National Polytechnic School

Preparatory Cycle - Algiers (El-Harrach)

Analysis 02 : Continuous Test (CC) 01
Date : 07 March 2024. Duration : 01h00.

Exercise 1
Z Z
2
Given the following two integrals I1 = x cos x dx and I2 = x sin2 x dx

1. Evaluate I1 + I2 and I1 − I2 , then deduce I1 and I2 .

2. Evaluate directly the two integrals I1 and I2

Exercise 2
Z π/4
The aim of this exercise is to evaluate the following definite integrale I = ln(1 + tan(x)) dx.
0
Z b Z b/2  
1. Prove by substitution that if f (x) is integrable on[0, b] f (x) dx = f (x) + f (b − x) dx.
0 0
tan(α) + tan(β)
2. Using the property proved in(1) and the identity tan(α + β) = , evaluate I.
1 − tan(α) tan(β)

Exercise 3
3x2 + 7x − 2
1. Find the partial fraction decomposition for f (x) = 3 .
x − x2 − 2x
Z
2. Evaluate the integral f (x) dx.

Exercise 4
Z b
The aim in this exercise is to try to evaluate J = cos x dx, using the Riemann sum with a partition
a
of [a, b] and using the right or the left point approximation.
n (2n+1)θ
X 1 sin( 2 )
1. Give a proof of the Lagrange Trigonometric Identity (LTI) : cos kθ = + θ
.
k=0
2 2 sin( 2
)
a 2a 3a
2. Derive the Riemann partial sum of J using the partition {0, , , ...., a} and the LTI.
n n n
Z a Z b
3. By letting n → ∞ in the Riemann sum, evaluate cos x dx and then deduce cos x dx.
0 a

1
 T ypi c
al Soluti on Wit h D et
ails

Solution of Exercise 1

1. Starting from the two integrals  Z

 I1 = x cos2 x dx,

Z
 I2 = x sin2 x dx.


We get two equations by adding and subtracting the two integrals,

 Z   Z
2 2
 I1 + I2 = x cos + sin dx = x dx,


Z   Z  
2 2 2 2
 I1 − I2 = x cos − sin dx = x cos − sin dx.



(0.25+0.25) pt/5

Integrating the first easily and the second by parts, we obtain two new equations,
1

 I1 + I2 = x2 + C1 ,

Z2   Z
x 1 x 1
 I1 − I2 = x cos 2x dx = sin 2x −
 sin 2x dx = sin 2x + cos 2x + C2 ,
2 2 2 4
(0.5+0.5) pt/5

from which we extract the two solutions I1,2 ,

2

 2I1 = x + x sin 2x + 1 cos 2x + C3 ,

22 2 4
x x 1
 2I2 = − sin 2x − cos 2x + C4 ,

2 2 4
(0.25+0.25) pt/5

and finally,
2

 I1 = x +
 x 1
sin 2x + cos 2x + C3 ,
42 4 8
x x 1
 I2 = − sin 2x − cos 2x + C4 .

4 4 8
(0.5+0.5) pt/5
Z Z
2. If we try to calculate the two integrals I1 = x cos2 x dx and I2 = x sin2 x dx directly, we can

first linearise the cos2 x, sin2 x terms using the trigonometric relations :

 cos2 x = 1 + cos 2x ,

2
2
 sin x = 1 − cos 2x
,

2

2 Prof. Dr. Madjid Lakhdar Hamou LADREM (ENP,2024)

 (0.25+0.25) pt/5

 Z Z   Z Z  
2 1 + cos 2x 1 1
 I1 = x cos x dx = x dx = x dx + x cos 2x dx,


Z  2  2Z 2Z  
−
Z
2 1 cos 2x 1 1
 I2 = x sin x dx = x dx = x dx − x cos 2x dx.


2 2 2
(0.25+0.25) pt/5

So the final expressions of I1,2 , after simply integrating the first term and integrating by parts of

the second, can be written as,

2

 I1 = x +
 x 1
sin 2x + cos 2x + C3 ,
42 4 8
 I2 = x −
 x 1
sin 2x − cos 2x + C4 ,
4 4 8
(0.5+0.5) pt/5

which are the same as the ones obtained in the first question above.

Solution of Exercise 2
π/4
The aim of this exercise is to evaluate the following definite integral I = int0 ln(1 + tan(x)) dx, using

a very specific approach.

Z b
1. In first we should to prove by substitution that if f (x) is integrable on[0, b] f (x) dx =
Z b/2   0

f (x) + f (b − x) dx. We will start by rewriting the first member of the equation using the
0
CHASLES relation,
Z b Z b/2 Z b
f (x) dx = f (x) dx + f (x) dx. (1)
0 0 b/2

(0.5) pt/5

After making the following substitution,


 y = b − x,


 dy = −dx.


(0.5) pt/5

3 Prof. Dr. Madjid Lakhdar Hamou LADREM (ENP,2024)

 we obtain the required result,
Z b Z b/2 Z 0
f (x) dx = f (x) dx + f (b − y) (−dy) =
0 0 b/2
Z b/2 Z b/2 Z b/2  
f (x) dx + f (b − y) dy = f (x) + f (b − x) dx.
0 0 0

(2)

(1.) pt/5

Finally we have the relation

Z b Z b/2  
f (x) dx = f (x) + f (b − x) dx. (3)
0 0

tan(α) + tan(β)
2. Using the property proved in(1) and the identity tan(α+β) = , we will evaluate
1 − tan(α) tan(β)
I.

Z π/4 Z π/8  
π
I= ln(1 + tan x) dx = ln(1 + tan x) + ln(1 + tan( − x)) dx =
0 0 4
Z π/8   Z π/8  
1 − tan x 2
ln(1 + tan x) + ln(1 + ) dx = ln(1 + tan x) + ln( ) dx =
0 1 + tan x 0 1 + tan x
Z π/8  
π ln 2
ln(1 + tan x) + ln 2 − ln(1 + tan x) dx = . (4)
0 8

(1.+1.+1.) pt/5

The numerical value of the definite integral defined in this way is given by :
Z π/4
π ln 2
I= ln(1 + tan(x)) dx = . (5)
0 8

Solution of Exercise 3

3x2 + 7x − 2
1. The partial fraction decomposition for f (x) = , is done in four steps.
x3 − x2 − 2x
(a) Step 1 : Factoring the denominator, we find that x3 − x2 − 2x = x(x + 1)(x − 2) in which a

simple linear factor appear.

(0.5) pt/5

4 Prof. Dr. Madjid Lakhdar Hamou LADREM (ENP,2024)

 (b) Step 2 : The partial fraction decomposition has one term for each factor in the denominator

3x2 + 7x − 2 3x2 + 7x − 2 A B C
f (x) = 3 2
= = + + . (6)
x − x − 2x x(x + 1)(x − 2) x x+1 x−2

(0.25) pt/5

The goal is to find the undetermined coefficients A, B, and C.

(c) Step 3 : A reduction to the same denominator is made on the second member of the equation

above lead to,

3x2 + 7x − 2 = (A + B + C)x2 + (−A − 2B + C)x − 2A. (7)

(0.5) pt/5

(d) Step 4 : We now equate coefficients of the same order on both sides of the equation in Step 3.




 A + B + C = 3,

 −A − 2B + C = 7,


 −2A = −2.


(0.25+0.25+0.25) pt/5

Solving for A, B and C, we conclude that





 A = 1,

 B = −2,



 C = 4.

(0.25+0.25+0.25) pt/5

Substituting the values of A, B, and C into equation (3), the partial fraction decomposition is

1 2 4
f (x) = − + . (8)
x x+1 x−2

(0.25) pt/5

2. Integration of f (x) is now straightforward :

Z Z   Z   Z   Z  
1 2 4 1 2 4
f (x) dx = − + dx = dx − dx + dx. (9)
x x+1 x−2 x x+1 x−2

5 Prof. Dr. Madjid Lakhdar Hamou LADREM (ENP,2024)

 (0.5) pt/5

Simple integration of the three terms will only generate logarithmic functions,

|x|(x − 2)4
Z
f (x) dx = ln|x| − 2ln|x + 1| + 4ln|x − 2| + C = ln + C. (10)
(x + 1)2

(0.5+0.5+0.5) pt/5

Solution of Exercise 4
Z b
The aim of this exercise is to try to evaluate cos x dx, using the Riemann sum with a partition of [a, b]
a
and the right or left point approximation. The calculation is simpler than it looks. Since the function

cos x is continuous, it is integrable in the Riemann sense. The definite integral of cos x from a to b is

exactly given by the infinite limit of the Riemann partial sum,

Z b n
X
J(a, b) = cos x dx = lim (cos xk ) ∆x, (11)
a n→∞
k=0

(0.5) pt/5

b−a
where ∆x = and xk = a + k∆x.
n

(0.5) pt/5

n (2n+1)θ
1 sin( 2 )
X
1. First, we will prove Lagrange’s Trigonometric Identity (LTI) : + cos kθ =
θ
. This
k=0
2 2 sin( 2
)
identity will be very useful later. We start with the complex form of cos kθ. This is given by the

well-known relation,
ejkθ + e−jkθ
cos kθ = , (12)
2

(0.5) pt/5

which is inserted into the sum,

n n  n n
ejkθ + e−jkθ

X X 1 X jkθ X −jkθ
cos kθ = = e + e
k=0 k=0
2 2 k=0 k=0
1 ej(n+1)θ − 1 e−j(n+1)θ − 1
 
= + . (13)
2 ejθ − 1 e−jθ − 1

6 Prof. Dr. Madjid Lakhdar Hamou LADREM (ENP,2024)

 The two sums in the above relationship are nothing more than partial sums of the geometric

Riemann series, whose general form is given by :

n
X rn+1 − 1
ark = a . (14)
k=0
r−1

where a is the coefficient of each term and r the common ratio between adjacent terms. After a

slight calculation, we obtain

n 1 θ 1 θ 
1 ej(n+ 2 )θ − e−j 2 e−j(n+ 2 )θ − e+j 2
X 
cos kθ = θ θ − θ θ =
k=0
2 ej 2 − e−j 2 e−j 2 − ej 2
1 1 θ θ 
1 ej(n+ 2 )θ − e−j(n+ 2 )θ + ej 2 − e−j 2

θ θ =
2 ej 2 − e−j 2
1 1 θ θ 
1 ej(n+ 2 )θ − e−j(n+ 2 )θ + ej 2 − e−j 2

θ θ =
2 ej 2 − e−j 2
1 1
1 ej(n+ 2 )θ − e−j(n+ 2 )θ 1 sin(n + 12 )θ
   
θ θ +1 = + 1 . (15)
2 ej 2 − e−j 2 2 sin( 2θ )

The result obtained is simply the LTI

n
1 sin(2n + 1) 2θ
 
X 1
cos kθ = θ
+ . (16)
k=0
2 sin( 2 ) 2

(0.5) pt/5
Z a
a 2a 3a
2. Writing the J(0, a) = cos x dx as a Riemann sum using a partition {0, , , ...., a} is
0 n n n
obvious. For the Left Point Approximation (LPA), we have,
Z a n−1
X
J(0, a) = cos x dx = lim (cos xk ) ∆x, (17)
0 n→∞
k=0

and with the Right Point Approximation (RPA) we have,

Z a n
X
J(0, a) = cos x dx = lim (cos xk ) ∆x. (18)
0 n→∞
k=1

(0.5) pt/5

a
With the particular case of xk = kαn and ∆x = = αn . Using LTI (with θ = αn ) the partial
n
sum of the Riemann series becomes
n−1 n−1
αn sin(2n − 1) α2n
 
X X αn
(cos xk ) ∆x = αn (cos kαn ) = αn + . (19)
k=0 k=0
2 sin( 2
) 2

(0.5) pt/5

7 Prof. Dr. Madjid Lakhdar Hamou LADREM (ENP,2024)

 Z a
3. By letting n → ∞ in the partial Riemann sum to evaluate cos x dx
0

n−1 n−1
( )
a
αn sin(2n − 1) α2n
Z  
X X αn
J(0, a) = cos x dx = lim (cos xk ) ∆x = lim (cos kαn ) = lim + .
0 n→∞
k=0
n→∞
k=0
n→∞ 2 sin( α2n ) 2
(20)

With the well-known results of two limits,


 lim αn = 0

n→∞
sin αn
 lim
 = 1,
n→∞ αn

(0.5+0.5) pt/5

and, after the calculation, we get the long awaited final result
(  )
αn sin(2n − 1) α2n
Z a 
αn
J(0, a) = cos x dx = lim + = sin(a), (21)
0 n→∞ 2 sin( α2n ) 2

(0.5) pt/5
Z b
Finally, we can evaluate the most general definite integral, J(a, b) = cos x dx using the CHASLES
a
relation :
Z b Z 0 Z b
cos x dx = cos x dx + cos x dx = 0 = J(0, b) − J(0, a) = sin(b) − sin(a). (22)
a a 0

(0.5) pt/5

N.B.

The difference between the two approximations (LPA and RPA) is illustrated on the last page by the

plots (FIGURE 1,2,3,4), which are the real graphs produced by the MATHEMATICA software. The

first two (FIGURE 1 and FIGURE 2) correspond to the case of n = 10, however, the case of n = 50 is

illustrated in the figures (FIGURE 3 and FIGURE 4). They show a tendency to be identical, especially

when the infinite limit is reached, which is reminiscent of the property of integrability in the Riemann

sense, as explained with details in the course.

8 Prof. Dr. Madjid Lakhdar Hamou LADREM (ENP,2024)

 1.0

0.8

0.6

0.4

0.2

0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

Figure 1 – LPA in Riemann Sum of cos x in Figure 2 – RPA in Riemann Sum of cos x in

[0, π/2] with n = 10. [0, π/2] with n = 10.

1.0

0.8

0.6

0.4

0.2

0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

Figure 3 – LPA in Riemann Sum of cos x in Figure 4 – RPA in Riemann Sum of cos x in

[0, π/2] with n = 50. [0, π/2] with n = 50.

9 Prof. Dr. Madjid Lakhdar Hamou LADREM (ENP,2024)