x) = lim 2 + = 2. Therefore [limn→∞ fn (x)]′ = f ′ (x) = 2.
n→∞ n→∞ n
CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS 235
# 1 # 1 # 1
x 1
d) Note that lim fn (x)dx = lim 2x + dx = lim 1 + = 1 and f (x)dx = 1.
n→∞ 0 n→∞ 0 n n→∞ 2n 0
Therefore # 1 # 1
lim fn (x)dx = f (x)dx = 1.
0 n→∞ 0

Solution 11.10

a) Let fn (x) = xn , for x ∈ [0, 1], as in the following figure:

Figure 11.1

Then {fn } converges pointwise to

/
0 0 ≤ x < 1,
f (x) =
1 x = 1.

Each fn is continuous and differentiable on [0, 1], but f is neither continuous nor differentiable
at x = 1.

b) The pointwise limit of integrable functions is not necessarily integrable. Indeed, let
/ p
1 if x = m ∈ Q, m ≤ n, when written in reduced form,
fn (x) =
0 otherwise,

for n ∈ N. Then {fn } converges pointwise to

/
1 x ∈ Q,
f (x) =
0 otherwise.

Since each of the fn has only a finite number of nonzero points, it is integrable on [0,1] with
integral zero. However, f is not integrable on [0,1] as can be seen from Problem 7.2.
236 CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS

Solution 11.11

a) Let fn (x) = xn /n, for n ≥ 1, and f (x) = 0. Then fn → f pointwise on [0,1]. Each fn is
differentiable and fn′ (x) = xn−1 . Thus, for x ∈ [0, 1),

lim f ′ (x) = lim xn−1 = 0.

n→∞ n n→∞

When x = 1, however,
lim f ′ (1) = lim 1n−1 = 1.
n→∞ n n→∞

Thus, for x = 1,
lim f ′ (1) = 1 ̸= 0 = f ′ (1).
n→∞ n

b) For x ∈ [0, 1], set f1 (x) = 1 and for n ≥ 2

⎧ 2
⎨ n x 0 ≤ x ≤ 1/n,
fn (x) = 2n − n2 x 1/n < x ≤ 2/n,
⎩
0 2/n < x ≤ 1,

as depicted in the following figure:

Figure 11.2

Since fn encloses an area of a triangle with base 2/n and altitude n, we have
# 1
fn (x)dx = 1 for all n > 2
0

which can also be computed directly. Therefore,

# 1
lim fn (x)dx = 1.
n→∞ 0
CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS 237

However, fn → f where f (x) = 0 for x ∈ [0, 1]. This is because for x = 0, fn (0) = 0 for every
n, and for any x ∈ (0, 1], we also have fn (x) = 0 for large enough n. Therefore,
# 1( )
lim fn (x) dx = 0.
0 n→∞

Thus we have the result that

# 1 # 1( )
lim fn (x)dx = 1 ̸= 0 = lim fn (x) dx.
n→∞ 0 0 n→∞

Solution 11.12
n
! ∞
!
Let sn (x) = fk (x) be the nth partial sum. Since Mn converges, for all ε > 0, ∃N such that
k=0 n=0
if n ≥ m ≥ N , then
Mm + Mm+1 + · · · + Mn < ε.
Thus if n ≥ m ≥ N , we have

|sn (x) − sm (x)| = |fm+1 (x) + · · · + fn (x)|

≤ |fm+1 (x)| + · · · + |fn (x)|
≤ Mm+1 + · · · + Mn < ε

∞
!
for all x ∈ D. It follows from Problem 11.7 that {sn } converges uniformly on D. Hence fn also
n=0
converges uniformly on D.

Solution 11.13

a) For any n ≥ 1, we have

nn
sup |sn (x) − sn−1 (x)| = sup |fn (x)| ≥ |fn (n)| = ≥ 1.
x∈R x∈R n!

The Cauchy criterion is not satisfied and thus, convergence is not uniform.

b) First note that we do have pointwise convergence. Next notice that

2 3 π4
x ≤ sin(x) , for any x ∈ 0, .
π 4
√
Let n ≥ 10, and h ∈ N such that 2n ≤ n + h < n nπ/4. Then for any k ∈ N with
√
n ≤ k ≤ n + h, we have k/n n < π/4. Hence
& '
2 k 1
√ ≤ sin k √ ,
πn n n n
238 CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS

which implies
n+h
! n+h
! 1 & '
1 2 k 1
√ √ ≤ √ sin k √ .
k πn n k n n
k=n k=n

But √
n+h
! 1 2 k
n+h
!2 k ! 2 √n
n+h
2h 2
√ √ = √ ≥ √ = ≥ ·
k π n n π n n π n n π n π
k=n k=n k=n

This obviously shows that - -

-n+h - 2
-! 1 -
sup - √ sin(kx)- ≥
-
x∈[0 ,2 π] k=n k - π
√
for any n ≥ 10 and h ∈ N such that 2n ≤ n + h < n nπ/4. Therefore the convergence will
not be uniform on [0, 2π].

c) Let Mn = 1/n2 , for n ≥ 1, then - 2 -

- cos (nx) -
- - ≤ Mn ,
- n2 -
!∞
1
since | cos(nx)| ≤ 1. Because is a p-series with p = 2 > 1, it converges. Hence, by the
n2
n=1
Weierstrass M-test convergence is uniform (see Problem 11.12).

Solution 11.14
∞
!
Let x ∈ [−t, t], then |an xn | ≤ |an |tn . Since 0 < t < r, the series |an |tn is convergent. Thus by
n=0
∞
!
the Weierstrass M-test, an x converges uniformly on [−t, t].
n

n=0
Solution 11.15
∞ & n '2
! x
Let t ∈ R. Let us first show that converges uniformly on [−t, t]. Set
n!
n=0
& '2
tn
Mn = , for n ≥ 1.
n!

Since & ' 2 & '2 & '2

Mn+1 tn+1
n! t
lim = lim · n = lim = 0 < 1,
n→∞ Mn n→∞ (n + 1)!
t n→∞ n + 1

!∞ !∞ & n '2
x
the ratio test forces Mn to be convergent and by the Weierstrass M-test, converges
n!
n=0 n=0
uniformly on [−t, t]. Uniform convergence will guarantee that f (x) is continuous because each of
the partial sums is continuous on [−t, t]. Since t was arbitrary, we have continuity of f (x) on R.
CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS 239

Solution 11.16

Let {gn } = {fn − f }. It is sufficient to show that gn ⇒ 0 on A. Given ε > 0, we want to find N
such that |gn (x)| < ε for all n ≥ N and x ∈ A. By hypothesis b), we know that {gn } converges to 0
pointwise on A. That is, for x ∈ A there exists Nx = N (ε, x) such that for n ≥ Nx , 0 ≤ gn (x) ≤ ε/2.
By the continuity of gNx there is a neighborhood U (x, Nx ) such that
ε
|gNx (y) − gNx (x)| < for any y ∈ U (x, Nx ).
2
The neighborhoods U (xi , Nxi ) form a cover for the compact set A. That is, there are finitely many
points x1 , x2 , . . . , xn ∈ A such that

A ⊂ U (x1 , Nx1 ) ∪ U (x2 , Nx2 ) ∪ · · · ∪ U (xn , Nxn ).

Let N = max(Nx1 , Nx2 , . . . , Nxn ). For any x ∈ A, there exists xi such that x ∈ Nxi . Hence for any
n ≥ N we have
ε ε
0 ≤ gn (x) ≤ gNxi (x) ≤ gNxi (x) − gNxi (xi ) + gNxi (xi ) ≤ + = ε.
2 2
Therefore |gn (x)| < ε for n > N , x ∈ A. This completes our proof.

Solution 11.17

(1) Let
1
fn (x) = x ∈ (0, 1), n = 1, 2, . . . .
1 + nx
It is the case that
fn+1 (x) ≤ fn (x).
However, using Problem 11.2, we see
- -
- 1 -
an = sup -- − 0-- = 1,
1 + nx
x∈(0 ,1)

so convergence is not uniform, because A = (0, 1) is not closed, hence not compact.

(2) The assumption that {fn } is a monotone sequence is also necessary. If we consider the function
defined in Problem 11.11 (b), then fn → f pointwise on [0,1]. But the sequence {fn } is not
monotonic and convergence is not uniform.

(3) Continuity of each fn cannot be omitted. For instance, let

/
1 x ∈ (0, n1 ),
fn (x) =
0 x = 0 or x ∈ [ n1 , 1].

Then each fn is not continuous. However, they form a monotonic sequence converging point-
wise to zero on [0,1], but again the convergence is not uniform.

(4) Finally, the continuity of the limit function is also needed. The sequence fn (x) = xn for
x ∈ [0, 1] defined in Problem 11.10 (a) has a discontinuous limit f (x) and fn (x) fails to
converge uniformly on [0,1].
240 CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS

Solution 11.18

a) First observe that

∞
! ∞
!
x 1
= x
(1 + x) n (1 + x)n
n=1 n=1
and when |1/(1 + x)| < 1 or equivalently when |1 + x| > 1 we have that
∞
! x x 1
= · 1 =1
(1 + x)n 1 + x 1 − 1+x
n=1
∞
!
so in particular fn (x) is convergent for x ∈ [1, 2].
n=1

b) A = [1, 2] is compact and fn (x) → 0 pointwise. Clearly, if k ≥ ℓ holds, then fℓ (x) ≥ fk (x).
All the hypotheses of Dini’s theorem are satisfied and thus convergence is uniform.

c) Since the convergence is uniform we can interchange the integral and the summation. Thus
the equality holds.

Solution 11.19

a) Consider {fi : 1 ≤ i ≤ n} where each fi : [0, 1] → R is continuous. Since [0,1] is compact, each
fi is uniformly continuous. Let ε > 0. Since fi is uniformly continuous, there exists δi > 0
such that if |x−y| < δi , then |fi (x)−fi (y)| < ε for 1 ≤ i ≤ n. Now let δ = min{δ1 , δ2 , . . . , δn },
then for any x, y ∈ [0, 1] such that |x − y| < δ, we have

|fi (x) − fi (y)| < ε, for 1 ≤ i ≤ n.

Hence {fi } is equicontinuous.

b) Let ε > 0. Since {fn } is uniformly convergent, there exists N ∈ N such that for any n, m ≥ N
we have
ε
sup |fn (x) − fm (x)| < ·
x∈[0 ,1] 3
In particular, we have for any n ≥ N
ε
sup |fn (x) − fN (x)| < ,
x∈[0 ,1] 3

for any n ≥ N . The first part shows that the family {f1 , f2 , . . . , fN } is equicontinuous. Hence
there exists δ > 0 such that
ε
|fi (x) − fi (y)| <
3
for any i ≤ N and x, y ∈ [0, 1] such that |x − y| < δ. Let n ≥ N , then we have

|fn (x) − fn (y)| ≤ |fn (x) − fN (x)| + |fN (x) − fN (y)| + |fN (y) − fn (y)| < ε
CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS 241

for any x, y ∈ [0, 1] such that |x − y| < δ. Hence for any n ∈ N and any x, y ∈ [0, 1] with
|x − y| < δ, we have
|fn (x) − fn (y)| < ε.
Therefore {fn } is equicontinuous.

Solution 11.20

a) Extend g(x) to all of R by requiring g(x) = g(x + 2) for all x. This is depicted in the following
figure:

Figure 11.3

b) Now, given x, y ∈ R, we have

- -
- -
|g(x) − g(y)| = -|x| − |y|- ≤ |x − y|.

By translation, this also applies to any pair of points that are no more than two apart. On
the other hand, if |x − y| > 2, then

|g(x) − g(y)| ≤ |g(x)| + |g(y)| ≤ 1 + 1 = 2 < |x − y|.

Hence, |g(x) − g(y)| ≤ |x − y| for all x and y, and so it follows that g is continuous on R.
For each integer n ≥ 0, let gn (x) = (3/4)n g(4n x). For instance,

g0 (x) = g(x)
g1 (x) = 34 g(4x)
g2 (x) = 169 g(16x)
..
.

Notice that gn oscillates four times as fast as gn−1 and at 3/4 the height of gn−1 . Now
∞
!
f (x) = gn (x) is defined on R. Moreover, for all x ∈ R, we have that
n=0
& 'n
3
|gn (x)| ≤ ∀n ∈ N.
4
Thus we can apply the Weierstrass M-test to conclude that
∞
!
gn (x) converges uniformly on R. Since each gn is continuous on R, then f (x) is continuous
n=0
on R.
242 CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS

4 −m
c) To see that f is nowhere differentiable, we fix x ∈ R and let hm = ± 2 (note that 4m |hm | =
1
2 ). Next, we claim that - -
- f (x + hm ) − f (x) -
- -
- hm -

are bounded below by a sequence that diverges to +∞ as m → ∞. In order to accomplish

this, we first observe that
∞ & 'n
! 3
f (x + hm ) − f (x) = [g(4n x + 4n hm ) − g(4n x)].
4
n=0

Now we need to examine g(4n x + 4n hm ) − g(4n x) for each n.

Case 1: If n > m, then 4n hm = ± 4n−m /2 is an even integer. Since g(t) = g(t + 2) for all t, it
follows that g(4n x + 4n hm ) − g(4n x) = 0.
Case 2: If n < m, then since |g(r) − g(s)| ≤ |r − s| for all r, s ∈ R, we have that

|g(4m x + 4m hm ) − g(4m x)| ≤ 4m |hm |.

Case 3: If n = m, then since there is no integer between 4m x and 4m (x + hm ), the graph of g

between these points is a straight line of slope ± 1. Thus

|g(4m x + 4m hm ) − g(4m x)| = 4m |hm |.

Combining all of these three cases together, we obtain

- - - -
-!m & 'n n x + 4n h ) − g(4n x) -
- f (x + hm ) − f (x) - - 3 g(4 m -
- - = - -
- hm - - 4 hm -
n=0
! & 3 'n 4n |hm |
m−1
≥ 3 −
m
4 |hm |
n=0
m−1
!
= 3m − 3n
n=0
1 − 3m 1
= 3m − = (3m + 1).
1−3 2

Since (3m + 1)/2 diverges to +∞ as m → ∞, we have shown that f is not differentiable at x.

Since x was arbitrary, f is nowhere differentiable.
CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS 243

Solution 11.21

1. Let x0 ∈ (0, 1). We have

# x
fn (x) = fn (x0 ) + fn′ (u)du
x0

for x ∈ (0, 1). If we take the limit as n → ∞, we will get

# x
f (x) − f (x0 ) = lim (fn (x) − fn (x0 )) = lim fn′ (u)du
n→∞ n→∞ x
0
# x
= g(u)du
x0

because uniform convergence

# x allows for interchanging the limit with the integral. The equa-
tion f (x) − f (x0 ) = g(u)du implies f ′ (x) = g(x), for any x ∈ (0, 1).
x0

2. Consider the sequence {fn } defined by

x
fn (x) = ·
1 + nx2

Then {fn } converges uniformly to 0 on R. Indeed let ε > 0. Choose N ∈ N such that
1/ε2 < N . Then for any n ≥ N and x ∈ R we have

• Case 1: if |x| < ε, then

|x|
|fn (x)| = ≤ |x| < ε.
1 + nx2

• Case 2: if |x| ≥ ε, then

|x| |x| 1
|fn (x)| = 2
≤ 2
≤ < ε,
1 + nx nx εn

since n ≥ N .

Hence
sup |fn (x)| < ε , for any n ≥ N.
x∈R

On the other hand, we have

1 − nx2
fn′ (x) = ·
(1 + nx2 )2
Hence {fn′ } converges pointwise to g where
/
1 if x = 0,
g(x) =
0 if x ̸= 0.

Notice that fn ⇒ f uniformly, but fn′ → g pointwise and f is differentiable but f ′ ̸= g.

244 CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS

Solution 11.22

This is known as the Arzelà–Ascoli Theorem in its simplest form, if we omit the uniform bound-
edness of the derivatives and replace it by the equicontinuity of the functions. Indeed since fn′ are
uniformly bounded on (0,1), there exists M > 0 such that |fn′ (x)| ≤ M , for any x ∈ (0, 1). Using
the Mean Value Theorem, we get
|fn (x) − fn (y)| ≤ M |x − y|, for any x, y ∈ [0, 1], and n ∈ N.
So if ε > 0 is given, set δ = ε/M . Then |fn (x) − fn (y)| < ε, for any n ∈ N, provided |x − y| < δ.
So {fn } is equicontinuous. Let us prove that {fn } has a subsequence which converges uniformly.
Set {rn } = [0, 1] ∩ Q. Since {fn } is uniformly bounded, so {fn (x)} is bounded, for any x ∈ [0, 1].
Therefore the Bolzano–Weierstrass Theorem shows that there exists a subsequence {fn1 } of {fn }
such that {fn1 (r1 )} is convergent. By induction, we can construct a sequence of subsequences {fni }
such that
(1) {fni+1 } is a subsequence of {fni };
(2) {fni (rk )}, k = 1, . . . , i is convergent.
Consider the subsequence {fnn }. We have
(1) {fnn } is a subsequence of {fn };
(2) {fnn (rk )}, for any k ∈ N, is convergent.
Let ε > 0. Since {fn } is equicontinuous, then there exists δ > 0 such that |fn (x) − fn (y)| < ε/3,
for any n ∈ N, provided |x − y| < δ. Since [0, 1] is compact, there exists {rn1 , . . . , rnK } such that
for any x ∈ [0, 1] there exists 1 ≤ i ≤ K such that |x − rni | < δ. Since {fnn (rk )}, for any k ∈ N, is
convergent, there exists N ∈ N such that for any n, m ≥ N ,
ε
|fnn (rni ) − fmm (rni )| <
3
for any i = 1, . . . , K. Let x ∈ [0, 1], then there exists 1 ≤ i ≤ K such that |x − rni | < δ. Hence
|fnn (x) − fmm (x)| ≤ |fnn (x) − fnn (rni )| + |fnn (rni ) − fmm (rni )| + |fmm (rni ) − fmm (x)| < ε.
So the subsequence {fnn } satisfies the uniform Cauchy criteria which imply that {fnn } is uniformly
convergent.

Solution 11.23

B is closed, bounded, and equicontinuous. Therefore, , b by the Arzelà–Ascoli Theorem B is compact.

Next we claim that T : B → R defined as T (f ) = a f (x)dx is continuous. Let {fk } be a sequence
of functions in B converging to some f uniformly. For ε > 0, we can choose N such that k ≥ N
ε
implies that |fk (x) − f (x)| < . Then
b−a
-# b # b - -# b -
- - - -
- f (x)dx − f (x)dx - = - f (x) − f (x)dx -
- k - - k -
a a a
# b
≤ |fk (x) − f (x)| dx
a
ε
≤ (b − a) = ε.
b−a
CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS 245

T is continuous on a compact set B, hence T (B) is also compact. Since compact subsets of R have
a maximum, there exists f0 ∈ B such that

T (f0 ) = max{T (f ); f ∈ B}.

Solution 11.24

a) The family Φ is uniformly bounded and equicontinuous, since

|f´(x)| ≤ |ab| + |cd| ≤ 2K 2 and |f (x)| ≤ |a| + |c| ≤ 2K

for all x ∈ [0, π] and any f ∈ Φ. Furthermore Φ is a closed subset of C[0, π], by the Arzelà–
Ascoli Theorem we conclude that Φ is compact in C[0, π].

b) Take g ∈ C[0, π], consider the distance d(g, Φ) = inf{d(g, h) : h ∈ Φ}. Since Φ is compact
this infimum is attained. In other words there exist values of a, b, c, and d in [−K, K] such
that
d(f, g) = max |g(x) − (a sin bx + c cos dx)|
0 ≤x≤π

is a minimum.

Solution 11.25

a) It is clear that the map Bn is linear, since Bn (f + g) = Bn f + Bn g and Bn (αf ) = αBn f .

Notice that when f ≥ 0, then Bn f is also positive. In particular, |f | ≤ g means −g ≤ f ≤ g
and hence −Bn g ≤ Bn f ≤ Bn g. This also proves that |Bn f | < Bn g if |f | < g (since it is
straightforward to show that f ≥ g ⇒ Bn f ≥ Bn g).
" $ %
b) Bn (1) = nk=0 1 nk xk (1 − x)n−k = 1 (From the Binomial Theorem). Next notice that
& ' & '
k n k n! (n − 1)! n−1
= = = .
n k n k!(n − k)! (k − 1)!(n − k)! k−1

Using the Binomial Theorem again, we have

n
! & '
k n k
Bn (x) = x (1 − x)n−k
n k
k=0
n &
! '
n − 1 k−1
=x x (1 − x)n−k
k−1
k=0

= x(x + (1 − x))n−1 = x.
246 CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS

c)
B1 (x) = f (0)(1 − x) + f (1)x = (1 − x) + ex = 1 + (e − 1)x.
1
B2 (x) = f (0)(1 − x)2 + 2f ( )x(1 − x) + f (1)x2 = (10x)2 + 2e1/2 x(1 − x)ex2
2
( )2 ( )2
= (1 − x) + e1/2 x = 1 + (e1/2 − 1)x .
More generally, we have
n & '
! n
Bn (x) = (e1/n x)k (1 − x)n−k = (e1/n x + (1 − x))n = (1 + (e1/n − 1)x)n .
k
k=0

d) It can be shown that Bn (ex ) may be written as (1 + x

n + cn
n2
) where 0 ≤ cn ≤ 1 and hence
Bn (ex ) converges uniformly to ex .

Solution 11.26

a) Let ε > 0. By the Weierstrass Approximation Theorem, there exists a polynomial q(x) such
that ∥q − f ∥ < ε/2. Suppose q has degree r, and

r
!
q(x) = ak xk
k=0
where some or all of the coefficients a0 , a1 , . . . , ar may be irrational. For each coefficient ak
we find a rational number bk such that |bk − ak | < ε/2(r + 1). Let p be the polynomial given
by

r
!
p(x) = bk xk .
k=0

Then, for all x ∈ [0, 1] we have

- r -
-! - ! r r
! ε ε
- k-
|p(x) − q(x)| = - (bk − ak )x - ≤ |bk − ak ||x|k ≤ = .
- - 2(r + 1) 2
k=0 k=0 k=0

Then ∥p − q∥ ≤ ε/2, so ∥p − f ∥ ≤ ∥p − q∥ + ∥q − f ∥ < ε.

b) To show C[a, b] separable we must show it contains a countable dense set. Let
n
!
A = {p(x) = bi xi | bi ∈ Q, n ∈ N}.
i=0

Then define

n
!
An = {p(x) = bi xi | bi ∈ Q},
i=0
CHAPTER 11. SEQUENCES AND SERIES OF FUNCTIONS 247

and we have
5
A= An .
n∈N

There is an obvious bijection between An and Qn+1 which sends each coefficient of a polyno-
mial in An to one coordinate in Qn+1 . Thus An is countable, and A is countable since it is a
countable union of countable sets.
So we must show that A is dense in C[a, b]. We follow the same proof as above:
Let ε > 0. By the Weierstrass Approximation Theorem, there exists a polynomial q(x) such
that ∥q − f ∥ < ε/2. Suppose q has degree r, and

r
!
q(x) = ak xk
k=0

where some or all of the coefficients a0 , a1 , . . . , ar may be irrational. Let c = max { sup {|x|i }}.
0 ≤i≤r a≤x≤b
Then for each coefficient ak we find a rational number bk such that |bk − ak | < ε/2c(r + 1).
Let p be the polynomial given by

r
!
p(x) = bk xk .
k=0

Then, for all x ∈ [0, 1] we have

- -
-!r - ! r r
! |x|k ε
r
! ε ε
- k-
|p(x) − q(x)| = - (bk − ak )x - ≤ k
|bk − ak ||x| ≤ ≤ = .
- - 2c(r + 1) 2(r + 1) 2
k=0 k=0 k=0 k=0

Then ∥p − q∥ ≤ ε/2, so ∥p − f ∥ ≤ ∥p − q∥ + ∥q − f ∥ < ε. Thus for any ε > 0 and f ∈ C[a, b]

we can find a p ∈ A with ∥p − f ∥ < ε, so A is dense in C[a, b].

Solution 11.27

Let Pn (x) = a0 + a1 x + · · · + an xn be a polynomial. Then by the given hypothesis,

# b # b # b # b
f (x)Pn (x)dx = a0 f (x)dx + a1 x.f (x)dx + · · · + an xn .f (x)dx = 0.
a a a a

By the Weierstrass Approximation Theorem, we know that any continuous function on [a, b] can be
uniformly approximated by polynomials. Hence there exists a sequence {Pn (x)} which converges
uniformly to f (x) on [a, b]. Since Pn ⇒ f , then f Pn ⇒ f 2 . Therefore
# b # b # b
lim f (x)Pn (x)dx = lim f (x)Pn (x)dx = f 2 (x)dx = 0.
n→∞ a a n→∞ a

Continuity of f (x) implies that f = 0 constantly (see Problem 7.19).