MAT3023/MT302 Real Analysis III Semester II - 2022/2023

Chapter 1: Sequences and Series of Functions

Lecturer: Udeni D. Wijesooriya

Disclaimer: These notes have not been subjected to the usual scrutiny reserved for formal
publications. They may be distributed outside this class only with the permission of the Instructor.

1.1 Sequences of Functions

A sequence of real numbers is a list (of infinite length) of real numbers, and we usually
denote
 them
 by
 (an ) or (x n ) where an (or (xn )) represents the general term. For exam-
1 1 1
ple, = 1, , , ... is a sequence of real numbers. We say the sequence (an ) is
n 2 3
convergent if lim an exists. A series of real numbers is an infinite sum of real numbers,
n→∞
X∞
denoted by an , where an represents the general term. Again, we say a series is conver-
n=1
gent/divergent if the sequence of nth partial sums is convergent/divergent.

A sequence of functions defined on E ⊂ R has the form (fn ) where fn is a function

defined on E for all n ∈ N. For example, (xn ) is a sequence of functions on R, where
fn (x) = xn for all n ∈ N. i.e, f1 (x) = x, f2 (x) = x2 and so on for all x ∈ R. Likewise, a
series of functions defined on E ⊂ R is an infinite sum of functions defined on the same
X∞
set E, denoted as fn , where each fn is a well defined function on E.
n=1

In this chapter, we focus on the convergence of sequences and series of functions. There
are two types of convergence we consider; pointwise convergence and uniform conver-
gence.

1-1
Lecture 1: Sequences and Series of Functions 1-2

1.1.1 Pointwise Convergence

Suppose (fn ) is a sequence of functions defined on some set E ⊂ R. For a fix x ∈ E, the
sequence (fn (x)) is a sequence of real numbers. If the sequence of real numbers (fn (x))
converges for each x ∈ E, we say the sequence of functions (fn ) converges pointwisely.

Definition 1.1. Let E ⊆ R and let (fn ) be a sequence of functions defined on E. If for each
x ∈ E, the sequence (fn (x)) converges, then we say that the sequence (fn ) of functions converges
pointwise on E.

In this case we define f on E by, f (x) = lim fn (x) for all x ∈ E and we simply say that
n→∞
the sequence (fn ) of functions defined on E converges pointwise to the function f on E.
This is also denoted by,
fn → f pointwise on E.

Here, f is called the pointwise limit function of the sequence (fn ). A more rigorous
definition can be given as:

Definition 1.2. Let E ⊆ R and let (fn ) be a sequence of functions defined on E. We say that the
sequence (fn ) of functions converges pointwisely to f on E, if for each x ∈ E and for each  > 0,
there exists N (, x) ∈ N such that |fn (x) − f (x)| <  for all n > N (, x).

Example 1.1. Let fn (x) = xn for all x ∈ [0, 1] and for all n ∈ N. The sequence (fn ) converges
pointwisely to the function given by,
(
0: 0≤x<1
f (x) = .
1: x=1

Exercise 1.1. Find the pointwise limit of (fn ), if

x
(i) fn (x) = for all x ∈ R and n ∈ N;
n
(ii) fn (x) = nx for all x ∈ R and n ∈ N; and

(iii) fn (x) = nx(1 − x2 )n for all x ∈ [0, 1] for all n ∈ N.

Lecture 1: Sequences and Series of Functions 1-3

One of the main problems arises when understanding of the converegence of sequences
of functions is to determine whether the important properties such as continuity, differ-
entiability, integrability, etc are ppreserved under the limit. In other words, if each fn is
continuous (or differentiable or integrable) and if (fn ) converegs to f pointwisely, is f
continuous (or differentiable or integrable)?

We will now illustrate through examples that those important properties are not pre-
served under the pointwise limit.

Example 1.2. Consider the sequence of functions in the Example 1.1. Each fn is continuous on
[0, 1], and (fn ) converges pointwisely to f on [0, 1]. However f is not continuous on [0, 1].

Therefore, the pointwise limit of a sequence of continuous functions need not to be

continuous.

Example 1.3. Consider the sequence (fn ) where fn (x) = nx(1−x2 )n for all x ∈ [0, 1] and n ∈ N.
Z 1 Z 1
n n 1
Observe that fn (x)dx = for all n ∈ N. So, lim fn (x)dx = lim = .
0 2(n + 1) n→∞ 0 n→∞ 2(n + 1) 2
Z 1
Let f be the pointwise limit of f (found in Exercise 1.1). Observe that f (x)dx = 0. Therefore,
Z 1 Z 1 Z 1 Z 10
lim fn (x)dx 6= f (x)dx. Equivalently, lim fn (x)dx 6= lim fn (x)dx
n→∞ 0 0 n→∞ 0 0 n→∞

i.e the limit of the integral is not equal to the integral of the pointwise limit limit.

sin (nx)
Example 1.4. Now consider (fn ) where fn (x) = √ for x ∈ [0, 2π] and n ∈ N and let f
n √
be the pointwise limit (to be determined). Note that f√n0 (x) = n cos (nx) for all x ∈ [0, 2π] and
n ∈ N and f 0 (x) = 0. However, lim fn0 (0) = lim n = ∞ = 6 f 0 (0). Therefore, (fn0 ) does not
n→∞ n→∞
d d
converge to f 0 pointwise. Equivalently, lim fn (x) 6= lim fn (x)
n→∞ dx dx n→∞

i.e the limit of the derivative is not equal to the derivative of the pointwise limit.

From these examples, its apparent that the pointwise convergence does not preserve
the important properties the functions int he sequence carried out. What hold greater
significance is that understanding what could go wrong if one interchange the process of
limit and integration or the process of limit and differentiation carelessly.
Lecture 1: Sequences and Series of Functions 1-4

1.1.2 Uniform Convergence

With that understanding, we now define a new mode of convergence, called uniform con-
vergence, a stronger version of the pointwise convergence to guarantee rather positive
conclusions for above. We will approach to the definition though the following exam-
ples:

Example 1.5. Using the  − N definition for pointwise convergence,

x
(i) prove that fn → 0 pointwisely on R, where fn (x) = for all x ∈ R and n ∈ N, and
n
sin (nx)
(ii) prove that fn → 0 pointwisely on [0, 2π], where fn (x) = √ for all x ∈ [0, 2π].
n

In the example (ii), unlike in the example (i), we can choose our N such that it depends
only on . i.e in example (ii), chosen N works for all x ∈ [0, 2π].

If we can pick an N such that it works for all x in the domain we are working on, then
we say that the convergence is uniform.

Definition 1.3. Let E be a non-empty subset of R and let (fn ) be a sequence of functions defined
on E. We say that (fn ) converges uniformly on E to the function f if, for each  > 0 there exists
N () ∈ N such that,

|fn (x) − f (x)| <  whenever n > N () and for all x ∈ E.

Note: Obviously, if fn → f uniformly on E, then fn → f pointwise on E. Is the

converse true? Further note that if fn → f uniformly on E, then f must be the pointwise
limit.

cos (nx)
Example 1.6. Show that the (fn ) converges uniformly on R, where fn (x) = for all
n
x ∈ R and for all n ∈ N.

x
Example 1.7. Let fn (x) = for all x ∈ R and for all n ∈ N. We already know that fn → 0
n
pointwise on R. We show that (fn ) does not converge to f uniformly.
Lecture 1: Sequences and Series of Functions 1-5

Suppose fn → 0 uniformly on R. We shall get a contradiction. Take any  > 0. Since fn → 0

uniformly, there exist N () ∈ N such that, |fn (x) − 0| <  for all n > N () and for all x ∈ R. So,
x
<  for all n > N () and for all x ∈ R. i.e,
n
x
− < <  for all n > N () and for all x ∈ R.
n

This is a contradiction. So (fn ) does not converge uniformly on R. That is, pointwise convergence
does not imply the uniform convergence.

Remark: Suppose fn → f uniformly on E. For each  > 0, there exists N () ∈ N such
that, |fn (x) − f (x)| <  for all n > N () and for all x ∈ E. i.e,

− < fn (x) − f (x) <  for all n > N () and for all x ∈ E.

i.e f (x) −  < fn (x) <  + f (x) for all n > N () and for all x ∈ E.

In other words, given an , functions in the tail of the sequence (fn ) can be trapped in
a region of average height 2, centered around f .
Lecture 1: Sequences and Series of Functions 1-6

Theorem 1.4. Let E ⊆ R and (fn ) be a sequence of functions defined on E such that fn −→ f
pointwise on E and let
Kn = sup |fn (x) − f (x)| .
x∈E

fn −→ f uniformly on E if and only if Kn −→ 0 as n −→ ∞.

Proof: Suppose fn −→ f uniformly on E. For each  > 0 there exists N () such that

|fn (x) − f (x)| < whenever n > N () and for all x ∈ E. Now
2

Kn = sup |fn (x) − f (x)| ≤ whenever n > N ().
x∈E 2

So, Kn ≤ <  whenever n > N (). Therefore, Kn −→ 0 as n −→ ∞.
2

Conversely, suppose Kn −→ 0 as n −→ ∞. For each  > 0 there exist N () ∈ N such that
|Kn − 0| <  whenever n > N (). i.e sup |fn (x) − f (x)| <  whenever n > N (). Therefore,
x∈E
|fn (x) − f (x)| <  whenever n > N () and for all x ∈ E. So, fn −→ f uniformly.

Example 1.8.

x
(i) Let fn (x) = for all x ∈ R, and for all n ∈ N. It is already proven that fn −→ 0 pointwise
n
on R. Is the convergence uniform?
nx
(ii) fn (x) = where x ∈ [0, 1]. Show that, fn → 0 pointwisely but not uniformly on
1 + n2 x2
[0, 1].
Lecture 1: Sequences and Series of Functions 1-7

x
(iii) Let fn (x) = for x ∈ [0, 2] and for all n ∈ N. Show that (fn ) converges uniformly
1 + nx2
on [0, 2].

Theorem 1.5. Let (fn ) be a sequence of continuous functions defined on E. If (fn ) converges
uniformly to a function f on E, then f is also continuous on E.

We may use this Theorem to show that given a sequence of function is not uniformly
convergent. The Contrapositive of the Theorem 3.2 is: If f is not continuous on E, then
fn 6→ f uniformly on E or there exist n ∈ N such that fn is not continuous.

(
0 ; x ∈ [0, 1)
Example 1.9. Let fn (x) = xn on [0, 1] for all n ∈ N andf (x) = . We already
1;x=1
know that fn → f pointwise on [0, 1]. Note that fn (x) = xn is continuous on [0, 1] for all n ∈ N.
However, f (x) is not continuous on [0, 1]. So fn 6→ f uniformly on [0, 1]

Proof:(of Theorem 1.5) Take any c ∈ E. To prove that f is continuous on E, we must

show that for each  > 0 there exist δ(, c) such that |f (x) − f (c)| <  when |x − c| < δ(, c).
Take any  > 0. Since fn → f uniformly, there exist N () ∈ N such that

|fn (x) − f (x)| < whenever n > N () and for all x ∈ E. (1.1)
3
Let N1 = N () + 1(> N ()) and let n = N1 in equation (1.1) to have

|fN1 (x) − f (x)| < for all x ∈ E. (1.2)
3
Since fN1 is continuous on E, it is continuous at x = c. Therefore, there exist δ(, c) such
that

|fN1 (x) − fN1 (c)| < whenever |x − c| < δ(, c). (1.3)
3
Observe that,
|f (x) − f (c)| = |f (x) − fN1 (x) + fN1 (X) − fN1 (c) + fN1 (c) − f (c)|
≤ |f (x) − fN1 (x)| + |fN1 (x) − fN1 (c)| + |fN1 (c) − f (c)|
<  whenever |x − c| < δ(, c) by equations (1.1), (1.2) and (1.3).
So f is continuous at c. Therefore, f is continuous on E.

NOTE: Suppose fn → f pointwisely on E and f is continuous. However, fn being

continuous on E for all n ∈ N does not imply the convergence is uniform.
Lecture 1: Sequences and Series of Functions 1-8

x2 + nx
For example, consider fn (x) = for all x ∈ R for all n ∈ N. Let f (x) = x on
n
R. Observe that (fn ) converges to f pointwisely, but not uniformly on R. However, each
fn and f are continuous on R. This example gives sequence of continuous functions that
converges pointwisely to a continuous function, however the convergence is not uniform.

Theorem 1.6. Let (fn ) be a sequence of Riemann integrable functions on [a, b]. If fn → f uni-
formly on [a, b], then f is also Riemann inegrable on [a, b], and
Z b Z b
lim fn (x)dx = f (x)dx.
n→∞ a a

that is, Z b Z b
lim fn (x)dx = lim fn (x) dx.
n→∞ a a n→∞

What this simply says is, if the convergence is uniform, interchanging the process of
limit and integration has no harm.

1
cos10 (nx)
Z
Example 1.10. Let fn (x) = for all x ∈ [0, 1] and for all n ∈ N. Find lim fn (x)dx.
n n→∞ 0

Note: Consider the sequence of functions given by fn (x) = xn on [0, 1] and for all n ∈ N.
Let (
0;0≤x<1
f (x) = .
1;x=1
We already know that fn converges to f pointwisely on [0, 1] but, not uniformly on [0, 1].
Z 1
Further, each fn and f are Riemann integrable on [0, 1]. Observe that lim fn (x)dx =
n→∞ 0
Z 1 Z 1 Z 1
1
lim = 0 and f (x)dx = 0. Therefore, lim fn (x)dx = f (x)dx. However
n→∞ (n + 1) 0 n→∞ 0 0
the convergence is not uniform.

Theorem 1.7. Let (fn ) be a sequence of functions on [a, b]. If

(i) each fn is differentiable on [a, b],

(ii) fn → f pointwise on [a, b] and

(iii) (fn0 ) converges uniformly on [a, b],

Lecture 1: Sequences and Series of Functions 1-9

then (fn ) converges uniformly to f on [a, b] and

lim fn0 (x) = f 0 (x) for all x ∈ [a, b].

n→∞

We conclude the discussion on the convergence of sequence of functions by address-

ing one of the very powerful tools we have in polynomial approximation theory, Stone-
Wiestrass Theorem.

Theorem 1.8. (Stone-Wiestrass Theorem) Every continuous function defined on [a, b] can be
written as a uniform limit of a sequence of polynomials.

That is, if f is a continuous function on [a, b], then there exist a sequence of polyno-
mials (Pn ) such that (Pn ) → f uniformly on [a, b]. This is the first form of this theorem
discovered by Karl Weistrass in 1885. Later Marshall H. Stone generalised this result, and
currently many versions are available.

1.2 Series of Functions

Now we discuss the convergence of series of functions. First recall that, given a series of
∞
X ∞
X
real numbers, an , we say that an converges say to s ∈ R, if the sequence (Sn ) of
n=1 n=1
∞
X
partial sums converges to s. i.e if lim Sn = s, then an = s. We adopt this definition
n→∞
n=1
to define pointwise and uniform convergence for series of functions, in terms of the se-
quence of partial sums formed by functions in the series.

Definition 1.9. Let (fn ) be a sequence of functions defined on E ⊆ R and let

n
X
Sn (x) = fk (x) for x ∈ E.
k=1

∞
X
We say that the series fn (x) converges pointwisely (uniformly) on E, if the sequence of partial
n=1
sums, (Sn ), converges pointwisely (uniformly) on E.
Lecture 1: Sequences and Series of Functions 1-10

Theorem 1.10. (Weistrass M-test) Let (fn ) be a sequence of functions defined on E and suppose
that there is a sequence of positive numbers (Mn ) such that
|fn (x)| ≤ Mn for all x ∈ E and n ∈ N.
∞
X ∞
X
If Mn converges, then fn (x) converges uniformly on E.
n=1 n=1

Example 1.11. Show that the following series converge uniformly on the given domain.

∞
X sin (nx)
(i) on R.
n=1
n2
∞
X
(ii) 2−n cos (3n x) on R.
n=0
∞
xn
 
X 1
(iii) on 0, .
n=1
n 2

∞
X
Theorem 1.11. Let fn (x) be a series of functions defined on E(⊆ R). If each fn is continuous
n=1
∞
X
on E and fn converges uniformly on E to a function f , then f is also continuous on E.
n=1

n
X
Proof: Take any n ∈ N. Let Sn = fk . Since f1 , f2 , f3 , .... are continuous on E,
k=1
∞
X
Sn , a finite sum of continuous functions is also continuous on E. Since fn converges
n=1
uniformly to f on E, by the definition, (Sn ) converges uniformly to f on E. So each Sn is
continuous on E and Sn → f uniformly on E. By Theorem 1.5, f is continuous on E.

∞
X
Theorem 1.12. Let fn (x) be a series of functions defined on [a, b]. If each fn is Riemann
n=1
∞
X
integrable on [a, b] and fn converges uniformly on [a, b] to a function f, then f is Riemann
n=1
integrable on [a, b] and
Z b ∞ Z
X b
f (x)dx = fn (x)dx.
a n=1 a
Lecture 1: Sequences and Series of Functions 1-11

In other words,
∞
! ∞ Z
Z b X X b
fn (x) dx = fn (x)dx.
a n=1 n=1 a

n
X
Proof: Take any x ∈ [a, b] and n ∈ N. Let Sn = fk (x). Since each fk is Riemann
k=1
∞
X
integrable on [a, b], Sn is Riemann integrable on [a, b]. Since fn converges uniformly
n=1
on [a, b] to a function f on [a, b], by the definition, Sn converges uniformly to f on [a, b]. By
the theorem (3.3), f is Riemann integrable on [a, b] and,
Z b Z b
f (x)dx = lim Sn (x)dx
a n→∞
Z ab X
n
= lim fk (x)dx
n→∞ a k=1
n Z
X b
= lim fk (x)dx
n→∞ a
k=1
∞ Z
X b
= fk (x)dx.
k=1 a

∞
! ∞ Z
Z b X X b
So, fn (x) dx = fn (x)dx.
a n=1 n=1 a

Example 1.12.

∞ π ∞
(−1)n
Z
X cos (nx) 2 X
1. Let f (x) = . Show that f (x)dx = .
n=1
n2 0 n=0
(2n + 1) 3

Z ∞
πX
sin((2n + 1)x)
2. Evaluate dx.
0 n=1
n3 + 1