Mathematics 1

08 - Antiderivatives

Luciano Battaia

[email protected]

Curriculum Business Administration and Management

A.Y.2024/25
Integration as inverse of differentiation

Indefinite integral
Given a function f : D ! R, any differentiable function F : D ! R such that
F 0 (x ) = f (x ) for all x 2 D is called an antiderivative of f .
-

If F is an antiderivative, also F (x ) + c , where c is an arbitrary real constant, is an

antiderivative of f .
We call indefinite integral the set of all antiderivatives:
Z
f (x ) dx = {F : D ! R : F 0 (x ) = f (x )}. St
O
If, moreover, D is an interval, the set of all antiderivatives is simply {F (x ) + c}, where

this case we write: Z

#
c is an arbitrary constant: this means that two antiderivatives differ by a constant. In

f (x ) dx = F (x ) + c. FWO ANTID .

DIFFER BY
and c is called the additive constant of integration. A consi
L.Battaia 08 - Antiderivatives A.Y.2024/25 1 / 15
 0)v(0 y)
f(x) t ( +
D x,
-

= ,
=

AN INTERNA
DISNOT

⑰
0
en (x) X =

⑫(lEXT5
e

430

X30

fax-fax =

20
Luciano Battaia
 To
IT IS POSSIBLE

COMPUTE THE DERIVATIVE

AT A Pant
Of A FCT

# Of Antiperivate
THE IDEA

MARES NO SENSE AT A POINT

Luciano Battaia
 ALWAYS
ANTI DERIVATIVE IS
IM A SET /NORMALLY in ,

in INTERVAL
an exercises) on

Luciano Battaia
 MEANS
INTEGRATION
TOPICS
3 Diff -

raTIVES
15 : ANTIDERI
(techindly very difficult

Ind :
RIEMANN INTEGRALS
CASEAS)
Itechnically simple (
32d : fats defined by in te got
Luciano Battaia
Functions not defined in an interval - Remarks
-

Functions not defined in an interval can have antiderivatives not differing by a constant. For example
1
consider the function f (x ) = x , whose domain is R \ {0}. The two functions
ß ß
ln( x ) + 1, if x < 0; ln( x ) + 3, if x < 0;
F 1 (x ) = ; F 2 (x ) =
ln(x ) + 2, if x > 0. ln(x ) 1, if x > 0.

are both antiderivatives, and we have

F 1 (x ) F 2 (x ) =
ß
2, if
se x < 0;
3, se x > 0.

that is not constant.

g
However, even for functions not defined in an interval, but usually in a finite union of intervals, we use
the notation Z
f (x ) dx = F (x ) + c

implying that we restrict the problem to one of the subintervals of the domain. Moreover, in the
exercises we’ll write the indefinite integral skipping the additive constant of integration. For example
we ca write Z Z
1 1
dx = ln |x| + c or, briefly, dx = ln |x|.
x x
Besides that we can use “Indefinite integral” as a synonym of “antiderivative”.
L.Battaia 08 - Antiderivatives A.Y.2024/25 2 / 15
Important remark

It can be proved that all continuous functions defined in an interval have an

antiderivative (and hence a family of antiderivatives), but in general the antiderivative
of an elementary function is not an elementary function.
2 R 2
A famous and very important example is the Gauss’ function f (x ) = e x : e x dx
exists, but it can’t be computed in terms of elementary functions. This function is of
fundamental
p importance in statistics. But there are lots of examples of such functions:
ex 1
1 + x 4 , x , ln x . In practice it can be said that only for a very limited subset of
elementary functions the antiderivative is an elementary function. Observe that, in
contrast with what happens for antiderivatives, the derivative of an elementary
function is always an elementary function.

As a consequence of this, in general, computing antiderivatives is a very difficult, if not

impossible, task and only simple situations can be dealt with using the rules for
antiderivatives that we’ll consider in the next slides.

We’ll also consider a straightforward important strategy by means of which, using the
antiderivatives, we can obtain new non elementary functions.
L.Battaia 08 - Antiderivatives A.Y.2024/25 3 / 15
Rules of differentiation vs rules of integration ↑Blux) =

3((x))
Assume a, b 2 R and f , g differentiable (in case of differentiation), continuous
(in case of integration).
1
#af (x) + bg (x) 0 = af 0 (x) + bg 0 (x),
R R R
… af (x ) + bg (x ) dx =P
a f (x ) dx + b g (x ) dx + c .

to
:
0 1 0 1
2 x a = ax a ; ln(x) = x ,
R a x a+1
… x dx = a+1 + c, if a 6= 1,
R 1 R 1
… x dx = x dx = ln |x| + c, if a = 1.
3

…
0
eax = aeax ,
R
8 1
eax dx = a eax + c .
-e
So"Setere
0 e
X 4 ax = ax ln a, a > 0, a 6= 1,
R
O+ c, a > 0, a 6= 1.
x
a
… ax dx = ln a
de a

L.Battaia 08 - Antiderivatives A.Y.2024/25 4 / 15

Differentiation/Integration of a product
The derivative of the product of two differentiable functions is a simple rule:
0
f (x )g (x ) = f 0 (x )g (x ) + f (x )g 0 (x ).

Unfortunately the corresponding rule for integrals is not simple at all.

Let f (x )g (x ) be the product of two continuous functions, and suppose we are able to
R
compute an antiderivative of one of the two, for example f (x ) dx = F (x ). Then the
following formula holds:
Z bZ
f (x )g (x ) dx = F (x )g (x ) F (x )g 0 (x ) dx.

In practice this formula does not provide the integral of the product: it only reduces the
problem of integrating f (x )g (x ) to the problem of integrating F (x )g 0 (x ). If this second
integral is feasible, the formula is useful, otherwise it does not provide a result.
This formula is called Integration by parts because it splits the problem in two parts: the
integration of f (x ) and the integration of F (x )g 0 (x ). It can be formulated in other ways, but,
probably, the simplest way is the one written above.
Remark In words the formula reads: the integral of a product is the integral of the first times
the second minus a so called “residual integral”. Only the first part is similar to the formula
for differentiation.
L.Battaia 08 - Antiderivatives A.Y.2024/25 5 / 15
Integration by parts - Remarks
1 If, in the product of two functions, both have a simple antiderivative, there
R is no guarantee that
the product has a simply computable antiderivative. For example while ex dx = ex and
R 1
x dx = ln |x|, the integral
Z Z
1 ex
ex · dx = dx is not an elementary integral.
x x

2 Sometimes the formula by parts is applied to a single function f (x ), by writing f (x ) = 1 · f (x ).

The most important case concerns the natural logarithm.
Z Z Z Z
1
ln x dx = 1 · ln x dx = x ln x x dx = x ln x 1 dx = x ln x x + c.
x

3 When, in a product f g , both functions have a simple antiderivative you have two possibilities
for the first step: compute F or compute G . Unfortunately in many circumstances only one way
leads to a result. For example consider the product xex .
8 R
Z < x2 x
e
x2 x
e dx harder than the original
x 2 2
xe dx = R .
:
ex x ex · 1 dx = ex x ex + c

L.Battaia 08 - Antiderivatives A.Y.2024/25 6 / 15

 Integration by parts - Remarks (cont.)
4 Sometimes the rule must be applied several times in succession.

ledy
%
Z Z ÅZ ã
ex x 2 dx = ex x 2 ex 2x dx = ex x 2 2 ex x dx =
Å Z ã
= ex x 2 2 ex x ex · 1 dx = ex x 2 2ex x + 2ex + c.

5 Sometimes application of the formula leads back to original integral.

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

B
This looks like the formula A = B A)A= 2
. So
Z
1 ln2 x
ln x dx = .
x 2

I
6 For some products this formula is not useful. We’ll see, for example, that the following integral
can be simply computed, without using by parts formula, that leads to unnecessary
complications.
Z 2
2 ex
xex dx = + c.
2

L.Battaia 08 - Antiderivatives A.Y.2024/25 7 / 15

Sydx : By Par

-(ex m
Res in murde
intege
mi completed
Than the un final
Luciano Battaia
The symbol of the integral
Stady
We give a brief explanation of the originZ and meaning of the symbol used for integrals.
↑ 4
The symbol contains three parts: 1 : , 2 : f (x ), 3 : dx .
Z
1 The symbol is a deformation of the capital S , or of ⌃ (“the Summation symbol”), and stands
for “Sum”. The reason for this use will become clear while studying Riemann Integral.
2 f (x ) is the “integrand function”.
3 dx has different meanings and uses.
… It can serve as a “right delimiter” of the integrand function. For example
Z Z
x + ln x dx is the same as x + ln x dx.
-

… It highlights the integration variable when we have functions of many variables and we
want to use only one of them to proceed with integration. For example
Z Z
e2t x2
xe2t dt = x · +c while xe2t dx = e2t ·
+ c.
2 2
This is common in economic applications, where many variables are used.

d = In … Recall that, given a function f , its derivative satisfies the relation dy = f 0 (x ) dx . The dx
that appears in the integral sign can be considered as the differential of x and we’ll see
how this can be used in the integration by substitution.

L.Battaia 08 - Antiderivatives A.Y.2024/25 8 / 15

 x+ y
f(x y) , =

aut
f =
(f(x y)
,
=
2x

fu on

En =
/exus)' =
sy
Luciano Battaia
Integration by substitution
Z
gxnxxe
Assume you need to compute g (x ) dx .

The integrand function g (x ) is normally composed of various parts, for example various
factors, but not only: pick out a part of 00 it
-

g (x ), say h(x ), and put u = h(x ). The choice of

h(x ) inside g (x ) is the most difficult part of this technique and only hard practice can help.
-
- 2

Now you can use two complementary procedures: sometimes both work, but sometimes only
one works and, again, only hard practice can help to choose the right path.

Integration by substitution: procedure 1

1 Compute the derivative of h(x ), written as
-
Edu
dx
0

R
0
= h (x ) and rewrite as du = h (x ) dx .
2 If, using u = h(x ) and du = h0 (x ) dx , you can rewrite the integral as f (u ) du , for a
R
suitable function f , and if this last integral is solvable, f (u ) du = F (u ), then to obtain
the original integral you only need to replace u by h(x ) in F (u ).
Z
g (x ) dx = F (h(x )) + c.

L.Battaia 08 - Antiderivatives A.Y.2024/25 9 / 15

Integration by substitution (cont.)

Integration by substitution: procedure 2

1 There is one additional step with respect to procedure 1: invert the formula u = h(x ),
obtaining x = m(u ), where m = h 1 .
-

dx
2 Compute the derivative of m(u ), written as du
= m0 (u ) and rewrite as dx = m0 (u ) du .
R
3 If, using x = m(u ) and dx = m0 (u ) du , you can rewrite the integral as f (u ) du , for a
R
suitable function f , and if this last integral is solvable, f (u ) du = F (u ), then to obtain
the original integral you only need to replace u by h(x ) in F (u ).
Z
g (x ) dx = F (h(x )) + c.

R ln x
Example: Compute x dx (already computed also by parts).
-

Procedure 1 Procedure 2
du 1 1 dx
u = ln x , u 0 (x ) = dx
= x , du = x dx . u = ln x ) x = eu , du
= eu , dx = eu du .
un ne e
Z Z Z Z
u2 ln2 x u2 ln2 x
80-
ln x u u ln x
u du = ) dx = + c. e du = ) dx = + c.
2 x 2 eu 2 x 2
-

This exercise highlights one of the more subtle difficulties connected to the calculation of integrals:
the correct strategy is not evident a priori: only practice can help to suggest the correct path.
L.Battaia 08 - Antiderivatives A.Y.2024/25 10 / 15
jady
U = bu X

duz
su 1 dy
b

Judu =
-

Luciano Battaia
Practice integration rules Stedve Sxdy
R x 3 +x 2 +x +1 R R R R R

10
dx = 1x +1+ x + x 2 dx = x dx 1 dx +
1 1 1 2
1
x2
:
x dx + x dx =
x2 x 1
= 2
+ x + ln |x| + 1
+ c.
Note: from now on, when no confusion can arise, we skip the dx in the integral
symbol, except in the integration by substitution.
=

R p
3 p
x R 1/12 x 1/12+1
2 p = ··· = x = + c = ...
R
4
x -

R
1/12+1
↓
x 2 2
x 1 x 2
1R x2 1 x2
3 x ln x = 2
ln x 2 x
= 2
ln x 2 x = 2
ln x 2 2
+ c = ...
R x +1 2
4
x 2 +2x dx . (Hint: put u = x + 2x ). du = (2x + 2) dx = 2(x + 1) dx )
du R 1 du ln u ln(x 2 +2x )
(x + 1) dx = 2 ) u 2 = 2 ) 2
+ c.
R 1 p -
5 p
x 1
dx . (Hint: put u = x 1 or u = x 1).
R 1 R u 1/ 2 p
u = x 1, du = dx . pu du = u 1/2 = 1/2 ) 2 x 1 + c .

I
…
p R1 p
… u = x 1 ) x = 1 + t 2 , dx = 2t dt . t 2t dt = 2t ) 2 x 1 + c .

L.Battaia 08 - Antiderivatives A.Y.2024/25 11 / 15

 X
dux
u=
+ 2x

(x) d
=

due
eu

ten (x 2x)
+
=

Luciano Battaia
=
Jeux - x dy = (X(x x) X
-

((xex xy
=
-
2 =

(xex x
&
= x ex - x2 -

+ =

=
X (x - x2 -

(x(x +
X

2/exox = x(x - x +
= ...

Luciano Battaia
Practice integration rules (cont.) Jx = 2

6
R
Dot -
p 1
1+ x
p dx .(Hint: put u =
p
1+
p

R 1
x.

dx = 2(u 2 1)2u du . u 4u(u 2 1) du = 3

p
x = u2 1. x = (u 2
4u 3
4u .
1)2 .

R p p p
p 1 p dx = 43 1 + x( x 2) + c .
1+ x
R ln x R 1/2 1/ 2
R 1/ 2 1 p p
7 p = x
x
ln x = 2x ln x 2x x = · · · = 2 x ln x 4 x + c .
R x p
8 p
4 x
dx . (Hint: put u = 4 x ). 4 x = u 2 . x = 4 u 2 .
R (4 u 2 ) 2u 3
dx = 2u du . u ( 2u) du = 8u + 3
.
R x 2 p
p
4 x
dx = · · · = 3 (x 2) x + 1 + c .
R 2 2x 3 1
9 x e dx . (Hint: put 2x 3 = u . 6x 2 dx = du . x 2 dx = 6 du .
R eu eu R 3 e2x
3

6
du = 6 . x 2 e2x dx = 6 + c .

L.Battaia 08 - Antiderivatives A.Y.2024/25 12 / 15

 No INT .
BY
PART
sedy du
n =

=
-

-
x

= -x2
e
- -
=

Luciano Battaia
Remark on procedure 1
There are some situations where procedure 1 leads almost always to a conclusion.
1 When in the integrand function there is a factor of the form eg (x ) , with a suitable
function g (x ), a first attempt is to put u = g (x ). If this does not work, other strategies
must be used.p
Re x p 1 R R epx p
Example. p dx . Put u = x , du = 2px dx . 2eu du = 2eu . px dx = 2e x + c .
x
2 When in the integrand function there is a power of a certain function (f (x ))a , a first
attempt isRto p
put u = f (x ). R
3
Example. x x 2 1 dx = x (x 2 1)1/3 dx . Put u = x 2 1, du = 2x dx .
R u 1/ 3 1 u 1/3+1 R p 1 (x 2 1)4/3
2
du = 2 1/3+1
. x 3 x2 1 dx = 2 4/ 3
+ c = ....
3 A special and important case of application of procedure 1 is the following:
Z
f 0 (x )
= ln |f (x )| + c
f (x )
This last formula is so common that it is advisable to learn it by heart as a fundamental
rule of integration.
Z
3x 2 5
Example. = ln |x 3 5x| + c .
x 3 5x

L.Battaia 08 - Antiderivatives A.Y.2024/25 13 / 15

 Jedx in ampe
enough
=
22
E

>J Exine an
-
impossible
of the
:

entden.
no-daw
fets with
.

Luciano Battaia
 Je -2
Some applications
fN) = = +
-

Find a formula for f (x), knowing the following conditions.

1 f 0 (x) = 12x 3 2x 2 + 1 , f (1) = 4. D

R 2x 3
f (x) = (12x 3 2x 2 + 1) = 3x 4 3
+ x + c . As f (1) = 4, we obtain
2 2
---
3 3 + 1 + c = 4, that is c = 3 .
2x
2 f 0 (x) = (1+x 2 )3 , f (0) = 1.
R 2x 1 3
f (x) = (1+x 2 )3 dx = · · · = 2(1+x 2 )2 + c . As f (0) = 1, we obtain c = 2 .
3 f 00 (x) = ln x + e x
, f (1) = 1 , f 0 (1) = 0.
R e+1
… f 0 (x ) = (ln x + e x ) = · · · = x ln x x e x + c . f 0 (1) = 0 ) c = e .
R 1 3x 2
… f (x ) = x ln x x e x + c = · · · = 2 x 2 ln x 4
+ e x + cx + d .
3 2
f (1) = 1 ) d = 4 e .

L.Battaia 08 - Antiderivatives A.Y.2024/25 14 / 15

 lidy
d = 1x

dx = 2xdx

-2

Jad S udu
:

zu2

T
(EY' = -
e
Jetdx =
-e

Luciano Battaia
Homework

Z
1 Given a 2 R, a > 1, find x a ln x dx .
Z
x/2
2 Compute (x 2)e dx . (Hint: by parts. . . )
Z
ln(x + 2)
3 Compute dx . (Hint: put u = ln(x + 2) . . . )
2x + 4
Z
x p
4 Compute p dx . (Hint: put u = x 1. . . )
x 1
Z p
(1 + x)(1 x)
5 Compute p dx .
23x

L.Battaia 08 - Antiderivatives A.Y.2024/25 15 / 15

 u =
-

X/2

du-dy
Tu
2
J(
X
-

n -
1) ehydu =

dx =
-
Tu

YQue"
=

Pe =

Luciano Battaia
PTe n -Je" 13
.
+
Pe
=

Pe"
=
u
-

eh] +
4e" = ane

Y/2
* ~

1fz)e --xxe

Luciano Battaia
By perts Sx =
-

Jxlx = x -
(
I

-
Luciano Battaia
=ex-

Luciano Battaia