Lecture 7: Integration Techniques

Antiderivatives and Indefinite Integrals

1. In differential calculus, we were interested in the derivative of a given real-valued function,

whether it was algebraic, exponential or logarithmic. Here we are concerned with the inverse of

the operation of differentiation. That is, the operation of searching for functions where

derivatives are a given function.

2. Consider any arbitrary real-valued function

f: X → R

defined on a subset x of the real line, i.e. X ⊂ R. By the antiderivative (primitive) of f; we mean

any differentiable function

F: X → R

whose derivative is the given function f. Hence,

dF
= F′(x) = f(x).
dx

3. Let c be any real number c ∈ R, (which is constant). Now since the derivative of c with

respect to x or dc/dx = 0, we have:

Theorem 1. If a function F: X → R is an antiderivative of f: X → R, so is the function F + c: X

→ R defined by

(F + c) (x) = F(x) + c,

for every x ∈ X ⊂ R.

Remark 1: Hence, the derived function f(x) is traceable back to an infinite number of possible

primary, antiderivative or primitive functions of the form F(x) + c.

Remark 2: However, here there is a subtle point. Every antiderivative fn of f is of the form F(x)

+ c if X ⊂ R is a connected subset of the real line R.

1
Def 1: A set X ⊂ R is connected if and only if, for any two real numbers a, b ∈ X with a < b, X

contains the closed interval [a, b].

Example: Consider the function f(x) = x2, where f: X → R, X = [0, ∞). It has an antiderivative

function defined by

1 3
F(x) = x + 10,
3

since F′(x) = dF/dx = x2. However, it also has an antiderivative function defined by

G(x) = 1/3 x3 + 100,

dG ( x) 2
since = x . In fact, any function of the form
dx

1 2
F(x) = x + c,
3

for any c ∈ R is an antiderivative function of f(x) = x2.

Theorem 2. If F: X → R is an antiderivative function of a real-valued function f: X → R, where

X ⊂ R is connected, then {F + c: c ∈ R} is the set of all antiderivatives of f.

Def. 2. The set of all antiderivative functions of a real valued function f: X → R, X ⊂ R, is

called the indefinite integral of f and is denoted by

∫ f ( x)dx.
Remark 3: The symbol ∫ is the integral sign. If f has no antiderivative (nonintegrable), then its

indefinite integral is ∅. The f(x) part is the integrand. The f(x) dx may be taken as the

differential dF of a primary or antiderivative function.

Remark 4: The process of determining the indefinite integral of a given real-valued function say

f: X → R is termed the indefinite integration of f. We have that

d
dx
F(x) = f(x) implies ∫ f ( x)dx = F( x) + c

2
for any real c, where X connected. The constant x will be referred to as the arbitrary constant of

integration. If X is not connected, then

∫ f ( x)dx = F + c

means only that F + c is an antiderivative of f for some x.

Counterexample.

Let x denote the subset of R which consists of all non-zero real numbers:

X = {x ∈ R: x ≠ 0}

(1) Consider the fn f(x) = 1/x2, f: X → R.

Then the fn. F(x)

F(x) = -x-1, F: X → R

is an antiderivative of f.

(2) Consider the function G(x), G: X → R,

− x −1 + 1 if x>0
G( x)  − 1
− x − 1 if x<0

Note dG/dx = G′(x) is given by:

G′(x) = 1/x2 = f(x).

∴ G(x) is also an antiderivative of f.

However,

1 1
G-F=- + 1 + ; if x > 0
x x

 1  1
G - F = − − 1 + ; if x < 0
 x  x

Hence

+ 1 if x>0
G-F= 
− 1 if x<0

3
Thus, G - F ≠ c, where c represents a constant function, for all x. Hence, G is not of form

F(x) + c, for all x.

Basic Rules of Integration

1. Theorem 1. If F is an antiderivative of f and G is an antiderivative of g, then F + G is an

antiderivative of f + g. Hence,

∫ [f(x) + g(x)] dx = ∫ f(x) dx + ∫ g(x) dx.

(holds for finite sum)

2. Theorem 2. If F is an antiderivative of f, then cF is an antiderivative of cf for every c ∈ R.

Thus,

∫ cf(x) dx = c∫ f(x) dx.

3. Theorem 3. For every N ≠ 1, xN+1/N + 1 is the antiderivative of xN. Thus,

1
∫ xN dx = XN+1 + c.
N +1

4. Examples:

#1 Let f(x) = x4, then

x5
∫ x dx =
4
+ c.
5

check:

d (x5 / 5)
= x4.
dx

#2 Let f(x) = x3 + 5x4, then

∫ [x3 + 5x4] dx = ∫ x3 dx + 5∫ x4 dx

= x4/4 + 5(x5/5) + c

= x4/4 + x5 + c

check.

4
 d 4
(x /4 + x5) = x3 + 5x4.
dx

#3 Let f(x) = x2 - 2x

∫ [x2 - 2x] dx = ∫ x2 dx + ∫ -2x dx

= x3/3 - 2∫ x dx + c′

= x3/3 - 2(x2/2) + c

= x3/3 - x2 + c.

5. Theorem 4. The antiderivative of ex is given by

∫ ex dx = ex + c.

Theorem 5. The antiderivative of x-1, x ≠ 0 is given by

1
∫ dx = ln|x| + c.
x

Remark 1: (Why ln|x| instead of ln x.) Note that the function f(x) = x-1 is such that

f: R - {0} → R.

That is, f(x) = x-1 is defined on the set of non-zero real numbers. Now, the natural logarithmic

function

G(x) = ln x, is such that G: P → R, where P ≡ {x: x ∈ R, x > 0}. That is, ln x is defined only on

the positive real numbers. Thus, G(x) = ln x, G: P → R, can not be an antiderivative function of

f(x) = x-1, where f: R - {0} → R. This is true since the antiderivative of f must map from the

same domain. In this case, clearly, R- {0} ≠ P. To be able to define an antiderivative function of

x-1, we must extend the function ln x to the set R - {0} somehow. Thus, we define the function

F(x) = ln(|x|), where, of course F: R-{0} → R. That is, ln (|x|) is defined on all non-zero real

numbers.

Example: Let f(x) = 2ex - x-1.

∫ f(x) dx = 2∫ex dx - ∫ x-1 dx = 2ex - ln|x| + c.

5
The Methods of Integration by Substitution and by Parts.

1. Method of Substitution: To determine the indefinite integral of a function f(x), we choose by

inspection a suitable differentiable fn u(x) such that the given function f(x) can be expressed as

the product g(u(x)) ⋅ u′(x), where g(u(x)) is a function of u and constants. Then we have

∫ f(x) dx = ∫ g(u) u′(x) dx = ∫ g(u) du.

Remark : We have that u = u(x); hence, du = u′(x) dx.

2. Examples

x
#1 Let f(x) = , find ∫ f(x) dx.
(x 2
+ 1)
1/ 2

Choose u = u(x) = x2 + 1, then we have that u′(x) = 2x, such that

du = u′(x) dx = 2xdx

Now

f(x) = g(u) u′(x)

x x 1 1
f(x) = = 1/ 2 = u′ ( x)
(x 2
+ 1)
1/ 2
u 2 u1 / 2

Here, g(u) = (1/2)(u-1/2). Thus

x 1 1
∫ f(x) dx = ∫ dx = ∫ u′ ( x) dx
(x 2
+ 1)
1/ 2
2 u1/ 2

but du = u′(x) dx, hence

1 1
∫ f(x) dx = ∫ g(u) du = ∫ du.
2 u1/ 2

Now integrate the last result.

∫ f(x) dx = 1
2 ∫u-1/2du

6
  1  1/2
 1 u +c
1
= 2
 − 2 + 1

 1  1/2
 1 u + c
1
= 2
 2

= u1/2 + c

Substitute for u = x2 + 1, then

x
∫ dx = (x2+1)1/2 + c.
(x 2
+ 1)
1/ 2

To check, find

d 2
(x +1)1/2 = ½ (x2+1)-1/2 2x
dx

x
=
( x + 1)1/ 2
2

2 x3 − x
#2 Let f(x) = . Find ∫f dx.
(x4 − x2 + 1)1/ 3
(1) Choose u = (x4 - x2 + 1)

∴ u′(x) = 4x3 - 2x, such that

du = (4x3 - 2x) dx.

(2) Construct the product g(u) u′(x) such that f(x) = g(u) u′(x)

2 x3 − x 2 x3 − x
f(x) = =
(x4 − x2 + 1)1/ 3 u1/ 3

( 12 )u′( x)
= 1
u3

u ′ ( x)
= 1
2u 3

1
Here g(u) = 1
2u 3

7
 (3) Thus we have that

∫ f(x) dx = ∫ g(u) u′(x) dx

1 1
=∫ 1 u′(x) dx = ∫ 1 3 du
2u 3
2u

Take last integral:

1 -1/3
∫ f(x) dx = ∫ u du
2

1  1  23
=  u + c
2  23

3 23
= u +c
4

(
3 4
x − x2 + 1) + c
2
3
=
4

To check take d/dx

1 4 2
= (x - x + 1)-1/3 (4x3 - 2x)
2

4 x3 − 2 x
=
2( x4 − x2 + 1)
1/ 3

x3 − x
=
(x4 − x2 + 1)1/ 3
2. Method of Integration by Parts. To determine the indefinite integral of the fn f(x), we choose

by inspection two differentiable functions u(x), v(x) such that

∫ f(x) dx = ∫ u dv = uv - ∫ v du.

Remark. If we can find u and v such that f(x) = u v′(x), then dv = v′(x) dx and

∫ f(x) dx = ∫ u v′(x) dx = ∫ u dv.

To see that ∫ u dv = uv - ∫v du, let z ≡ vu, then

dz = u dv + v du

8
Integrate both sides

∫ dz = ∫ u dv + ∫ v du

z = ∫ u dv + ∫ v du

∫ u dv = z - ∫ u du

∴ ∫ u du = uv - ∫ v du.

Examples.

#1 Let f(x) = xeax

(1) Choose dv so that it is the most complicated expression,

but is easy to integrate

Let dv = eax dx

Let u = x such that du = dx

1
(2) ∫ u dv = x eax - ∫ v du
a

Now since

dv = eax dx

∫ dv = v = ∫ eax dx

1 ax
v = e
a

Also since

u =x

du = 1 dx

x ax
∫ u dv = e - ∫ v du
a

x ax 1
= e - ∫ eax dx
a a

9
 x ax 1 ax
= e - ∫ e dx
a a

x ax 1 ax
= e - 2 e +c
a a

x ax 1 ax
∴ ∫ x eax dx = e - 2 e + c.
a a

check

d  x ax 1 ax  1 ax 1
 e − 2 e  = e + xeax − eax = xeax
dx  a a  a a

2
+2
#2 Let f(x) = 6xe x . Integrate by parts.

2
+2
(1) Let v = e x

2
+2
then dv = 2x e x dx

Hence, let

u =3

du = 0 dx = 0

(2) Then ∫ f(x) dx = ∫ u dv

2
+2 2
+2
= 3 ex - ∫ ex du = 0

Hence,

2
+2 2
+2
∫ 6x e x dx = 3 e x +c

To check:

d
(3 e x + 2 ) = 6x e x + 2 .
2 2

dx

Definite Integrals

1. Let f(x) be continuous on an interval X ⊂ R, where f: X → R. Let F(x) be an antiderivative of

f, then ∫ f(x) dx = F(x) + c.

10
2. Now choose a, b ∈ X such that a < b. Form the difference

[F(b) + c] - [F(a) + c] = F(b) - F(a).

(note indep. of c)

This difference F(b) - F(a) is called the definite integral of f from a to b. The point a is termed

the lower limit of integration and the point b, the upper limit of integration.

3. Notation: We would write

∫ f ( x)dx ≡ F( x)]a ≡ F( x) a = F( b) − F( a )
b b b
a

4. All that we have described thus far is known as the fundamental theorem of calculus. We

state these results formally as follows:

Theorem 1: For any antiderivative function F(x), F: X → R, X ⊂ R, of a continuous function

f(x), f: X → R defined on a closed interval I = [a, b], we have

∫ f ( x)dx = F(b) - F(a).

b

Examples

#1 Let f(x) = x3, find

1 1 4 1 4
∫ x dx = 4 x = (1) − ( 0)
1
3 41
0
0 4 4

1
= .
4

∫ f ( x)dx ,
2 5
#2 f(x) = 2x e x , find
3

∫ 2xe
5 5
x2
= ex
2
= e25 - e9
3 3

5. The definite integral represents the area under f(x) between the points a and b.

6. Properties of Definite Integrals.

∫ f ( x) dx over [a, b], then

b
P. 1 If f(x) is such that ∃
a

11
 ∫ f ( x)dx = − ∫ f ( x)dx .
b a

a b

P. 2 If f(x) is defined and continuous at the point a, then

∫ f ( x)dx = 0 .
a

P. 3 If f(x) is defined and continuous on each of the closed intervals [x1, x2], …, [xN,

xN+1], where N, the number of subintervals, is finite and

U[x , x ] = [x , x ] , then
i
i i +1 1 N +1

∫ f ( x)dx = ∫ f ( x)dx + ∫ f ( x)dx+K+ ∫ f ( x)dx .

x N +1 x2 x3 x N +1

x1 x1 x2 xN

∫ f ( x) dx and ∃ ∫ g( x) dx, then

b b
P. 4 If f(x) and g(x) are such that ∃
a a

∫ kf ( x)dx = k ∫ f ( x)dx , for any k ∈ R

b b
(i)
a a

∫ [ f ( x) + g( x)]dx = ∫ f ( x)dx + ∫ g( x)dx .

b b b
(ii)
a a a

Improper Integrals

∫ f ( x)dx . If a or b or both are infinite or if f(x) is

b
1. Consider a fn. f(x) and the definite integral
a

undefined for some x ∈ [a, b], the above expression is termed improper.

∞
Def 1. If f(x) is defined for x ∈ [a, +∞), then the expression ∫ a
f ( x) dx is defined as

lim ∫ f ( x) dx. If f(x) is defined for x ∈ (-∞, b], then ∫ f ( x) dx is defined as lim + ∫ f ( x) dx.
b b b

b →∞ a −∞ a → −∞ a

+∞
Def 2. If f(x) is defined for x ∈ [a, b] then the expression ∫ −∞
f ( x) dx is defined as

lim − ∫a f (x) dx.

b

b→+∞
a →−∞ +

12
 ∫ f ( x)dx is
b
Def 3. If f(x) is defined for x ∈ [a, b), b ∈ R, but not defined for x = b, then
a

∫ f ( x)dx .
c
defined as lim− If f(x) is defined for x ∈ (a, b], a ∈ R, but not defined for x = a, then
c→ b a

∫ f ( x)dx is defined as lim+ ∫ f (x)dx .

b b
the expression
a c→ a c

Def 4. If the limit of the improper integrals called for in Def. 1, 2 or 3 exists, the improper

integral is said to be convergent, otherwise divergent.

2. Examples:

∞
#1 Evaluate ∫
2
x − 7 dx . Hence, by Definition 1,

∞
∫ x − 7 dx = lim− ∫ x − 7 dx
b

2 b →∞ 2

 1 b   1 1 1 1
= lim− − x − 6 2  = lim− − 6 +
b →∞  6  b→∞  6 b 6 26 

1  −6
=  2 − lim b− 6 
6 b →∞ 

1 −6
= 2
6

1
∴ our integral is convergent and =
( 6)26
1 1
∫x ∫x
−2 −2
#2 Evaluate dx = lim+ dx
0 a →0 a

a→0 a a→0
1
[ 
a→0
1
lim ∫ x2 dx = lim − x −1 a = lim− 1 + 
1

a]
1 
lim − 1 = ∞
a → 0 a 

Hence our integral is divergent and has no value.

+1
#3 Do #2 for ∫ −1
x − 2 dx

13
 +1 0 1
∫ x −2 dx = ∫x ∫x
−2 −2
dx + dx
−1 −1 0

but we found above that

1 
lim ∫ x − 2 dx = lim − 1 = ∞
1

a→0 a a → 0 a 

∴ it diverges.

#4 Let f(x) be a continuous probability density function with

support (domain) (-∞,+∞). We know that the following improper

integral is convergent.

+∞

∫ f (x)dx = 1 .
−∞

Both the expectation and the variance of the random variable x

are convergent improper integrals.

+∞
E(x) = ∫ xf (x )dx,
−∞

+∞

∫ ( x − E( x )) f ( x)dx.
2
Var(x) =
−∞

Likewise, the distribution function of x is a convergent improper

integral. This function is defined by

x'
F(x') = ∫ f (x)dx = Pr ob(x ≤ x' ).
−∞

Differentiation of an Integral

1. The following rule applies

q( y) q
∂
∫ ∫
f ( x , y)dx = f y ( x , y)dx + f (q, y)q ' ( y) − f (p, y)p' ( y) .
∂y p ( y ) p

14
2. Example: In Economics, we study a consumer's demand function in inverse form

p = p(Q),

where Q is quantitiy demanded and p denotes the maximum uniform price that the consumer is

willing to pay for a given quantity level Q. We assume that p' is negative. The definite integral

∫ p(z)dz = TV(Q)
0

is called total value at Q. It gives us the maximum revenue that could be extracted from the

consumer for Q units of the product. The dollar amount

TV(Q) – p(Q)Q = CS(Q)

is called consumer's surplus. Let C(Q) be the cost of supplying Q units. If a firm could extract

maximum revenue from the consumer, its profit function would be

∫ p(z)dz − C(Q).
0

The output level maximizing the firm's profit sets the derivative of the previous definite integral

equal to zero. This implies

p(Q) = C'(Q).

Some Notes on Multiple Integrals

1. In this section, we will consider the integration of functions of more than one independent

variable. The technique is analogous to that of partial differentiation. When performing

integration with respect to one variable, other variables are treated as constants. Consider the

following example:

d b

∫∫ f (x, y)dxdy .
c a

15
We read the integral operators from the inside out. The bounds a,b refer to those on x, while the

bounds c,d refer to y. Likewise, dx appears first and dy appears second. The integral is

computed in two steps:

∫
#1. Compute f ( x , y)dx = g ( y).
a

d d b

∫
#2. Compute g ( y)dy =
c
∫∫ f (x, y)dxdy .
c a

If there were n variables, you would follow the same recursive steps n times. Each successive

integration eliminates a single independent variable.

2. Some Examples.

Example 1: Suppose that z = f(x,y). We wish to compute integrals of the form

d b

∫∫ f (x, y)dxdy.
c a

Consider the example f = x2y, where c = a = 0 and d = 2, b = 1. We have

2 1

∫∫ x
2
ydxdy.
0 0

Begin by integrating with respect to x, treating y as a constant

1
1 1
∫ yx dx = 3 yx |0 =
2 3 1
y.
0
3

Next, we integrate the latter expression with respect to y.

2
1 11 1 2
∫ 3 ydy = 3 2 y |0 = 4= .
2 2

0
6 3

Example 2: Compute

∫∫∫(3xy + 2z)dxdydz,
where the limits of integration are 0 and 1, in each case. Begin with x.

16
 3x 2 y 3y
( + 2zx ) |10 = + 2z.
2 2

Next, y

3y 2 3
( + 2zy) |10 = + 2z.
4 4

Finally, z

3z 7
( + z 2 ) |10 = .
4 4

Example 3. Let f(x,y) be a joint probability density function defined on the continuous random

variables (x,y) with support R2. The unconditional expectation of x is defined as the double

integral

+∞ +∞
E(x) = ∫ ∫
−∞ −∞
xf ( x , y)dydx .

This expectation can be rewritten in terms of the marginal probability density function for x

+∞
f1(x) = ∫
−∞
f ( x , y)dy .

Thus,

+∞
E(x) = ∫ xf (x)dx .
−∞
1

The conditional expectation of x for a given value of y is quite different. The conditional density

of x, given y, is

g1(x | y) = f(x,y)/f2(y),

and the conditional expectation is

+∞
E(x | y) = ∫ xg (x | y)dx .
−∞
1

17