Chapter 2

Riemann integrals

Motivation
Consider the function f : x 7→ e x .
• How to calculate the area A

(1) We approximate this area with sums of the areas of rectangles located below the curve.

• Let ngeq1 be a natural number. Consider the subdivision d = {0, n1 , n2 , · · · , kn , · · · , n− 1

n , 1} of
h the interval
i
−
[0, 1]. • We consider the ’lower rectangles’ denoted by Rk , each having as its base the interval k− 1 k
k , n and

19
20 CHAPTER 2. RIEMANN INTEGRALS

k −1
a height of f ( k− 1
n )=e
n , k ∈ {1, · · · , n}. • The area of R−
k is ’base × height’:

k−1
 
k k −1 1 k −1
− ×e n = e n .
n n n

• The sum of the areas of R−

k is calculated as a geometric sequence:

n k −1 n 1 1
e n 1 1 1 1 − (e n )n
∑ n
=
n ∑ ( e n ) k −1 = n 1 − e n1
= 1
n
(e − 1) → e − 1 n → +∞.
k =1 k =1 en − 1
x
For the limit, we recognized the expression of the form e x−1 → 1, as x → 0 (here x = n1 ).
(2) Now we approximate this area with sums of areas of rectangles located above the curve.

k
• Now consider the ’upper rectangles’ R+ k −1 k k
k , having the same base [ n , n ] but the height f ( n ) = e . A similar
n
k
calculation shows that ∑nk=1 en
n → e − 1 as n 7→ +∞.

we have
 21

n k −1 n k
e n en
∑ n
<A< ∑
n
k =1 k =1

however
n k −1 n k
e n en
lim
n→+∞
∑ n
= lim
n→+∞
∑ n
= e − 1,
k =1 k =1

then
A = e − 1.
22 CHAPTER 2. RIEMANN INTEGRALS

Introduction
Let f : [ a, b] → R be a positive bounded function.

• We aim to measure the area of the surface S bounded by the graph of f and the x-axis (see Figure). • This area,
Rb
as we will see in the following, denoted by a f ( x)dx is referred to as the integral of f . • Note that the Riemann
integral, which we will develop later, allows us to integrate more than just positive functions. The case of
positive functions is used solely to illustrate the intuition behind the construction of the Riemann integral.
• The concept of the Riemann integral involves dividing the region S into increasingly smaller rectangular
subregions, which we use to approximate the area of S.

The sign of the integral

Rb
a f ( x)dx corresponds to the area of the region in the plane between the graph of f and the x-axis, counted:
• positively for the part above the x-axis,
• negatively for the part below.

2.1 Construction of the Riemann Integral

Definition 2.1(1)

[( Partition of an interval)] A partition P of a closed bounded interval [ a, b] is a finite sequence of points

in [ a, b] containing the points a and b. The following notation will be used to denote such a partition:

P = { a = x 0 < x 1 < · · · < x n −1 < x n = b }.

• Occasionally, the term ’partition’ will be used synonymously with ’subdivision’.

2.1. CONSTRUCTION OF THE RIEMANN INTEGRAL 23

Definition 2.1(2)

Let P = { a = x0 < x1 < · · · < xn−1 < xn = b} be a subdivision of the interval [ a, b].
• The step or norm of the subdivision P is defined as the maximum of the lengths of the subintervals
[ xi , xi+1 ], 0 ≤ i ≤ n − 1, comprising the subdivision. In other words:

max ( x i +1 − x i ) .
i ∈{0,1,··· ,n−1}

he step of P is denoted by µ(P ).

• When the subintervals [ xi , xi+1 ], 0 ≤ i ≤ n − 1 have the same length, we say that the subdivision is
regular or uniform.

• Let Q be another subdivision of [ a, b], we say that Q is finer than P (or Q is a refinement of P) if
P ⊂ Q.

Example 2.1(1)

(1) The simplest example of a subdivision of an interval [ a, b] is P = { a = x0 < x1 = b}.

(2) LThe sequence P = {0, 14 , 13 , 23 , 1} is a subdivision of the interval [0, 1] with step µ(P ) = 31 . This
subdivision is not regular.
(3)Let Q = {0, 61 , 14 , 10
3 1 3 2 4
, 3 , 5 , 3 , 5 , 1} be a subdivision of [0, 1] finer than P. The longest subinterval in
1 3
this subdivision is [ 3 , 5 ] hence
3 1 4
µ( Q) = − = .
5 3 15
(4) Let n ∈ N∗ and [ a, b] be an interval in R. By dividing the interval [ a, b] into n subintervals of equal
length, we obtain a regular subdivision

Pn = { a = x0 < x1 < · · · < xn−1 < xn = b}.

24 CHAPTER 2. RIEMANN INTEGRALS

where
b−a
xk = a + k , ∀k ∈ {0, 1, · · · , n}.
n
The step of this subdivision is
b−a
µ(P) = .
n

Definition 2.1(3)

[(Darboux Sums)] Let f : [ a, b] → R be a bounded function and p = { a = x0 < x1 < · · · < xn−1 <
xn = b} be a partition of the interval [ a, b]. we define for each i ∈ {1, 2, · · · , n}

mi = inf f ( x) and Mi = sup f ( x ).

x∈[ xi−1 ,xi ] x∈[ xi−1 ,xi ]

Then, the lower Darboux sum of f relative to the partition p is defined as:

n
dp( f ) = ∑ ( x i − x i −1 ) m i
i =1

And the upper Darboux sum of f relative to p is defined as:

n
Dp ( f ) = ∑ ( xi − xi−1 ) Mi
i −1

Remark 2.1(1)

1. Note that the condition that f be bounded is necessary here to ensure that the Darboux sums
are well-defined. Indeed, since f is bounded on [ a, b], and therefore bounded on each subin-
terval [ xk−1 , xk ], k ∈ {1, 2, · · · , n}, the supremum principle assures us that infx∈[ xk−1 ,xk ] and
supx∈[ xk−1 ,xk ] , exist and are finite.

2. The upper Darboux sum (resp. lower Darboux sum) simply calculates the sum of the areas of the
rectangles with base [ xk−1 , xk ] and height supx∈[ xk−1 ,xk ] ( resp. infx∈[ xk−1 ,xk ] ).

• Returning to the context of a positive function f , let S denote the region in the plane bounded by the
x-axis, the vertical lines x = a and x = b, and the graph of the function f . One can visually convince
oneself that for all partitions p and q of the interval [ a, b], we have:

d p ( f ) ≤ area o f S ≤ D p ( f ).

• As the subdivisions p and q become finer, the lower Darboux sum d p ( f ) ) increases while approaching
2.1. CONSTRUCTION OF THE RIEMANN INTEGRAL 25

aire(S),and the upper Darboux sum D p ( f ) decreases towards aire(S).

• Figures 5 and 6 illustrate these points well. In these figures, d p and D p respectively denote the values
of the lower and upper Darboux sums of the function f ( x) = x2 on the interval [0, 4] with respect to a
uniform subdivision. Compare these values with the area of the surface bounded by the graph of f and
the segment [0, 4], which is equal to 64
3 .
26 CHAPTER 2. RIEMANN INTEGRALS

All the aforementioned statements are summarized in the following proposition.

Properties 2.1(1)

Let f : [ a, b] 7→ R be a bounded function, and let p and q be two partitions of the interval [ a, b].

1. We have
D p ( f ) ≥ d p ( f ).

2. If q is finer than p, then

d p ( f ) ≤ dq ( f ) and Dq ( f ) ≤ D p ( f ).

3. We have
d p ( f ) ≤ Dq ( f ) .

Let f : [ a, b] 7→ R be a bounded function. Suppose that m ≤ f ( x) ≤ M, ∀ x ∈ [ a, b], If p and q are two

arbitrary partitions of [ a, b], then

m ( b − a ) ≤ d p ( f ) ≤ Dq ( f ) ≤ M ( b − a ) .
2.1. CONSTRUCTION OF THE RIEMANN INTEGRAL 27

Definition 2.1(4)

Let f : [ a, b] 7→ R be a bounded function. Let [ a, b] denote the set of all subdivisions of [ a, b]. The
upper Darboux integral of f is defined as:
Z b
f ( x)dx = inf D p ( f )
a p∈[ a,b]

And the lower Darboux integral is defined as:

Z b
f ( x)dx = sup d p ( f )
a p∈[ a,b]

Remark 2.1(2)

• The property stated in Corollary 1 and the supremum principle ensure the existence of the upper
and lower Darboux integrals.

• The reader can easily verify that

Z b Z b
f ( x)dx ≤ f ( x)dx
a a

The following proposition will be useful in the upcoming sections (you can ignore it for now and return to
it only if necessary).

Properties 2.1(2)

Let f , g : [ a, b] 7→ R be two functions, and c ∈ R a real constant. The following properties hold:

1.
Z b Z b Z b
( f ( x) + g ( x))dx ≤ f ( x)dx + g ( x)dx
a a a

2. Z b Z b Z b
f ( x)dx + g ( x)dx ≤ ( f ( x) + g ( x))dx
a a a

Rb Rb Rb Rb
3. a c f ( x)dx = c a f ( x)dx and a c f ( x)dx = c a f ( x)dx if c ≥ 0.
Rb Rb Rb Rb
4. a c f ( x)dx = c a f ( x)dx and a c f ( x)dx = c a f ( x)dx if c ≤ 0.

In light of all the foregoing, one might think that the upper Darboux integral is equal to the lower Darboux
integral. In reality, this is not always the case (see Example 2), which leads us to the following definition:

Definition 2.1(5)

[( Riemann Integral)] Let f : [ a, b] → R be a bounded function. We say that f is Riemann integrable

over [ a, b] if
Z b Z b
f ( x)dx = f ( x)dx.
a a
28 CHAPTER 2. RIEMANN INTEGRALS

Rb
We then define the Riemann integral of f , denoted a f ( x)dx by

Z b Z b Z b
f ( x)dx = f ( x)dx = f ( x)dx.
a a a

Example 2.1(2)

• Show that the constant function f ( x) = c, c ∈ R, is Riemann integrable over [a, b].

Exercice 2(1)

The Dirichlet function is defined on [0, 1] as


1 si x ∈ [0, 1] ∩ Q,
f ( x) =
0 si x ∈ [0, 1] Q

Calculate the lower and upper Darboux sums of f .

Remark 2.1(3)

In truth, it is quite rare to resort to the definition to show that a certain function is Riemann integrable.
In the following sections, we will see several sufficient conditions for integrability that are easy to
verify, and the Fundamental Theorem of Calculus (discussed in Section 6) provides a straightforward
method for calculating the value of a Riemann integral.

2.2 Integrability Criterion

The following theorem, which provides a necessary and sufficient condition for integrability, will be very
useful in proving certain properties of the Riemann integral. Some refer to it as Cauchy’s integrability criterion
due to its resemblance to Cauchy’s criterion for the convergence of real sequences.

Theorem 2.2(1)

A bounded function f : [ a, b] 7→ R is Riemann integrable if, for every ϵ > 0, there exists a partition p
of [ a, b] such that
| D p ( f ) − d p ( f )| < ϵ.

2.2.1 Step Functions

An example of very simple functions that is interesting to study for Riemann integrability is the step functions.

Definition 2.2(1)

A function f : [ a, b] → R is called a ”step function” if there exists a partition p = { a = x0 < x1 <

· · · < xn = b} of [ a, b] such that f is constant on each subinterval ] xk−1 , xk [ for all k ∈ {1, · · · , n}.
2.2. INTEGRABILITY CRITERION 29

• Such a partition p is said to be associated with f .

• Note that if q = { a = y0 < y1 < · · · < yn = b} is another partition of [ a, b] finer than p, then f will
also be constant on each subinterval ]yk−1 , yk [ of q. Thus, any partition finer than p is also associated
with the function f .

Example 2.2(1)

[(Step Function)] The floor function, denoted by E( x), which associates each x ∈ R with its floor value,
is a step function. For example, on the interval [0, 25 ], we have

0

 si x ∈ [0, 1[,
E( x) = 1 si x ∈ [1, 2[,

si x ∈ [2, 25 .

2

30 CHAPTER 2. RIEMANN INTEGRALS

Properties 2.2(1)

Let f : [ a, b] → R be a step function, and let p = { a = x0 < x1 < · · · < xn = b} be a partition

of [ a, b] associated with f . For every k ∈ {1, · · · , n}, let ck denote the value of f on the sub-interval
] xk−1 , xk [.is Riemann integrable, and we have
Z b n

a
f ( x)dx = ∑ c k ( x k − x k −1 ) .
k =1

Remark 2.2(1)

The result of the previous proposition was entirely expected since we already know how to calculate
the area of the region bounded by f and the x-axis when f is a step function. This region is a union of
rectangular sub-regions, as illustrated in Figure 7.

Example 2.2(2)

I ( f ) = λ0 ( x1 − x0 ) + λ1 ( x2 − x1 ) + λ2 ( x3 − x2 ) + λ3 ( x4 − x3 ) + λ4 ( x5 − x4 ) + λ5 ( x6 − x5 )
= −S1 + S2 − S3 + S4 + S5 + S6

Remark 2.2(2)

The real number I ( f ) evaluates the algebraic area bounded by the graph of f , the x-axis (Ox), and the
lines with equations x = a and x = b.
2.3. RIEMANN SUMS 31

2.3 Riemann Sums

Definition 2.3(1)

1. Let [ a, b] be a closed and bounded interval. A marked subdivision σ = ( p, τ ) of [ a, b] is the

specification of a partition p = { a = x0 < x1 < · · · < xn = b} and a finite sequence of
elements τ = {t1 < t2 < · · · < tn } such that tk ∈ [ xk−1 , xk ], ∀k ∈ {1, · · · , n}.

2. For a function f : [ a, b] → R, the Riemann sum of f relative to the marked subdivision σ =

( p, τ ) is defined as:
n
Sσ ( f ) = ∑ ( x k − x k −1 ) f ( tk ) .
k =1

Example 2.3(1)

Often, one chooses tk among the endpoints of the interval [ xk−1 , xk ]. Let pn = { a = x0 < x1 < · · · <
xn = b} be the regular subdivision of [ a, b] given in Example 1, with

b−a
xk = a + k , ∀k ∈ {0, 1, · · · , n}.
n

For example, choosing tk = xk , ∀k ∈ {0, 1, · · · , n}, the Riemann sum of f relative to pn , is then
expressed as
n
b−a n b−a
 
S p n ( f ) = ∑ ( x k − x k −1 ) f ( x k ) =
n k∑
f a + k .
k =1 =1
n

Theorem 2.3(1)

Let f : [ a, b] → R be a bounded function, and (σn )n a sequence of marked subdivisions of [ a, b].

Suppose that limn→+∞ µ(σn ) = 0. If f is Riemann integrable over [ a, b], then
Z b
lim Sσn ( f ) = f ( x)dx.
n→+∞ a

In particular, considering the sequence of subdivisions pn from the previous example, we obtain

n Z b
b−a b−a
 
lim
n→+∞ n ∑ f a+k
n
=
a
f ( x)dx...............(∗)
k =1

• One can use the limit (*) to approximate the value of the integral of a function. In order to study the
effectiveness of such a method, we must first estimate the error incurred when attempting to approximate the
integral of f with such a sum. This will be the subject of the following exercise:

Example 2.3(2)

Apply the results of the previous exercise to the function f : [ a, b] → R defined by

1
f ( x) = .
1+x

f is Riemann integrable because it is continuous (see Theorem 12).

32 CHAPTER 2. RIEMANN INTEGRALS

According to Theorem 6,

n
1
lim Sn ( f ) = lim
n→+∞ n→+∞ n ∑ f ( xk )
k =1
n
1 1
= lim
n→+∞
∑ n k
k =1 1 + n
n
1
= lim
n→+∞
∑ n + k
k =1
Z 1
dx
= .
0 1+x

And we have, by Exercise 2

Z 1 n
dx 1 1
− ∑ ≤ .
0 1 + x k =1 n + k 2n
R1
So, to obtain an approximate value of 0 f ( x)dx with an accuracy of 10−2 , one should choose n = 500!.
We then calculate:
Z 1 500
dx 1
≈ ∑ = 0.6926.
0 1+x k =1
500 +k

2.4 Properties of the Riemann Integral

Theorem 2.4(1)

[(Linéarité)] Let f , g : [ a, b] → R be two Riemann integrable functions, and c ∈ R real constant. Then,
f + g and c f are Riemann integrable, and the following hold:
Rb Rb Rb
1. a ( f ( x) + g ( x))dx = a f ( x)dx + a g ( x)dx.
Rb Rb
2. a c f ( x)dx = c a f ( x)dx.

Properties 2.4(1)

Let f and g be two Riemann integrable functions on [ a, b]. The following properties hold:

1. Positivity of the Integral: If f ( x) ≥ 0, ∀ x ∈ [ a, b], then

Z b
f ( x)dx ≥ 0.
a

2. Monotonicity If f ( x) ≤ g ( x), ∀ x ∈ [ a, b], then

Z b Z b
f ( x)dx ≤ g ( x)dx.
a a

3. | f | is integrable over [ a, b], and we have:

Z b Z b
f ( x)dx ≤ | f ( x)|dx.
a a
2.4. PROPERTIES OF THE RIEMANN INTEGRAL 33

R
4. If m = x∈[ a,b] f ( x) and M = supx∈[a,b] f ( x) then

Z b
m(b − a) ≤ f ( x)dx ≤ M (b − a).
a

Theorem 2.4(2)

Let c ∈] a, b[, If f is integrable over [ a, c] and [c, b], then f is integrable over [ a, b], and the Chasles
relation holds: Z b Z c Z b
f ( x)dx = f ( x)dx + f ( x)dx.
a a c

Definition 2.4(1)

For b < a, we define:

Z b Z a
f ( x)dx = − f ( x)dx,
a b

and for b = a, we have Z a

f ( x)dx = 0.
a

The previous theorem takes the following more general form:

Theorem 2.4(3)

Let a, b, c ∈ R be real numbers. The following relation holds:

Z b Z c Z b
f ( x)dx = f ( x)dx + f ( x)dx.
a a c

for any integrable function over the interval [min( a, b, c), max( a, b, c)].

Theorem 2.4(4)

[Mean Value Theorem] Let f : [ a, b] → R be a continuous function. Then,

Z b
1
∃c ∈ [ a, b], f ( x)dx = f (c).
b−a a

Proof. Since f is continuous on [ a, b], it is integrable over [ a, b]. Moreover, f is bounded and attains its
bounds, meaning there exist : ∃ x1 , x2 ∈ [ a, b] such that

f ( x1 ) = inf f ( x) and f ( x2 ) = sup f ( x)

x∈[ a,b] x∈[ a,b]

and

f ( x1 ) ≤ f ( x ) ≤ f ( x2 )
Z b
⇒ ( b − a ) f ( x1 ) ≤ f ( x)dx ≤ (b − a) f ( x2 )
a
Z b
1
⇒ f ( x1 ) ≤ f ( x)dx ≤ f ( x2 ).
b−a a
34 CHAPTER 2. RIEMANN INTEGRALS

According to the Intermediate Value Theorem, there exists

Z b
1
∃c ∈ [ a, b], f (c) = f ( x)dx.
b−a a

2.5 Sufficient Conditions for Integrability

In this section, we’ll state a few theorems that, in most cases, are sufficient by themselves to show that a
function is Riemann integrable.

Theorem 2.5(1)

Let f : [ a, b] → R be a monotone function. Then f is Riemann integrable on [ a, b].

Theorem 2.5(2)

Any continuous function f : [ a, b] → R is Riemann integrable on [ a, b].

Theorem 2.5(3)

Let f , g : [ a, b] → R be two functions that take the same values except possibly at a finite number
of points in [ a, b]. Then, f is integrable if and only if g is integrable, and in the case where they are
integrable, we have
Z b Z b
f ( x)dx = g ( x)dx.
a a

Corollary 2.5(1)

Let f : [ a, b] → Rbe a function. Suppose that the set of discontinuity points of f is finite. Then, f is
Riemann integrable. Denoting c1 , c2 , · · · , cn as the points of discontinuity of f , we have the following
formula: Z b Z c1 Z b
f ( x)dx = f ( x)dx + · · · + f ( x)dx
a a cn

Proof. Let’s assume that f is discontinuous at a single point a < c < b, the general case will
follow easily. f is continuous on [ a, c[. Let’s define g : [ a, c] → R by

 f ( x) si x ∈ [ a, c[
g ( x) =
limx→c f ( x) si x = c.

Clearly, g is continuous on [ a, c] and therefore integrable. Since f differs from g at only one point
c, we deduce that f is integrable on [ a, c]. Similarly, we can show that f is also integrable on
[c, b], and the result follows from Theorem 9.
2.6. FUNDAMENTAL THEOREM OF INTEGRAL CALCULATIONS 35

2.6 Fundamental Theorem of Integral calculations

Definition 2.6(1)

Let D ⊂ R and f : D → R be a function. We call a ’primitive’ of f on D any function F : D → R

satisfying:

1. F is differentiable on D.

2. F ′ = f in D

Properties 2.6(1)

If F and G are primitives of f , then ( F − G ) is a constant on any interval I ⊂ D.

Proof. Let I be an interval in D. On I, we have

( F − G )′ ( x) = F ′ ( x) − G ′ ( x)
= f ( x ) − f ( x ).

Then
∀ x ∈ I, ( F − G )′ ( x) = 0.

According to the mean value theorem, we have

F − G = c,

on I, where c is a constant.
The following theorem and its corollary highlight the relationship between differentiation and integration,
which, to some extent, can be seen as inverse operations.

Theorem 2.6(1)

Let f : I → R be a continuous function defined on an interval I, and a ∈ I. We define F : I → R by

Z x
F ( x) = f (t )dt.
a

Then F is differentiable, and F ′ = f .

An immediate consequence of the previous theorem is that any continuous function defined on an interval
has infinitely many primitives. The following corollary states that any two primitives of f on an interval differ
only by a constant. Let f : I → R be a continuous function defined on an interval I of R and a ∈ I. If
F : I → R is a primitive of f , then there exists a constant C ∈ R such that
Z x
F ( x) = f (t )dt + C.
a

Theorem 2.6(2)

[(Fundamental Theorem of Calculus)] Let f : [ a, b] → R bea continuous function and F : [ a, b] → R

36 CHAPTER 2. RIEMANN INTEGRALS

is a primitive of f , then
Z b
f (t )dt = F (b) − F ( a).
a

We can replace the continuity condition in the previous theorem with the integrability of f .

Theorem 2.6(3)

Let f : [ a, b] → R be a Riemann integrable function on [ a, b] and F : [ a, b] → R is a primitive of f ,

then Z b
f (t )dt = F (b) − F ( a).
a

• Although this theorem is more general than theorem 19, its use is limited. In theorem 19, the continuity
assumption ensures the existence of a primitive for f in [ a, b] (see theorem 17), which is not the case with the
integrability condition. The following function, for example:

1 si x ∈ [0, 1],
f ( x) =
 −1 si x ∈ [−1, 0[.

is integrable on [−1, 1], however, it does not have a primitive.

Definition 2.6(2)

The ’indefinite integral’ of a function f : [ a, b] → R is the set of all primitives of f , denoted by

R
f ( x)dx.

Remark 2.6(1)

• If F is a primitive of f , then we write:

Z
f ( x)dx = F ( x) + C,

where C is a real constant.

• Knowing a primitive of f on [ a, b] allows us to calculate the integral of f over the interval [ a, b].

• In practice, one often seeks a primitive using various methods instead of using Riemann sums.

2.7 Rules for Calculations on Riemann Integrable Functions

2.7.1 Primitives of Common Functions

By differentiation, we can verify that
α +1
1. xα dx = xα+1 + C; α ∈ R {−1}, C ∈ R;
R

x = ln | x | + C, C ∈ R;
R dx
2.

3. sin xdx = − cos x + C, C ∈ R;

R

4. cos xdx = sin x + C C ∈ R;

R
2.7. RULES FOR CALCULATIONS ON RIEMANN INTEGRABLE FUNCTIONS 37

dx
= − cot x + C C ∈ R;
R
5.
sin2 x
dx
= tan x + C C ∈ R;
R
6. cos2 x
dx
= ln | tan 2x | + C C ∈ R;
R
7. sin x
dx
= log |tan( π4 + 2x )| + C C ∈ R;
R
8. cos x

x+ a
dx 1
+ C C ∈ R;
R
9. a2 − x2
= 2a ln x− a

dx 1 x
+ C C ∈ R;
R
10. a2 + x2
= a arctan a

√ dx 1
arctan |ax | + C C ∈ R;
R
11. = a
a2 − x2
√
√ dx x2 − a2 ) + C C ∈ R;
R
12. = ln( x +
x2 − a2
√
√ dx a2 + x2 ) + C C ∈ R;
R
13. = ln( x +
a2 + x2

2.7.2 Integration by Parts

Properties 2.7(1)

Let f and g be two C 1 functions on [ a, b]. Then,

Z b Z b
f ′ ( x) g ( x)dx = [ f ( x) g ( x)]ba − f ( x) g ′ ( x)dx.
a a

Example 2.7(1)
Z 1
I1 = arctan xdx
0

We integrate by parts. Let’s set f ′ ( x) = 1 et g ( x) = arctan x.

f ′ ( x) = 1 ⇒ f ( x) = x

1
g ( x) = arctan x ⇒ g ′ ( x) = .
1 + x2
then Z 1 Z 1
x
I1 = arctan xdx = [ x arctan x]10 − dx.
0 0 1 + x2
Z 1
1 2x π 1
I1 = (arctan 1 − arctan 0) − 2
dx = − 0 − [ln(1 + x2 )]10
2 0 1+x 4 2
thus
π 1
I1 = − ln 2.
4 2

Example 2.7(2)
Z
I2 = arcsin xdx

We integrate by parts. Let’s set f ′ ( x) = 1 et g ( x) = arcsin x.

f ′ ( x) = 1 ⇒ f ( x) = x
38 CHAPTER 2. RIEMANN INTEGRALS

1
g ( x) = arcsin x ⇒ g ′ ( x) = √ .
1 − x2
then
x
Z Z
I2 = arcsin xdx = x arcsin x − √ dx.
1 − x2

1 2x
Z
I2 = x arcsin x + 1 dx
2 (1 + x2 ) 2
−2x h′
= − (n−11)hn−1 + C, where C ∈ R.
R R
1 dx is of the form hn
(1+ x2 ) 2
Therefore
−2x 1
Z p
1 dx = − 1 + C = 2 1 − x2 + C, W here C ∈ R.
(1 + x2 ) 2 ( 12 − 1)(1 − x2 ) 2 −1

thus p
I2 = x arcsin x + 1 − x2 + C, where C ∈ R

2.7.3 By change of variables

1- f (e x )dx
R

We set t = e x ,
dt
t = e x ⇒ dt = e x dx = tdx ⇒ dx = .
t
Then
f (t )
Z Z
f (e x )dx = dt.
t

Example 2.7(3)

dx
Z
I3 =
3 + e− x
dt
By change of variables. Let’s set t = e x ⇒ dt = e x dx = tdx ⇒ dx = t . Then

dt
Z
I3 =
t (3 + 1t )
dt
Z
=
3t + 1
1 3
Z
= dt,
3 3t + 1
1
= ln(3t + 1) + C, where C ∈ R.
3
Thus,
1
I3 = ln(3e x + 1) + C, where C ∈ R.
3

Example 2.7(4)

dx
Z
I4 =
ex +1
2.7. RULES FOR CALCULATIONS ON RIEMANN INTEGRABLE FUNCTIONS 39

dt
By change of variables. Let’s set t = e x ⇒ dx = t . Then

1 dt
Z
I4 =
t+1 t
1+t−t
Z
=
t ( t + 1)
t+1 t
Z Z
= dt − dt,
t ( t + 1) t ( t + 1)
dt dt
Z Z
= − ,
t t+1
 
t
= ln t − ln(t + 1) + C = ln + C, where C ∈ R.
t+1

Thus,
ex
 
I4 = ln + C, où C ∈ R.
ex + 1

Remark 2.7(1)

To calculate the indefinite integral of the previous example, we can perform the following change of
variable:
√ 2t
t = 1 + e x ⇒ dx = 2 dt.
t −1

2- f (ln x) dx
R
x
dx
We set t = ln x ⇒ dt = x .
Thus,
dx
Z Z
f (ln x) = f (t )dt
x

Example 2.7(5)

dx
Z
I5 =
x(ln x)2
dx
By change of variables. Let’s set t = ln x ⇒ dt = x donc

dt 1 1
Z
I5 = = − +C = − + C, where C ∈ R.
t2 t ln x

Therefore
1
I5 = − + C, where C ∈ R.
ln x

Example 2.7(6)

dx
Z
I6 = p
x 1 − (ln x)2
dx
By change of variables. Let’s set t = ln x ⇒ dt = x then

dt
Z
I6 = √ = arcsin(t ) + C = arcsin(ln x) + C, where C ∈ R.
1 − t2
40 CHAPTER 2. RIEMANN INTEGRALS

Therefore
I6 = arcsin(ln x) + C, where C ∈ R.

R
3- f (sin x, cos x)dx
We set t = tan 2x .

x x (sin 2x cos 2x + sin 2x cos 2x )/ cos2 x

2 (2 sin 2x cos 2x )/ cos2 2x
sin x = sin + = =
2 2 (sin2 2x + cos2 2x )/ cos2 2x (sin2 2x + cos2 2x )/ cos2 2x
2 tan 2x
=
tan2 2x + 1
2t
= 2
t +1

Thus
2t
sin x = .
t2 + 1

x
x x (cos2 2 − sin2 2x )/ cos2 x
2 1 − tan2 x
2
cos x = cos + = x
= x
2 2 (cos2 2 + sin2 2x )/ cos2 x
2
1 + tan2 2
1 − t2
=
1 + t2

hence
1 − t2
cos x = .
1 + t2
x 2dt
t = tan ⇒ x = 2 arctan x ⇒ dx =
2 1 + t2
Special Case
R
1. f (sin x). cos xdx
We set t = sin x ⇒ dt = cos xdx. Then
Z Z
f (sin x). cos xdx = f (t )dt.

x
Alternatively, we use the change t = tan 2

R
2. f (cos x). sin xdx
We set t = cos x ⇒ −dt = sin xdx. Then
Z Z
f (cos x). sin xdx = − f (t )dt.

x
Alternatively, we use the change t = tan 2

sin p cosq xdx, where p, q, ∈ N.

R
3.
If either of the natural numbers p or q is even, we can reduce it to one of the two previous cases (
R R
f (sin x). cos xdx or f (cos x). sin xdx).

Example 2.7(7)
Z Z Z
sin3 x cos2 xdx = sin2 x cos2 x sin xdx = (1 − cos2 x) cos2 x sin xdx.
2.7. RULES FOR CALCULATIONS ON RIEMANN INTEGRABLE FUNCTIONS 41

If both are even, we set 

sin x eix −e−ix
= 2i
cos x eix +e−ix
= 2

Example 2.7(8)

1
Z
I7 = cos xdx
1 + sin2 x
R
This integral is in the form f (sin x) cos xdx.
By change of variables. Let’s set t = sin x ⇒ dt = cos xdx. Then

dt
Z
I7 = = arctan t + C = arctan(sin x) + C, where, C ∈ R.
1 + t2

Thus
I7 = arctan(sin x) + C, where, C ∈ R.

Example 2.7(9)
Z
I8 = (sin 2x)3 dx

We have
Z Z Z
I8 = (sin 2x)3 dx = (2 sin x cos x)3 dx = 8 sin3 x cos3 xdx
Z
=8 sin3 x cos2 x cos xdx
Z
=8 sin3 x(1 − sin2 x) cos xdx

R
This integral is in the form f (sin x) cos xdx.
By change of variables. Let’s set t = sin x ⇒ dt = cos xdx. Then
!
t4 t6 sin4 sin6
Z Z  
3 2 3 5
I8 = 8 t (1 − t )dt = 8 (t − t )dt = 8 − +C = 8 − + C, where, C ∈ R.
4 6 4 6

Thus
4
I8 = 2 sin4 x − sin6 x + C, where, C ∈ R.
3

2.7.4 Integration of Rational Fractions

Decomposition into Simple Elements

Definition 2.7(1)

P( x)
1. Let P and Q be two real polynomials. The function x 7→ Q( x)
, defined for x such that Q( x) ̸= 0,
is called a ’rational fraction’.
Bx+C
2. The functions x 7→ A
( x − x0 ) m
and x 7→ ( x2 +rx+s)m
where m, n ∈ N∗ ; A, B, C, x0 , r, s ∈ R and
∆ = r 2 − 4s < 0 , are called ’simple elements’ of the 1st and 2nd kinds, respectively.
42 CHAPTER 2. RIEMANN INTEGRALS

Example 2.7(10)

2
• x 7→ x +1 −→ simple element of the 1st kind.
5
• x 7→ ( x −2)3
−→ simple element of the 1st kind.
1
• x 7→ ( x2 − x +1)
−→ simple element of the 2nd kind.
− x +8
• x 7→ ( x2 + x +1)4
−→ simple element of the 2nd kind.
x
• x 7→ 2x2 − x+5
−→ simple element of the 2nd kind.
1
• x 7→ ( x −2)2
−→ simple element of the 1st kind.
− x +2
• x 7→ (( x−1)2 +4)2
−→ simple element of the 2nd kind.

Properties 2.7(2)

P P
If Q is a rational fraction such that deg(P ) ≥ deg( Q), then Q can be written uniquely in the form

P( x) R( x)
= A( x) + , with deg( R) < deg( Q),
Q( x) Q( x)

where A and R are respectively the quotient and remainder of the Euclidean division of P by Q.

Example 2.7(11)

x3 x
2
= x− 2 , 1 = deg( x) < deg( x2 + 1) = 2.
x +1 x +1

Theorem 2.7(1)

P
Consider the rational function Q such that deg(P ) < deg( Q)

Q ( x ) = ( x − x 1 ) m1 ( x − x 2 ) m2 · · · ( x − x p ) m p ( x 2 + r 1 x + s 1 ) n1 ( x 2 + r 2 x + s 2 ) n2 · · · ( x 2 + r q x + s q ) n q

with r 2j − 4s j < 0, ∀ j = 1, · · · , q. Then, the fraction can be expressed in the form

P( x) A11 A12 A1m1

= + +···+
Q( x) x − x1 ( x − x1 ) 2 ( x − x 1 ) m1
A21 A22 A1m2
+ + +···+
x − x2 ( x − x2 )2 ( x − x 2 ) m2
A p1 A p2 A pm1
+···+ + +···+
x − xp ( x − x p )2 ( x − x p )m p
B11 x + C11 B12 x + C12 B1n x + C1n1
+ 2 + 2 +···+ 2 1
x − r1 x + s1 ( x − r1 x + s1 )2 ( x − r 1 x + s 1 ) n1
B21 x + C21 B22 x + C22 B1n x + C1n2
+ 2 + 2 +···+ 2 2
x − r2 x + s2 ( x − r2 x + s2 )2 ( x − r 2 x + s 2 ) n2
Bq1 x + Cq1 Bq2 x + Cq2 Bqnq x + Cqnq
+···+ 2 + 2 +···+ 2
x − rq x + sq ( x − rq x + sq ) 2 ( x − rq x + sq )nq
2.7. RULES FOR CALCULATIONS ON RIEMANN INTEGRABLE FUNCTIONS 43

where A, B, and C are real constants.

Example 2.7(12)

3x−2
The rational function x 7→ F ( x) = ( x −1)2 ( x +2)
can be expressed as

1.
3x − 2 A1 A2 A3
= + + .
( x − 1)2 ( x + 2) x−1 ( x − 1)2 x+2
To determine A1 , A2 , A3 , we have two methods.
First Method[By identification]

3x − 2 A1 ( x − 1)( x + 2) + A2 ( x + 2) + A3 ( x − 1)2
=
( x − 1)2 ( x + 2) ( x − 1)2 ( x + 2)
( A1 + A3 ) x2 + ( A1 + A2 − 2A3 ) x − 2A1 + 2A2 + A3
=
( x − 1)2 ( x + 2)

by identification, we will have

 
8


 A1 + A3 = 0 

 A1 = 9
⇔ 1
A1 + A2 − 2A3 = 3 A2 = 3
 
A3 = − 98
 
2A1 + 2A2 + A3 = −2
 

Second Method[The multiplicative method]

To calculate A2 , multiply both sides of the equation by ( x − 1)2 and then let x approach 1. We
obtain
3x − 2 A3 ( x − 1)2
= A1 ( x − 1) + A2 +
x+2 x+2
3x − 2 A3 ( x − 1)2 1
A2 = − A1 ( x − 1) − , x → 1 ⇒ A2 = .
x+2 x+2 3
To calculate A3 , multiply both sides of the equation by ( x + 2) and then let x approach (−2).
We obtain
3x − 2 A1 ( x + 2) A2 ( x + 2)
= + + A3
( x − 1)2 x−1 ( x − 1)2
3x − 2 A1 ( x + 2) A2 ( x + 2) −8
A3 = − − , x → −2 ⇒ A3 = .
( x − 1)2 x−1 ( x − 1)2 9
Then
3x − 2 A1 1 8
2
= + 2
−
( x − 1) ( x + 2) x−1 3( x − 1) 9( x + 2)
To calculate A1 , multiply both sides by x and then let x approach +∞. We obtain

(3x − 2) x
 
A1 x x 8x 8
lim = lim + − ⇒ A1 =
x→+∞ ( x − 1)2 ( x + 2) x→+∞ x−1 3( x − 1) 2 9( x + 2) 9

Thus
3x − 2 8 1 8
2
= + 2
− .
( x − 1) ( x + 2) 9( x − 1) 3( x − 1) 9( x + 2)
44 CHAPTER 2. RIEMANN INTEGRALS

2x2 +3
2. The rational function x 7→ F ( x) = ( x−1)( x2 +2x+2)
can be expressed as

2x2 + 3 A Bx + C
2
= + 2
( x − 1)( x + 2x + 2) x−1 x + 2x + 2

To determine A, B, and C, we have two methods.

First Method[By identification]
We reduce to a common denominator, and we get

2x2 + 3 A( x2 + 2x + 2) + ( Bx + C )( x − 1)
2
=
( x − 1)( x + 2x + 2) ( x − 1)( x2 + 2x + 2)
( A + B) x2 + (2A − B + C ) x + 2A − c
=
( x − 1)( x2 + 2x + 2)

by identification, we will have


 A+B = 2 
A=B=1

 
2A − B + C = 0 ⇔

  C = −1
2A − C = −3


Second Method[The multiplicative method]

To calculate A, multiply both sides of the equation by ( x − 1) and then let x approach 1. We
obtain
2x2 + 3 ( Bx + C )( x − 1)
2
= A+ ,
( x − 1)( x + 2x + 2) x2 + 2x + 2
Then
2x2 + 3 ( Bx + C )( x − 1)
A= − ,x→1 ⇒A=1
( x − 1)( x2 + 2x + 2) x2 + 2x + 2
To calculate B, multiply both sides of the equation by x and then let x approach ∞. We obtain

(2x2 + 3) x
 
x ( Bx + C ) x
lim = lim + 2 ⇒ B = 1.
x→+∞ ( x − 1)( x2 + 2x + 2) x→+∞ x−1 x + 2x + 2

then
(2x2 + 3) x 1 x+C
2
= + 2 .
( x − 1)( x + 2x + 2) x−1 x + 2x + 2
To calculate C, we replace x with 0. We obtain

3 C C 1
− = −1 + ⇒ = − ⇒ C = −1,
2 2 2 2
thus
2x2 + 3 1 x−1
2
= + 2 .
( x − 1)( x + 2x + 2) x−1 x + 2x + 2

Primitive of a Rational Fraction

Let P, Q, H, P1 , Q1 be five real polynomials.

2.7. RULES FOR CALCULATIONS ON RIEMANN INTEGRABLE FUNCTIONS 45

• A simple element of the 1st kind is of the form

A
u( x) = ,
( x − x0 ) m

Where A, x0 are two real numbers, and m ∈ N∗ . A primitive U of u is defined by


U ( x ) = A 1
I f m ̸= 1
1− m ( x − x0 ) m −1 ,
U ( x) = A ln | x − x |, If m = 1
0

• A simple element of the 2nd kind is of the form

Bx + C
v( x) = ,
( x2 + rx + s)n

where
√ B, C, r, and s are real numbers, n ∈ N∗ , and ∆ = r 2 − 4s < 0. The change of variable x =
−∆ r
2 t − 2 leads to the calculation of primitives of two types

t dt
Z Z
In = dt and Jn =
(1 + t 2 ) n (1 + t 2 ) n

we obtain immediately 
 In = − 1 1
, I f n > 1,
2( n −1) (1+ t 2 ) n −1
 I = 1 ln |1 + t 2 |, If n = 1
n 2

Jn is calculated by a recurrence formula (obtained by integrating by parts Jn )


 J1 = arctan t,
t
 2nJn+1 = (2n − 1) Jn + (1+ t 2 ) n
, ∀n ≥ 1.
46 CHAPTER 2. RIEMANN INTEGRALS

Example 2.7(13)

Calculate the following integral

3x − 2
Z
I= dx.
( x − 1)2 ( x + 2)
−2
The fraction F ( x) = ( x−3x
1)2 ( x +2)
is a proper fraction. To calculate its integral, we decompose it into
partial fractions. According to Example ..., we have

3x − 2 8 1 8
= + − ,
( x − 1)2 ( x + 2) 9( x − 1) 3( x − 1)2 9( x + 2)

Then
3x − 2 8 dx 1 dx 8 dx
Z Z Z Z
I= dx = + −
( x − 1)2 ( x + 2) 9 x−1 3 ( x − 1)2 9 x+2
8 1 8
= ln | x − 1| − − ln | x + 2| + C,
9 3( x − 1) 9

where C ∈ R. Therefore,

3x − 2 8 1 8
Z
dx = ln | x − 1| − − ln | x + 2| + C, where C ∈ R.
( x − 1)2 ( x + 2) 9 3( x − 1) 9

Example 2.7(14)

Calculate the following integral

2x2 + 3
Z
I= dx.
x3+ x2 − 2
It concerns the integral of a proper fraction, therefore using the method of decomposition into partial
fractions. We write x3 + x2 − 2 as a product of irreducible polynomials. We have

x3 + x2 − 2 = ( x − 1)( x2 + 2x + 2),

2x2 + 3 2x2 + 3
= .
x3 2
+x −2 ( x − 1)( x2 + 2x + 2)
According to Example ..., we have

2x2 + 3 1 x−1
= + 2 .
( x − 1)( x2 + 2x + 2) x−1 x + 2x + 2

Donc
2x2 + 3 1 x−1
Z Z Z
I= dx = dx + dx.
( x − 1)( x2 + 2x + 2) x−1 x2 + 2x + 2
We have
1
Z
dx = ln | x − 1| + C1 , C1 ∈ R.
x−1
and
x−1 1 2x + 2 − 4 1 2x + 2 dx
Z Z Z Z
2
dx = 2
dx = 2
dx − 2 .
x + 2x + 2 2 x + 2x + 2 2 x + 2x + 2 x2 + 2x + 2
we have,
2x + 2
Z
dx = ln | x2 + 2x + 2| + C2 , C2 ∈ R.
x2 + 2x + 2
2.7. RULES FOR CALCULATIONS ON RIEMANN INTEGRABLE FUNCTIONS 47

dx
R
We need to calculate x2 +2x+2
. We have

dx dx
Z Z
2
= .
x + 2x + 2 ( x + 1)2 + 1

Now, to calculate this integral, we set z = x + 1,

z = x + 1 ⇒ dz = dx.

dx dz
Z Z
=
( x + 1)2 + 1 z2 + 1
= arctan z + C3
= arctan( x + 1) + C3 , C3 ∈ R.

Therefore,

2x2 + 3 1
Z
dx = ln | x − 1| + ln | x2 + 2x + 2| − 2 arctan( x + 1) + C, where C ∈ R.
x3 + x2 − 2 2

Example 2.7(15)

Calculate the following integral

x3 + 1
Z
I= dx.
x2 + 1
It is the integral of an improper fraction, so using Euclidean division, we bring it to the integral of a
polynomial and another proper fraction. We have

x3 + 1 −x + 1 x 1
= x+ 2 = x− 2 + 2 ,
x2 + 1 x +1 x +1 x +1

then

x3 + 1 1 2x 1
Z Z Z Z
dx = xdx − dx + dx,
x2 + 1 2 x2 + 1 x2 + 1
1 2 1
= x − ln( x2 + 1) + arctan x + C, C ∈ R.
2 2
Therefore,
x3 + 1 1 1
Z
I= 2
dx = x2 − ln( x2 + 1) + arctan x + C, C ∈ R.
x +1 2 2