All LECTURES (CALCULUS 1)

LECTURE 1

Sequences
A sequence is an ordered set of numbers.
Each number in the set is called a term of the sequence.
An infinite sequence is a sequence with an infinite number of terms.
Sequence notation: a1 , a 2 , , , ,.a n or { a n }.
The first term of a sequence is denoted a1 . The second term a 2 . The term in the n-
th position, called the n th term, is denoted by a n .
Examples:
1) The sequence 1,4,9,16,25,…can be specified by the formula an  n 2 .
2) The sequence 2,4,6,8,10,… can be specified by the formula a n  2n .

Bounded sequences
A sequence { a n } is said to be bounded above if there is a number M such that
a n  M for all n.

In short, in terms of mathematics it can be written as follows:

{ a n }is bounded above  M n : a n  M .
A sequence { a n } is said to be bounded below if there is a number m such that
a n  m for all n .

Similarly, it can be written as follows:

{ a n } is bounded below  m n : a n  m .
Examples:
1) a n  n is bounded above.
2) an  n 2 is bounded below.
A sequence bounded both above and below is called bounded sequence. In
other words, a sequence { a n } is said to be bounded if there exist numbers M and
m such that m  a n  M for all n . In short, it can be written as { a n } is bounded

 M , m n : m  a n  M .

Examples:
1
1) Consider the sequence whose terms are defined by a n  . It is that
n
1
0  1 . Therefore, the given sequence a n is bounded.
n

2) A sequence is defined by an  (1) n .

We see that a1  1, a 2  1, a3  1, a 4  1, and so on. Since this sequence
assumes only two values -1 and 1 we can write  1  a n  1 . Therefore, the given
sequence a n is bounded.
Now we introduce another equivalent definition of bounded sequence.
A sequence { a n } is said to be bounded if there is a positive number C  0 such
that a n  C for all n .

{ a n } is bounded  C  0, n : an  C .
A sequence is said to be unbounded if it is unbounded either from above or
below. In other words, a sequence { a n } is said to be unbounded if given any
positive number C  0 there is a number n  N such that a n  C .
Examples of unbounded sequences are the sequences
an  n 2 , an  2n, an  n, an  (1) n  n, an  n . a

Limit of a sequence
Definition. A number a is said to be the limit of the sequence { a n },
a1 , a 2 ,..., a n ,... as n   if for all sufficiently large values of n the corresponding

values a n of the sequence { a n } become arbitrary close to the number a .

If a is the limit of the sequence { a n } as n   we write
lim an  a or an  a
n n

Now we give a mathematically more rigorous definition of the limit.

 Definition. A number a is said to be the limit of the sequence { a n }, as n   if
for any positive number  (however small) there is a positive number N such that
the inequality an  a   holds for all n  N. thus, we have
lim a n  a    0 N  0 n  N : a n  a  
n 

To understand the deep meaning of the concept of limit, it is advisable to

rephrase the definition in terms of “neighbourhoods.” Recal that the
neighbourhood of a point is any interval containing this point.
It is clear that
xn  a      xn  a   
a    xn  a   
xn  (a   , a   )  U (a),

where U (a ) is a neighbourhood of a point a .

We are now ready to define the concept of limit in terms of neighbourhoods.
Definition. A number a is said to be the limit of the sequence {xn} if for any
neighbourhood U (a ) of the point a there exist a positive number N such that
x n  U (a ) for all n  N .

lim xn  a  U (a) N n  N : xn  U (a)

n

In other words this means that.

A number a is said to be the limit of the sequence {xn} if for any
neighbourhood U (a ) of the point a outside it remain only a finite number of terms
of this sequence.
Example. The limit of the sequence
1 5 7 2n  1
, , , ..., , ...
2 7 8 n5

exists and is equal to 2. Indeed, for the absolute value of the difference between
2n  1
and 2, which is equal to
n5
2n  1 9
2 
n5 n5
to be less than a given positive number  , it is sufficient that the inequality
9 9
n5 should be fulfilled. This is equivalent to the condition n   5 , and
 
9
hence, given  , there exists N   5 such that for all n  N the absolute value of

2n  1
the above difference is less then  , which means that lim  2.
n  n5
Definition. A sequence that has a finite limit is called convergent.
If otherwise, the sequence is said to be divergent.
Definition. A sequence {an} is said to be infinitely large if, for any positive
number  (however large) there is a positive number N such that the inequality
a n   holds for all n  N .

Hence, lim a n      0 N  0 n  N : a n  
n 

Similarly, the definition of the limit of the sequence is paraphrased for the
case when this limit is infinite with a definite sign.
lim an      0 N  0 n  N : an  
n

lim an      0 N  0 n  N : an  
n

Obviously, if lim an   or lim an  

n  n 

then lim an  
n 

Examples of infinitely large sequences are

an  n, an  n, an  n 2 , an  (1) n  n, an  2n  1, an  (2n) 2n

an  ((2n  1)) 2n1 ; an  (n) n .

Moreover, it is clear that

lim n  , lim(2n) 2 n  
n  n 

lim(n)  , lim((2n  1) 2 n 1  

n  n 

lim n  ,
2
lim(n) n  
n  n 

lim(1)  n  
n
n 
 Infinitesimals
Definition. A sequence  n tending to zero as n   is called an infinitesimal.
1
{  n } is an infinitesimal  lim n  0 . Examples of infinitesimals are a n  ,
n n
n

, a n  , a n    .
(1) n 1 1 1 1
an  , an  2 , an 
n n n n! 2
Now we introduce a relation of equality between sequences and operations on
them.
Definition. {xn}={yn}  xn  y n n

{x n }  { y n }  {x n  y n }

{x n }  { y n }  {x n  y n }

{x n }  x n 
  , ( y n  0, n )
{ yn }  yn 

{xn }  {xn }

Theorem. A sum, a difference and a product of two infinitesimals is again an

infinitesimal.
In other words, if {  n } and {  n } are infinitesimals then { n   n }, { n   n } and
{ n   n } are also infinitesimals.

Theorem. The product of a bounded sequence by an infinitesimal is an

infinitesimal. { x n } is bounded and {  n } is an infinitesimal then { n x n } is a
infinitesimal.

Example. Let us consider the sequence 

sin n 
 which can be written as
 n 
sin n 1
 sin n 
n n
1
It is clear that xn  sin n is bounded since  1  sin n  1, and  n  is an
n
sin n
infinitesimal. Therefore the product  n x n  is an infinitesimal.
n

The theorems below are important for applications.

 Direct theorem. If a sequence has a finite limit it is representable as a sum of a
constant, equal to that limit, and an infinitesimal.
lim xn  a  xn  a   n where lim n  0
n n

Reciprocal theorem. If a sequence is representable as a sum of a constant and

an infinitesimal the constant summand is the limit of the sequence.
x n  a   n , where lim n  0  lim xn  a
n n

Properties of limits
1. If x n  C , C - constant n then lim xn  C . This means that the limit of a constant
n 

equals the constant.

2. If a sequence has a limit then this limit is unique, i.e., there no other limits of the
sequences.
3. If a sequence has a limit then it is bounded.
Proof. Let { x n } be convergent and lim xn  a . Take   1 . According to definition
n 

there exist a number N such that the inequality xn  a  1 holds for all n  N . Let
d  max1, x1  a ,..., x N  a . Then the inequality xn  a  d holds for all n .

The last inequality can be written as

 d  xn  a  d

or a  d  xn  a  d
denoting m  a  d and M  a  d we get
m  x n  M for all n .

This means that { x n } is bounded.

4. The limit of a sum of sequences is equal to the sum of the limits of these
sequences.
lim( xn  y n )  lim xn  lim y n
n n n

Proof. Suppose { x n }and { y n } are convergent and lim xn  a, lim y n  b

n n

According to the theorem we can write

x n  a   n and y n  b   n

where {  n }and { n } are infinitesimals.

It follows that
x n  y n  a  b   n   n or

x n  y n  a  b   n where  n   n   n and lim  n  0

n 

This means that lim( xn  y n )  a  b  lim xn  lim y n

n n n

5. The limit of a difference of sequences is equal o the difference of the limits of

these sequences:
lim( xn  y n )  lim xn  lim y n
n  n n

Proof. Let { x n } and { y n } be convergent and lim xn  a lim y n  b

n n

Then we have xn  a   n and y n  b   n

Where {  n } and {  n } are infinitesimals.
It follows that
xn  y n  a  b   n   n or
xn  y n  a  b   n

where  n   n   n and lim  n  0

n 

Hence, lim( xn  y n )  a  b  lim xn  lim y n

n n n

6. The limit of a product of sequences is equal to the product of the limits of these
sequences.
lim xn y n  a  b  lim xn  lim y n
n n n

Proof. Let { x n } and { y n } be convergent and lim xn  a and lim y n  b

n  n

Then we can write

x n  a   n and y n  b   n

where {  n } and {  n }are infinitesimals.

It follows that
xn y n  (a   n )  (b   n )  ab  ( n b   n a   n  n )  ab   n

where  n   n b   n a   n  n
According to the properties of infinitesimals it is clear that lim  n  0 . Hence
n 

lim xn y n  a  b  lim xn  lim y n .

n n n

7. A constant factor may be taken outside the sign of the limit.

lim Cx n  C  lim xn where C is constant.
n n

This property follows from properties 1 and 6 .

8. The limit of a ratio of sequences is equal to the ratio of the limits of these
sequences.
x n lim xn
lim  n  ( y n  0 and lim y n  0 )
n  y n lim y n n
n 

9. If xn  y n  z n for all n and lim xn  lim z n  a then lim y n  a

n  n  n

Proof. We have
lim x n  a    0 N1 n  N1 : x n  a  
n 

lim z n  a    0 N 2 n  N 2 : z n  a  
n 

Now putting N  max{N1 , N 2 } we get

xn  a   and z n  a   for all n  N .

It follows that
a    xn  y n  z n  a  

or a    y n  a  
or y n  a   for all n  N
Hence we get
  0 N n  N : yn  a   .

By definition, this means lim y n  a . and we get what we had to prove.

n
10. Theorem.

 a0
 b for p  q,
a 0 n  a1 n  ...  a k 
p 1
p 0

lim  , for p  q,

n  b n q  b n q 1  ...  b
0 1 s 0 , for p  q



Monotone sequences
Definition. A sequence { x n } is said to be increasing if the inequality x n  xn1 holds
for all n .
Examples of increasing sequences are
xn  n, xn  2 n , xn  n , xn  n 2

Definition. A sequence { x n } is said to be decreasing if the inequality x n  xn1 holds

for all n. Examples of decreasing sequences are
1 1 1
xn  , x n  n, x n  n , x n 
n 3 n

Definition. Increasing and decreasing sequences are called monotone.

Note. Not every sequence is monotone. For example, the sequences (1) n and
 1
(1)
n
 are not monotone.
 n

Theorem. A monotone and bounded sequence is convergent. This theorem can be

rephrased as follows.
Theorem. An increasing and bounded above sequence has a finite limit.
A decreasing and bounded below sequence has a finite limit.

Number e
n

Consider the sequence xn  1   , n  1,2, ,...

1
 n

We shall prove that { x n } is convergent.

By Newton’s binomial formula, we have
 n 1 n(n  1) 1 n(n  1)(n  2) 1 n(n  1)...( n  n  1) 1
n
 1
x n  1    1     2   3  ...   n 
 n 1! n 2! n 3! n n! n
1  1  1  1  2  1  1  2  n  1 
1  1  1    1  1    ...  1  1  1  
2!  n  3!  n  n  n!  n  n  n 

Replacing n by n  1 we obtain an analogous expression for xn1 .

n 1
 1  n 1 1 (n  1)  n 1 (n  1)  n  (n  1) 1
x n 1  1    1       ...
 n 1 1! n  1 (n  1) (n  1) 3
2
2! 3!
(n  1)  n(n  1)...( n  1  n) 1 1 1  1 1  2 
...   n 1
 1  1  1    1  1    ...
(n  1)! (n  1) 2!  n  1  3!  n  1  n  1 
1  1  2   n 
...  1  1  ...1  
(n  1)!  n  1  n  1   n  1 

Now, comparing these expressions for x n and xn1 we conclude that the first two
terms in both sums coincide and that every subsequent term in x n is less than the
corresponding term in xn1 since
k k
1  1 , k  1,2,..., n  1
n n 1

Therefore x n  xn1 for any n, that is x n is an increasing sequence.

Let us show that it is bounded. It is clear that
1 1 1 1 1 1
xn  1  1    ...   1  1   2  ...  n1 
2! 3! n! 2 2 2
1 1 1 1  1 1 1 
 1  1   2  ...  n 1  n  ...  1  1   2  ...  n  ... 
2 2 2 2  2 2 2 
The sum of the progression in the parentheses being equal to 2 we obtain xn  3 .
Thus, the increasing sequence { x n }is bounded above, hence according to the above
theorem, it possesses a finite limit.
Thus, we have proved the following theorem.
n

Theorem. The sequence xn  1   possesses a limit as n   .

1
 n
This limit plays a very important role in mathematics and is traditionally denoted
by e.
Definition. The number e is the limit
n
 1
lim1    e .
n 
 n
The number e is irrational and therefore cannot be expressed by a finite decimal.
Its approximate value is
e  2,71828182845...  2,718

TEST
n 3  2n  1
3
1. lim
n  n2
A) 7 B) 6 C)5 D)0 E)1

2. lim
n 2
1  n 
2

n  3
n6  1

A) 4 B) 5 C)6 D)7 E)8

1 1 1
1
  ...  n
3. lim 2 4 2
n  1 1 1
1    ... n
3 9 3

A)
5
B)
4 C)2 D)1 E)
1
3 3 3

1  2  3  ...  n n 
4. lim  
n
 n2 2

A) 2 B) -1 C) 
1 D)3 E)
1
2 4

9n 2  5
5. lim
n  6n  11

1 1 1 1 1
A) B) C) D) E)
6 5 4 3 2

1 n2 
6. lim   
n  n  3 
n

A)  B) 0 C)1 D)2 E)3

 n 1
7. lim
n 
n2  3

A) 5 B) 4 C)3 D)2 E)1

16n 2  3n  7
8. lim
n  (2n  1) 2

A) 3 B) 4 C)1 D)2 E)0

n  (1) n
9. lim
n  n  ( 1) n

A) 1 B) 2 C)3 D)4 E)5

10. lim n  1  n
n

A) 0 B) 2 C)3 D)6 E)-1

LECTURE 2
Subsequences. Fundamental sequences.
Subsequence.
Let us consider a sequence of numbers
an n1 : a1 , a2 ,..., an ,...
And an increasing sequence of positive integers
nk n1 : n1  n2  ...  nk  ...
  
Then the sequence a nk k 1 is said to be the subsequence of the sequence an n1


Examples:
Let us consider a sequence of positive integers nn1 : 1,2,3,…,n,…


Then each of following sequences

2k k 1 : 2,4,6,…
2k  1k 1 : 1,3,5,…
3k k 1 : 3,6,9,.. is the subsequence of the sequence n .
Because all terms of a subsequence are also terms of the original sequence, the
properties of the subsequence are closely tied to the properties of the sequence, and
so a great deal of information about one can be determined by studying the other.

 
Here are some obvious properties:
 
1. If a sequence xn is bounded then any of its subsequences xnk is also
bounded.
 2. If a sequence xn   is increasing ( respectively decreasing), then any of its
 
subsequences xnk is also increasing ( respectively decreasing).

Convergence of Subsequences

The convergence of a sequence can be characterized in terms of the convergence of

its subsequence.
Theorem. A sequence converges to a limit a if and only if every subsequence
also converges to the limit a .
Proof. Necessity .Suppose that lim a  a i.e. an  a and consider some
n  n

 
subsequence a nk , nk   n.
Because an  a , for every   0 , there is N  0 ,such that
n  N  an  a   .
Then, for ank  
k 1
, there must be some K such that k  K  nk  N , because
the
{nk }  are increasing. But if nk  N , then an  a   , so
k
an  a
k
as
k   i.e. lim a a
k  n k

Thus, any subsequence converges as well.

 
Sufficiency. Suppose that every subsequence a nk converges. In particular we can
take {nk }  n, so the sequence an   converges because every sequence can be
regarded as a subsequence of itself.
Bolzano-Weierstrass theorem. Every bounded sequence has a convergent

 
subsequence.
In other words, if xn is a bounded sequence, then there exist a subsequence
x  x  such that converges.
nk n

Proof. Let xn   

n 1
be bounded sequence. Then there exist an interval a1 ,b1 such
that a1  xn  b1 for all n
 a1  b1 
Let us divide the closed interval a1 ,b1  into two equal intervals a1 ,
 2 
 a1  b1 
and  , b1  . At least one of them contains infinitely many terms of
 2 
x .That is, there exists infinitely many n such that a
n n

is in a1 ,

a1  b1 
2 
or

 a1  b1   a b 
there exists infinitely many n such that an is in  , b1  . If a1 , 1 1 
 2   2 

 
contains infinitely many terms of xn , let a2 , b2   a1 ,
 a1  b1 
 2 
. Otherwise,

 a1  b1 
let a2 , b2    , b1  .
 2 
 a2  b2   a2  b2 
Either a2 ,
 
2   2
or  , b2  contains infinitely many terms of xn .If

 
 a2  b2 
a2 , 2  contains infinitely many terms of xn , let  
a3 , b3   a2 , a2  b2  .
 2 
 a2  b2 
Otherwise, let a3 , b3    , b2  .By mathematical induction, we can
 2 
continue this construction and obtain a sequence of intervals an , bn  such that
i) for each n , an , bn  contains infinitely many terms of xn ,  
ii) for each n , an1
, bn1   an , bn  and

iii) for each n , bn1  an1 

1
bn  an 
2

 
The nested intervals theorem implies that the intersection of all the
intervals an , bn  is a single point x. We will now construct a subsequence of xn

 
which will converge to x.
Since a1 ,b1  contains infinitely many terms of xn , there exists k1 such that
 
x k is in a1 ,b1 . Since a2 ,b2  contains infinitely many terms of xn , there
1

exists k 2 , k2  k1 , such that x k 2 is in a2 ,b2 . Since a3 ,b3  contains infinitely
 
many terms of xn , there exists k 3 , k3  k1 , such that x k3 is in a3 ,b3  .
Continuing this process by induction, we obtain a sequence xkn such that xkn is in
a , b  for each n. The sequence x  is a subsequence of x
n n kn n
 since k n1  kn
for each n. Since an  x , bn  x and an  xn  bn for each n. the squeeze
theorem xkn  x
Fundamental sequence(Cauchy sequence )
A Fundamental sequence( or a Cauchy sequence ) is a sequence whose elements
become arbitrarily close to each other as the sequence progresses. More precisely,
given any small positive distance, all but a finite number of elements of the

 
sequence are less than that given distance from each other.
Definition. A sequence xn : x1 , x2 , x3 ,... of real numbers is called a fundamental
sequence ( or a Cauchy sequence ) if for every positive real number  there is
positive integer N  N   such that for all naturel numbers m, n  N
xm  xn  
x 
n


n 1
is a fundamental sequence    0 N  N   :
m, n  N xm  xn  

   
The Cauchy Convergence Criterion
Theorem. If xn is a sequence of real numbers, then xn is convergent if and
  is a fundamental sequence.
only if xn
x  is convergent  x  is a fundamental sequence.
Proof. "" Suppose that x  is a convergent sequence of real numbers, It
n n

A , such that lim

means that there is the real numbers x  A . It follows that for
n  n

every   0 there exists a number a number N  N   such that for all


n  N  xn  A  . And so for m  N and n  N we have
2
 
x m  xn  xm  A  A  xn  xm  A  x n  A    
2 2
 
that means xn is fundamental.
Thus, we already proved that a convergent sequence of real numbers is

 
fundamental.
"" Suppose that xn is a fundamental sequence. Then for   0 there
exists an N  N   such that if m  N and n  N then.
 
Now look at the a subsequence xnk which converges to A . Thus there exists
a natural number k  N1 where k belongs to the set of indices n1 , n2 ,... such

that xk  A 
2
Since k  N1 then if we substitute m  k we have that for n  N1 :
  
xk  xn  or xn  xk 
2 2
And so for n  N1 we have that:
 
xn  A  xn  xk  xk  A  xn  xk  xk  A   
2 2
Thus, lim x  A , and therefore xn
n n
  is convergent to the real number A .
Now let us summarize the proved theorem (The Cauchy Convergence Criterion) as

 
follows:
1. Any sequence of real numbers xn that is convergent is also is a
fundamental sequence
2. Any fundamental sequence of real numbers is also a convergent sequence.

Examples:

1.The sequence xn   1 has no limit since it is not fundamental. Let us show

n

it.
Denying the statement that the sequence is fundamental looks like this
  0 n n  N : xm  xn  

In our case, it sufficient to put   1. Then for any N we have

x N 1  x N  2  1   1  2  1  
Therefore xn   1 is not fundamental
n

 
2.Consider the following sequence xn , where xn  1 
1 1
  ... 
2 3
1
n
Since for any n  N  1,2,...
1 1 1 1
x2 n  xn   ...   n 
n 1 nn 2n 2
So according to the Cauchy Convergence Criterion the sequence xn   has no
limit.
Partial limit
Partial limit of given sequence is the limit of some subsequence.
Let us discuss this concept in more detail. Let us consider a sequence xn  

n 1

and let a  R   ,

Definition.
A number a  R (either symbol a   or a   ) is called a partial limit
   
of the given sequence xn if it contains a subsequence xnk converging to
a.
Thus, if a is partial limit of x , then there exists subsequence
n

x   x such that lim x

nk n k  nk
a
Examples:
1. Suppose sequence xn   is defined as follows 1,2,3,1,2,3,...
Let us consider the following subsequences of xn 
:
x   x   1,1,1,...
nk

3 k  2 k 1

x   x   2,2,2,...
nk

3 k 1 k 1

x  x   3,3,3,...
nk

31 k 1

It is clear that
lim x  lim
k  3 k  2 k 
11
lim x  lim
k  3 k 1 k 
22
lim x  lim
k  3 k k 
33
Therefore, the numbers 1,2 and 3 are partial limits of xn .
2. Let us consider the sequence xn , where xn   1 .
n

 1,1,1,1,...
Consider the following subsequences of given sequence:
x2k 1: 1,1,1,... with lim x  1
k  2 k 1
x2 :1,1,1,... with lim x 1
k  2 k
and
Therefore, the numbers -1 and 1 are partial limits of the given sequence xn .
Theorem. Any sequence of numbers has at least one partial limits( finite or
infinite)
Proof. In the case when a sequence xn  is bounded the statement of the
theorem follows from the Bolzano-Weierstrass theorem.It remains to consider
the case when the sequence is unbounded.
Let xn  be unbounded. Then for k  N we will choose nk  N such that
xn  k and nk  nk 1
k

As a result we get a subsequences that tends to infinity.

Upper and Lower limits of a real sequence
 The upper and lower limits of a sequence of real numbers xn  can be defined
in several ways and are denoted, respectively as
lim x and lim xn
n  n n 
One possible way is following Definition:
Consider the set A of elements a  R   , for which there is
subsequences of xn  converging to a .
A  a : xn   a
k

It is clear that the set A is the set of all partial limits of the sequence xn .
Then
limx  max A
n n
and lim xn  min A
n
Example 3.

Consider xn   1 .
n

It is clear that A   1,1

Therefore lim x  1 and lim xn  1
n  n n 
The upper and lower limits can also be defined in several alternative ways.
In particular as follows.
Definition.
lim x  lim
k  k n 
sup xk , lim xn  lim
n 
inf xk
k n k  k n

Example 4.

xk   1 , k  N  1,2,3,...
k

lim xn  lim inf xk  lim inf  1  lim 1  1

k

k  n  n  k  n n 
k n

x  lim sup xk  lim sup  1  lim 11

k
lim
k  k n  n  n 
k n k n

Example 5.

xk  k , kN
lim xn  lim
n 
inf xk  lim inf k  lim
n  k  n n 
n  
k  k n

lim x  lim
k  k n
sup xk  lim
n 
sup k  lim
n 
n  
k n k n

Example 6.

1
xk  kN
k ,
 1 1 1 1 
lim xn  lim inf xk  lim inf  lim inf  , , ,...  lim 0  0
k  n
k n
n k n
k n  n n  1 n  2  n
1 1 1 1  1
lim xk  lim sup xk  lim sup  lim sup  , , ,...  lim 0
k  n k n n k  n
k n  n n  1 n  2  n n

Example 7.

xk   k 2 , kN
lim xn  lim
k  n 
inf xk  lim inf  k 2   lim
n  k  n
k n
n 

inf  n 2 ,n  1 ... 
2

lim
n 
    
lim
k 
x k  lim
n 
sup x k
k n
 lim
n 
sup
k n
 k 2
  lim
n 
sup n 2 2

,n  1 ,... 

lim
n 
 n 2    lim
n 
n 2  
Example 8.

xk   1 k , kN
k

xk : 1,2,3,4,5,6,...
lim xn  lim inf xk  lim inf  1 k  lim     
k

k  n n  k  n n
k n

x  lim sup xk  lim sup 1 k  lim     

2
lim
k  k n n n 
k n k n

Theorem.
The limit of xn  is a real number A (respectively  , ) if and only if the
upper and lower limit coincide and are a real number A (respectively  , )
Lecture 3

Limit of a function

Definition. A number A is said to be the limit of the function f (x ) as x  a if for

all the values of x lying sufficiently close to a the corresponding values of the
function f (x ) are arbitrary close to the number A. It can be written as
lim f ( x)  A or f ( x)  A
x a xa

It should be noted that this definition does not require that the function should be
defined at the point a itself. It is only necessary that the function should be defined
in an arbitrary neighborhood of this point. Moreover, the phrase “arbitrary close”
means “as close as you want”. Now let us introduce the following
Definition. A number A is said to be the limit of the function f (x ) as x  a if for
given any positive number  (however small), there is a positive number  such
that for all x different from a and satisfying the inequality 0  x  a   there holds
the inequality
f ( x)  A  

In short
lim f ( x)  A    0   0 x : 0  x  a    f ( x)  A  
x a

This definition is sufficiently rigorous because the imprecise terms “arbitrary

close” and “sufficiently close” in the previous definition have been replaced by a
precise statement using inequalities.
Example: Let f ( x)  5 x  3 and prove that lim f ( x)  2
x 1

To this end, choose a positive number  . For the inequality

f (x)  2   or (5 x  3)  2  

or the equivalent inequality

5  x 1  

to be fulfilled it is sufficient that the condition


x 1 
5

should hold.

Hence, for all x differing from 1 by less than the function f (x ) differs from 2 by
5

less than  where  is an arbitrary positive number.


In this example we can put   . Therefore
5

  0    0 x : 0  x  1    f ( x)  2   .
5

According to definition this means that lim f ( x)  2 . It should be noted that here it
x 1


is not necessary to stipulate the condition x  1 since inequality x  1  is also
5
fulfilled for x  1 .
In this example the limit of the function f (x ) as x approaches 1 is equal to the value
of the function at that point.
lim f ( x)  2  f (1) .
x1

Note. If the limit of a function f (x ) as x approaches a exists, this limit may not be
equal to f (a ) . In fact, f (a ) may not even be defined.
x2  4
Example. Let us consider the function f ( x)  which is defined for all x except
x2
2 ( x  2 ).
It is clear that
x2  4 ( x  2)( x  2)
lim f ( x)  lim  lim  lim( x  2)  4  f (2)
x2 x2 x  2 x 2 x2 x 2

In this example the limit exist and is not equal to the value of the function at the
given point 2. Moreover, f ( 2) is not defined.
y
Figure shows the graph of
x2  4 4
f ( x)  , x  2, near x  2. As x 0
x2
approaches 2 from the left or right, the
corresponding values of f (x ) becomes closer O 2 x

to 4.
Properties of limits
These are a number of facts that simplify the computation of limits.
1. If f ( x)  C (C-constant) then lim f ( x)  C .
x a

2. If f ( x)  x then for every number a, lim x  a

xa

3. If g ( x)  f ( x)  h( x) and there exist lim g ( x)  lim h( x)  A then lim f ( x)  A

x a x a x a

4. If lim f ( x)  A and lim g ( x)  B , then

x a x a

(a) lim( f  g )(x)  lim f ( x)  g ( x)  A  B

x a xa

(b) lim ( f  g )( x)  lim f ( x)  g ( x)  A  B

x a x a

(c) lim ( fg )( x)  lim f ( x) g ( x)  A  B

x a x a
 f  f ( x)  A
(a) lim  ( x)  lim   (for B  0)
xa  x a
g  g ( x)  B

(e) lim f ( x)  A ( for f ( x)  0 for all x near a )

x a

5. If f (x ) is a polynominal function, that is

f ( x)  a0 x n  a1 x n1  ...  an ,

Then lim f ( x)  f (a) for any real number a.

x a

6. Let f (x ) be a rational function, that is

a0 x n  a1 x n1  ...  an
f ( x) 
b0 x m  b1 x m1  ...  bm

and let a be a real number such that f (a ) is defined. Then lim f ( x)  f (a )

x a

 a0
b for n  m
a0 x n  a1 x n1  ...  an 
0

7. lim m    for n  m
x  b x  b x m1  ...  b
0 1 m 0 for n  m



Remarkable limits
sin x
Theorem. lim
x 0
1
x
Proof. Consider a circle of unit radius suppose that the control angle x

expressed in radians and 0  x  . From the figure it is clear that
2
AreaACO  Areasec torACO  AreaBCO

We have A B
1 1 1
AreaACO  OC  AD   1  sin x  sin x 1
2 2 2
x
1 1 1
Areasec torACO  OC 2  AC   1  x  x O D C
2 2 2
1 1 1
AreaBCO   OC  BC   1  tan x  tan x
2 2 2
 1 1 1
It follows that sin x  x  tan x
2 2 2

or sin x  x  tan x
Dividing all the terms of these inequalities by sin x we get
x 1
1 
sin x cos x
which can be written as
sin x
cos x  1
x

Since lim
x 0
cos x  1 and lim
x 0
1  1 passing to the limit as x  0 we get, in

sin x
accordance with properties of limits, the desired result: lim
x 0
 1.
x
Many other limits can be computed with the aid of this limit.
Examples
tan x sin x 1 sin x 1
1) lim  lim   lim  lim  1  1  1
x0 x x0 x cos x x0 x x0 cos x
sin kx k  sin kx sin kx
2) lim  lim  k lim  k 1  k
x0 x x 0 kx x 0 kx

1  cos 2 x
2
2 sin 2 x  sin x 
3) lim  lim  2  lim   2 1  2
 x 
x0 2  2 
x x 0 x x 0

arctan x arctan x
4) lim  lim
x0 x x 0 tan(arctan x)

Putting arctan x  t , it is clear that t  0 as x  0 we obtain

arctan x t 1 1
lim  lim  lim  1
x 0 x t  0 tan t t  0 tan t 1
t

arcsin x
5) lim
x 0
1
x
arcsin x arcsin x y
lim  lim  lim 1
x0 x x0 sin(arcsin x ) y 0 sin y

sin x
The limit lim  1 playing an important role in mathematics is called the
x0 x

first remarkable limit.

 x

Theorem. lim1    e
1
x 
 x

Many other limits can be computed with the aid of this limit. Examples:
x

1) lim1    e k
k
x
 x
kx

2) lim1    e k
1
x
 x

3) lim1  x x  e
1

x0

ln(1  x)
1
1
4) lim  lim ln(1  x)  lim ln(1  x) x  ln e  1
x0 x x0 x x0

a x 1
5) lim  ln a
x 0 x
Indeed, putting a x  1  t  0 as x 0 . We have a x  1  t hence
1
x ln a  ln(1  t ) therefore x   ln(1  t ) .
ln a

Making substitution we get

a x 1 t t
lim  lim  ln a  lim  ln a
x0 x t 0 1
 ln(1  t )
t 0 ln(1  t )
ln a
x

The limit lim1    e has a wide application and is called the second
1
x 
 x
remarkable limit.
One – sided limits
The function whose graph is shown below is defined for all values of x except
x=2. y

4 -2 O 2 4 x
 As x approaches 2 from the right, the graph shows that the corresponding
values of f (x) get very close to 1. Consequently,
lim f ( x)  1 or lim f ( x)  1
x2 x20
x 2

The notation x  2  0 indicates that only the values of x with x  2 are

considered. Similarly, as x approaches 2 from the left, the graph shows that the
corresponding values of f (x) get very close to 4. Consequently,
lim f ( x)  4 or lim f ( x)  4
x2 x20
x 2

Here, the notation x  2  0 indicates that only the values of x with x  2 are
considered.
These “one-sided” limits are a bit different than the “two sided” limits
discussed in previous sections. When x approaches 2 from the left and from the
right, the corresponding values of f (x) do not approach a single number, so the
limit as x  2, i.e. lim f ( x) does not exist.
x2

Let us introduce the following notations:

x  a  0  x  a and x  a
xa
x  a  0  x  a and x  a
ax
f (a  0)  lim f ( x)  lim f ( x)
xa 0 xa
x a

f (a  0)  lim f ( x)  lim f ( x)
xa 0 xa
x a

Definition. The numbers f ( a  0) and f ( a  0) are called respectively the left-

hand limit and the right-hand limit provided that the corresponding limits exist.
The computation of one-sided limits is greatly facilitated by the following
fact.
 All the results about limits in previous sections, such as the properties of
limits, limits of polynomial functions remain valid if “ x  a ” is replaced by either
“ x  a  0 ” or “ x  a  0 ”.
Now we state the following necessary and sufficient condition for existence of
limit.
Theorem.
 lim f ( x)  A  f (a  0) f (a  0) and f (a  0)  f (a  0)  A
x a

It means that if a function f (x) possesses a limit as x  a (i.e. x approaches a

in an arbitrary way) the left-hand and the right-hand limits also exist and coincide
with the limit of the function. Conversely, if one-sided limits f (a  0) and

f (a  0) exist and coincide f (a  0)  f (a  0)  A then the function f (x ) has the

same limit A as the argument x approaches a arbitrary.

Local comparison of functions
Big O notation
Let f be a real or complex valued function and g a real valued function. Let
both functions be defined on some unbounded subset of the real positive numbers,
and be strictly positive for all large enough values of x .One writes
f ( x)  Og x as x  
if the absolute value of f (x ) is at most a positive constant multiple
of g (x) for all sufficiently large values of x.
That is, f ( x)  Og x if there exists a positive real number M and a real
number x0 such that
f  x   Mg  x  for all x  x0

The notation can also be used to describe the behavior of f near some real number
a (often, a = 0): we say
f ( x)  Og x as x  a

if there exist positive numbers  and M such that for all x with 0  x  a   ,

f  x   Mg  x 
As g(x) is chosen to be non-zero for values of x sufficiently close to a, both of
these definitions can be unified using the limit superior:
f ( x)  Og x as x  a if

f x 
0  lim sup 
x a
g x 

In typical usage the O notation is asymptotical, that is, it refers to very large x. In
this setting, the contribution of the terms that grow "most quickly" will eventually
make the other ones irrelevant. As a result, the following simplification rules can
be applied:

 If f(x) is a sum of several terms, if there is one with largest growth rate, it
can be kept, and all others omitted.
 If f(x) is a product of several factors, any constants (terms in the product that
do not depend on x) can be omitted.

Example.

let f(x) = 6x4 − 2x3 + 5, and suppose we wish to simplify this function,
using O notation, to describe its growth rate as x approaches infinity.
This function is the sum of three terms: 6x4, −2x3, and 5. Of these three terms, the
one with the highest growth rate is the one with the largest exponent as a function
of x, namely 6x4. Now one may apply the second rule: 6x4 is a product
of 6 and x4 in which the first factor does not depend on x. Omitting this factor
results in the simplified form x4.
Thus, we say that f(x) is a "big O" of x4. Mathematically, we can write f(x) = O(x4).
One may confirm this calculation using the formal definition:
let f(x) = 6x4 − 2x3 + 5 and g(x) = x4. Applying the formal definition from above,
the statement that f(x) = O(x4) is equivalent to its expansion,
f  x   Mx 4
for some suitable choice of x0 and M and for all x > x0.
To prove this, let x0 = 1 and M = 13. Then, for all x > x0:
6 x 4  2 x 3  5  6 x 4  2 x 3  5  6 x 4  2 x 4  5 x 4  13 x 4
So
6 x 4  2 x 3  5  13 x 4
Properties
Product
  f1  Og1  and f 2  Og 2   f1 f 2  Og1 g 2 
 f  Og   O fg 
Sum
 f1  Og1  and f 2  Og 2   f1  f 2  Omax( g1 , g 2 )
This implies

f1  Og  and f 2  O g   f1  f 2  Og 

Multiplication by a constant
Let k be constant. Then:
 O k g   O g 
if k is nonzero
 f  Og   kf  Og 

Little-o notation

Intuitively, the assertion "f(x) is o(g(x))" (read "f(x) is little-o of g(x)") means
that g(x) grows much faster than f(x). Let as before f be a real or complex valued
function and g a real valued function, both defined on some unbounded subset of
the real positive numbers, such that g(x) is strictly positive for all large enough
values of x. One writes
f ( x)  og x as x  
if for every positive constant ε there exists a constant N such that
f  x   g  x  for all x  N

As g(x) is nonzero, or at least becomes nonzero beyond a certain point, the

relation f ( x)  og x is equivalent to
f x 
lim 0
x 
g x 
Little-o respects a number of arithmetic operations. For example,
 if c is a nonzero constant and f  og  then c  f  og 
 if f  oF  and g  oG  then f  g  oF  G 
It also satisfies a transitivity relation:
f  og  and g  oh then f  oh
 Comparison of infinitesimal Quantities
Suppose   x  and   x  are infinitesimal quantities as x  a .

1. If lim  0 ,then  is said to be an infinitesimal quantity of the higher order
x a

of smallness as compared to  . In this case we write a  o 


2. If lim  m , where m is nonzero number, then  and  are said to be
xa

infinitesimal quantities of the same order of smallness. In particular, if

lim  1 then the infinitesimal quantities  and  are said to be
xa

equivalent. The notation  ~  signifies that  and  equivalent
infinitesimal quantities.
 
3.If   , this means that lim  0 Thus,  is an infinitesimal quantity of
 x a

the higher order of smallness compared to  i.e   o 

4.If  k and  are infinitesimals of the same order of smallness, with k  0 ,

then the infinitesimal  is said to have the order k as compared to 
Example .Assume t to be an infinitesimal quantity. Compare the infinitesimal
  5t 2  2t 5 and   3t 2  2t 3
Solution. We find
 5t 2  2t 5 5  2t 3 5
lim  lim  lim 
t 0
 t 0 3t 2  2t 3 t 0 3  2t 3
Since the limit of the ratio of  to  is a number different from zero, then then
 and  are infinitesimals of the same order of smallness.
Example. Compare the infinitesimal quantities   t sin2 t and   2t sin t as
t 0
Solution. We find
  t sin 2 t 1
lim  lim  limsin t  0 i.e a  o 
t 0
 t 0 2t sin t 1 t 0
TEST
x2  x
1) lim
x1 x 1

A) 1 B) 2 C) 3 D) 4 E) 5

3
1 x  3 1 x
2) lim
x0 x
2 3 4 5 7
A) B) C) D) E)
3 4 5 6 8

3) lim
x

x2  1  x2 1 
A) 1 B) 2 C) 0 D) 3 E) 4

 x 1  
3
4) lim x 2 3
x3  1
x

A) 5 B) 4 C) 3 D) 2 E) 1
2 x  arcsin x
5) lim
x0 2 x  arctan x

1 1 1 1 1
A) B) C) D) E)
6 5 4 3 2

1  5x 5 x
1
6) lim
x0

A) e B) e 2 C) e 3 D)
1 E) 2
e

7) lim1  8x x
2

x0

A) e 7 B) e 3 C) e16 D) e8 E) e 2
1  cos x
8) lim
x0 x
A) 0 B) 2 C)
1 2 3
2 D) E)
2 2
 sin 3( x  1)
9) lim
x1 6( x  1)

A)
1
B)
1
C)
1 D) 1 E) 2
3 4 2
sin 4 x
10) lim
x x
8

A) 1 B)
8
C)
4 D) 4 E) 0
 
Lecture 4
Continuity of a function

Let a be a real number in the domain of a function f. Informally, the function f is

continuous at x  a if you can draw the graph of f at and near the point
(a, f (a)) without lifting your pencil from the paper. For example, each of the
following graphs is the graph of a function that continuous at x  a .

y y y
(a, f (a))
(a,f(a))

a х a х a х
(a,f(a))

Thus, a function is continuous at x  a if its graph around the point (a, f (a)) is
connected and unbroken.
On the other hand, none of the functions whose graphs are shown below is
continuous at x  a .

y y y
(a, f (a))

(a,f(a))

a х a х a х
 Each of the functions is discontinuous, that is, not continuous at x  a because
you cannot draw the graph of f at and near the point (a, f (a)) without lifting your
pencil from the paper. In other words, a function f is not continuous at x  a if its
graph has a jump, break, gap, or hole when x  a .
Now we state the following formal definition of continuity.
Definition. A function y  f (x) is said to be continuous at a point a if it is
defined in a neighbourhood of the of the point a and lim f ( x)  f ( a ) .
x a

This definition we can restate as follows.

Definition. A function y  f (x) is said to be continuous at a point a under the
following conditions:
1. f (a ) is defined
2. lim
xa
f ( x) exists

3. lim f ( x)  f ( a ) .
x a

Here is another equivalent definition of continuity:

Definition. A function y  f (x) is said to be continuous at a point a if for given any
  0 , there is   0 such that for all the values of x satisfying the condition
x  a   the inequality f ( x)  f (a)   is fulfilled.

Properties of continuous functions

Theorem. If the functions f and g are continuous at x  a , then each of the

following functions
1. The sum function f +g
2. The difference function f  g
3. The product function f g
 4. The quotient function f , g (a)  0
g

is also continuous at x  a .

Continuity of special functions

Theorem 1. Every polynomial function, that is
f ( x)  a0 x n  an x n1  ...  an ,

is continuous at every real number.

Theorem 2. Every rational function, that is
a0 x n  a1 x n1  ...  an
f ( x) 
b0 x m  b1 x m1  ...  bm

is continuous at every real number in its domain.

Theorem 3. Every exponential function y  a x is continuous at every real number.
Theorem 4. Every logarithmic function y  loga x is continuous at every positive
real number.
Theorem 5. The trigonometric function f ( x)  sin x and g ( x)  cos x are continuous
at every real number.
Theorem 6. The trigonometric function f ( x)  tan x and g ( x)  cot x are continuous
at every real number in its domain.
Definition. A function f is said to be continuous on an interval if it is continuous
at every point of that interval.
It can be shown that all the basic elementary functions are continuous on the
intervals where they are defined.

Points of discontinuity of a function

Definition. If, for a function y  f (x) , the condition of continuity, at a point a, is
violated, a is called a point of discontinuity of the function.
For better understanding, we introduce the following more precise definition.
Definition. If, for a function y  f (x) , at least one of the conditions of continuity,
at a point a,
 1. f (a ) is defined
2. lim
xa
f ( x) exists
3. lim f ( x)  f ( a )
x a

is violated, a is called a point of discontinuity of the function.

It is clear, that the condition of continuity is violated if one of the following cases
takes place:
I. The limit lim f ( x) exists but does not coincide with f (a ) i.e. lim f ( x)  f (a) .
xa xa

f(a)

f(a) is not defined

lim f ( x)  f (a)
x a

a x a x

II. The limit entering into condition of continuity is infinite or does not exist at all
(a function must not necessarily be defined at its point of discontinuity).

f(a+0) f(a
f(a+0) )

f(a-0) f(a-0) f(a-0) =f(a)

a a a

a a

lim f ( x) lim f ( x)   lim f ( x)

x0 x0 xa

does not exist does not exist

Let us state the following definition.
Definition. If a function f (x ) possesses the left-hand and the right-hand limits at its
point of discontinuity a
i.e.  f (a  0) and  f (a  0) this point a is called point of discontinuity of the fist
kind.
Definition. All the points of discontinuity which are not the points of discontinuity
of the first kind are referred to as the points of discontinuity of the second kind.
If the limit lim f ( x) exists but the function value f(a) is different (i.e. lim f ( x)  f (a) )
xa xa

or function f is even not defined at point a we say that function f has a removable
discontinuity at a. If the limit lim
xa
f ( x) exist but the function f is not defined at

point a, we say that function f has a gap at a.

A function f has a finite jump at a if both one-sided limits f (a  0) and
f (a  0) exist and they are different. The magnitude f (a  0)  f (a  0) is called a
jump of the function at a.
Now we state the following.
Theorem. For a function f to be continuous at a point x=a it is necessary and
sufficient that f (a  0)  f (a  0)  f (a) .
Example 1. Let function f be given by formula
x 2  2x  3
f ( x) 
x 1
In this case, f is not defined at x=1, but lim f x   4 . Thus, function f has a gap at
x 1

point x=1. This is a removable discontinuity, since we can define a function g with
 f ( x) for x  1
g ( x)  
4 for x  1

Which is continuous at point x=1

f(x)
4
gap

O 1 x
Example 2. Consider a function f given by the formula
x  1 for x  0
f ( x)  
x  1 for x  0

It is clear, that f is not defined at x=0.

We have f (a  0)  xlim
0
f ( x)  1

f (a  0)  lim f ( x)  1
x0

It follows that the limit lim

x0
f ( x) , does not exist.

Thus, function f has a jump at x=0 (discontinuity of the first kind) and the jump is
equal to f (0  0)  f (0  0)  2 .

y
1
jump

O x
-1

Example 3. Consider a function f given by the formula

sin x
f ( x) 
x

We have
sin x sin x
f (0)  lim f ( x)  lim  lim 1
x  0 x  0 x x  0 x

sin x sin x
f (0)  lim f ( x)  lim  lim  1
x  0 x  0 x x  0  x
Hence, function f has a jump at x=0 (discontinuity of the first kind) and the jump
is equal to 2.
1
Example 4. The function f ( x)  is not
x
defined at x=0 and has the discontinuity
of the second kind at x=0 since
1 O
lim does not exist.
x0 x

Basic properties of functions continuous on a closed interval

Theorem 1. (Weierstrass theorem)

A continuous function on a closed interval is bounded.
i.e. f is continuous on [a,b]  f is bounded on [a,b]
Theorem 2. (Weierstrass theorem)
A continuous function on a closed interval
attains its bounds. y
In other words, this theorem means that if a
function is continuous on a closed interval there exist
at least one point at which the function assumes the
O a b х
greatest value and at least one point at which it
assumes the least value on that interval.
Theorem 3. (Bolzano-Cauchy theorem)
y
If a function f is continuous on a closed interval
and assumes value of different signs at its end points
(i.e. f (a)  f (b)  0 ) there exists at least one point lying
inside the interval at which the function turns into O a C b х

zero.

TEST
 5
1) Find the points of discontinuity of the function y 
( x  2)( x  3)

A) {2,3} B) 5 C) -3 D) {-3,2} E) 2

2) Which of the following functions is continuous on the closed interval [1,4]?

A) y 
1
B) y 
1
C) y 
1
D) y  1 E) y 
1
1 x x4 x  5x  4
2
x2 x3

3) Which of the following has no points of discontinuity on the infinite interval

(,) ?

A) y 
1
B) y 
1 C) y  sgn x D) y  1 E) y 
1
x2  4 1  e 3 x 2x 1 x5

x 2  3x
4) Find the sum of the points of discontinuity of the function y
x 2  9 x  18

A) 8 B) 6 C) 5 D) -3 E) -9

x 2  7 x  10
5) Find the product of the points of discontinuity of the function f ( x) 
x 2  6x  8
A) 5 B) -8 C) 8 D) 10 E) 7

4 x  16
6) Find the interval of continuity of the function f ( x) 
3x  6

A) (4; ) B) (;2)  (2;) C) (2;) D) (;) E) (;2)

7 x  14
7) Find the point of discontinuity of the function f ( x) 
3 x  27

A) 1 B) 4 C) 0 D) 2 E)3

5x  7
8) Find the interval of continuity of the function f ( x) 
x 2  3x

A) (;4) B) (0;) C) (;3)  (3;0)  (0;) D) (3;0)  (0;3) E) (3;)

5 x  125
9) Find the point of discontinuity of the function f ( x) 
2  7 x  98
A) -2 B) 2 C) 3 D)1 E) 0
10) Find the point of discontinuity of the function f ( x)  3 xx  18
4 8

A) 0 B) 3 C) -6 D) 2,5 E) 1,5

Lecture 5
Continuous on an interval

A function f is continuous on an interval if it is continuous at every number in the

interval. If f is defined only on one side of an endpoint of the interval, we
understand continuous at the endpoint to mean continuous from the right or
continuous from the left
Example 1.Show that the function f  x   1  1  x is continuous on an
2

interval  1,1
Solution. If  1  a  1, then using the limit laws, we have
lim
x a
f x   lim
x a
(1  1  x 2 )  1  lim
x a
1  x 2  1  1  a 2  f a 
It means, f is continuous at a
If -1<a<1.Similar calculations show that
lim
x 1
f  x   1  f  1 and lim f  x   1  f 1
x 1

So f is continuous from the right at -1 and continuous from the left at 1.Therefore,
according to definition f is continuous on an interval  1,1
Basic properties of functions continuous on a closed interval

Theorem 1. (Weierstrass theorem)

A continuous function on a closed interval is bounded.
i.e. f is continuous on [a,b]  f is bounded on [a,b]
Proof. Let I = [a, b] be a closed bounded interval, and f : I → R be continuous on I.
Suppose that f is not bounded on I. Then for each n ∈ N, there exists xn  I such

that f  xn   n . As xn  I , so xn  is bounded, by Bolzano-Weierstrass Theorem,

there exists an accumulation point of x ,

n so there exists a subsequence of
x so that x
nk nk  x . Since a  xn  b , we also have a  x  b ., i.e., xn  I .
k
  
Since f is continuous on I, we must have f xn  f  x  But this is a contradiction
k

 n
since f xn k k  k kN
,
Theorem 2. (Weierstrass theorem)
A continuous function on a closed interval
attains its bounds. y
In other words, this theorem means that if a
function is continuous on a closed interval there exist
at least one point at which the function assumes the
O a b х
greatest value and at least one point at which it
assumes the least value on that interval.

Proof: To show that f has at least one root on I  a, b ,we will show that
c  I such that f c   0 . Without loss of generality, assume
that f a   0  f b .Let I1  a1 ,b1 ,where a1  a and b1  b .Then let

p1  a1  b1 .Now if f  p1   0 , then let c  p1 and so f  p1   f c  0 and

1
2
so we have found a root .If f  p1   0 ,then it follows that either f  p1   0
or f  p1   0 . If f  p1   0 then let a2  a1 and let b2  p1 ,and if f  p1   0 then
let a2  p1 and b2  b1 .Then let I 2  a2 ,b2 . Therefore I 2  I1 and f a2   0
while f b2   0 .
Continue this process. For the k interval, if f  pk   0 then let c  pk and we` re
th

done, otherwise if f  pk   0 then let ak 1  ak and let bk 1  pk ,and if

f  pk   0 then let ak 1  pk and let bk 1  bk .Then let I k 1  ak 1 , bk 1  so that
I k 1  I k .If this process terminates at some n  N , that is f  pn   0 ,then let
pn  c .If this process never terminates, then we obtain a sequence of nested,
closed, and bounded intervals I1  I 2  ...  I n  ... such that n  N we have
that f an   0 and f bn   0 .So by the Nested Intervals Theorem, for all
n  N there exists a c  I n .Therefore an  c  bn for all n  N .Now the length of
ba ba
each interval I n  bn  an  n 1 and so lim b  a   lim  0 which
2 n1
n n
n  n 
2
implies that lim an  lim bn . Therefore by The Squeeze Theorem, since
n  n 

an  c  bn we have that lim an  c  lim bn

n  n 

Now since f is continuous at c and since an  and bn  are sequences from the
domain I we have by the Sequential Criterion for the Continuity a Function that
lim
n 
f an   f c   lim
n 
f bn  .Since f an   0 for all n  N then we have that

lim
n 
f an   f c   0 .Similarly, since f bn   0 for all n  N then we have

that lim
n 
f bn   f c   0 .Combining these inequalities we get that lim f c   0 ,
n 

and so c  I is a root of f.

Bolzano`s Intermediate Theorem : Let I be an interval and let f : I  R be a

continuous function. If a, b  I and k  R is such that f a   k  f b , then
there ,where c is between a and b such that f c   k .
Proof: Let I be an interval and let f : I  R be a continuous function. First
consider the case where a  b , and define g x   f x   k .Then we have that
g a   f a   k  0 and g b  f b  k  0 .Therefore g a   0  g b.By the
theorem 3, there exists a c, such that a  c  b where 0  g c  .But
g c   f c   k and so 0  f c  k , in other words k  f c  .
Now consider the case where a  b ,and define g x   k  f x  .Therefore
g b  k  f b  0 and g a   k  f a   0 .Thus g b  0  g a , and once
again by the theorem 3 there exists a c, such that b  c  a where 0  g c  .But
g c   k  f c  and so 0  k  f c , in other words f c   k .
Corollary 1: Let I be an interval and let f : I  R be a continuous function. If
k  R satisfies inf f I   k  sup f I  , then there there exists a c such that
f c   k .
Uniform Continuity
First recall the definition of f being continuous at x0 :   0 ,
  0 x : x  x0    f  x   f  x0    .
In general, δ depends on both  and x0 , as function changes rapidly at some
points and flat at some other points. We start some examples to look into this.
Example. Let f : R  R and f x   2 x . Let x0  R . Consider
f  x   f  x0   2 x  2 x0  2 x  x0

From this we can see if we choose   , we
2
have x  x0    f  x   f  x0    In this case, δ depends only on  , it works
for all x0  R .

Example Let f : 0,   R f  x  

1
. Let x0  u  0 .
x
Consider
xu
f  x   f u  
xu
u u 3u
As x  u , so consider only x  u  , i.e.,  x  .
2 2 2
1 2 1  1  2  2
Then  . Hence      . Now given  > 0, choose
x u ux  u  u  u 2
 u u 2 
  min  ,  . So when x  u    f  x   f u    .
 2 2 
Here δ depends on both  and u. In fact, there is no δ for all u > 0, as then δ = 0.
See the graph of f  x  
1
.
x
Definition. Let f : D → R is uniformly continuous on E ⊂ D iff ∀  > 0∃δ > 0 ,
∀x, y ∈ E , x  y    f  x   f  y    ⇒. If f is uniformly continuous on D,
then f is uniformly continuous.
Remark. f uniformly continuous on E implies f is continuous on E. The converse
is not true.

Example 1. f : [2.5, 3] → R defined by f x  

3
.
x2
2. f : (0, 6) → R with f  x   x  2 x  5
2

3. f : (2, 3) → R with f x  
3
.
x2
Non-uniform Continuity Criteriia Let A ⊆ R and let f : A → R. Then the
following statements are equivalent.
1. f is not uniformly continuous on A.
2. ∃  0 > 0 such that for every δ > 0 there are points x , y  A such that
x  y   and  f  x   f  y    0 .
3. ∃  0 > 0 and two sequences xn and yn  in A such that lim xn  yn   0 and

f  x   f  y    0 for all n ∈ N.
Theorem Let I be a closed bounded interval and let f : I → R be continuous on I.
Then f is uniformly continuous on I.
Proof. If f is not uniformly continuous on I. From the above, ∃  0 > 0 and
xn , yn  I such that xn  yn  0 and f  xn   f  yn    0 for all n. As I is
 
bounded, so by Bolzana-Weierstrass, there is a subsequence xn k of xn that
converges to z ∈ I, as I is closed interval. In addition, from
y n  z  y n  xn  xn  z
y n  z as well. Now as f is continuous at z, so we have f xn   f  z 
k k k k

 
k k

and f y n  f  z  But this is a contradiction, as f  xn   f  y n    0 for all n

k

Lecture 6
Derivative

The concept of derivative is at the core of Mathematics. Let a function y  f (x) be

y
f ( x  x)

y
f (x )
x

x x  x x
given. If the independent variable x is given an increment x , the function y
receives the corresponding increment y  f ( x  x)  f ( x).
Definition. The limit of the ratio of the increment of a given function y  f (x) to
the increment of the independent variable as the latter tends to zero is called the
derivative of that function (provided that this limit exists):
y f ( x  x)  f ( x)
f / ( x)  lim  lim
x 0 x x  0 x

Notation. The derivative of a function with respect to a variable x is denoted either

df
f / ( x) or (read ‘f dash x’ or ‘f prime x’)
dx
Definition. The operation of finding a derivative is called differentiation.

Derivatives of some functions

Let us find the derivatives of some elementary

functions. C y f(x)=C

1) f ( x)  C , where C is an arbitrary constant. o x  x

x x
It is clear that y  f ( x  x)  f ( x)  C  C  0
y
Therefore f(x)=x
x  x
y 0
f / ( x)  (C ).  lim  lim  lim 0  0
x 0 x x  0 x x0
x
Thus,
That is, the derivative of a constant is equal to zero. o x x  x x

2) f ( x)  x
We have y  f ( x  x)  f ( x)  x  x  x  x
Hence,
y x
f / ( x)  ( x) /  lim  lim  lim 1  1
x  0 x x  0 x x  0

.
Thus, ( x ).  1
That is, the derivative of an independent variable is equal to one.
3) Consider a linear function f ( x)  kx  b where k y

and b are constant. We can find an expression for

y
y y  f ( x  x)  f ( x)  k ( x  x)  b  kx  b  kx .
f(x)=kx+b
It follows that x
o x
y kx x x  x
f / ( x)  (kx  b) /  lim  lim  lim k  k
x  0 x x  0 x x  0

Thus, the derivative of a linear function

y
f ( x)  kx  b is equal to k, that is f(x)=x2
(kx  b) /  k
y
4) f ( x)  x 2
We have o x x  x x
y  f ( x  x)  f ( x)  ( x  x) 2  x 2  2 x  x  (x) 2

Therefore
y x(2 x  x)
f / ( x)  ( x 2 ) /  lim  lim  lim (2 x  x)  2 x
x 0 x x  0 x x 0

Thus, the derivative of f ( x)  x 2 is equal to 2x that is ( x 2 ) /  2 x

5) Now let us consider a power function f ( x)  x n for n N where N is a set of all
positive integers and prove that its derivative is equal to nx n1
That is ( x n ) / .  nx n1

To this end, using the well-known Newton’s binomial formula we find an

expression for y
n(n  1) n2
y  f ( x  x)  f ( x)  ( x  x) n  x n  x n  nxn1  x  x (x) 2  ... 
2!
 n(n  1) n2 
 (x) n  x n  x nxn1  x x  ...  (x) n1 
 2! 

It follows that
y n(n  1) n  2
 nxn 1  x x  ...  (x) n 1
x 2!

It is clear, that the first term on the right-hand side of the last equality does not
depend on x and the others tend to zero as x  0 . Therefore,
 y
f / ( x)  ( x n ) /  lim  nxn1
x0 x

Thus, for every positive integer n, the power function f ( x)  x has the derivative
n

nxn1 , which is what we wished to prove. It can be shown that the formula
( x n ) /  nxn1 is valid when n is an arbitrary real number.

6) Now let us consider a trigonometric function f ( x)  sin x and prove that the
derivative of sin x is equal to cos x , that is
(sin x) /  cos x

First, we find an expression for y . We have

x  x 
y  f ( x  x)  f ( x)  sin( x  x)  sin x  2  sin  cos x   .
2  2 

Therefore,
x  x  x
2  sin  cos x   sin
y 2  2  2  lim cos x  x   cos x
f / ( x)  (sin x) /  lim  lim  lim
x0 x x0 x x 0  x x0
 2 
2

Thus, (sin x) /  cos x ,

which is what we had to prove.
7) Let us consider another trigonometric function f ( x)  cos x and prove that the
derivative of cos x is equal to  sin x that is
(cos x) /   sin x

Since y  f ( x  x)  f ( x)  cos(x  x)  cos x  2  sin x  sin x  x 

2  2 

We have
x  x  x
2  sin  sin x   sin
y 2  2    lim 2  lim sin x  x    sin x.
f / ( x)  (cos x) /  lim  lim 
x 0 x x 0 x x 0 x x 0  2 
2

Thus, (cos x) /   sin x

and we get what we had to prove.
8) f ( x)  a x
We have y  a xx  a x  a x (a x  1)
It follows that
y a x  1 a x  1
f / ( x)  (a x ) /  lim  lim a x   a x  lim  a x  ln a
x0 x x0 x x0 x

Thus, the derivative of a x is equal to a x ln a ,

That is, (a x ) /  a x ln a .
In particular, we have (e x ) /  e x

Geometrical meaning of the derivative

Let us consider the graph of a function y  f (x) and a point P( x0 , f ( x0 )) on it. Next
we will consider a nearby point Q( x, y ) which is also on the graph. Let us join the
point Q to the point P with a straight line which is called a secant.

y y

f (x ) Q Q1
y Q2

Q3
P P
f ( x0 )
 
x
o x0 x x x0 x

We can calculate the gradient (slope) of the secant.

y
Putting x  x  x0 and y  f ( x)  f ( x0 ) , we see that  tan  , i.e. the ratio is equal
x

to the gradient of the secant. If x  0 the point Q gets closer and closer to the
point P . In other words, if x  0 then the secant turns round the point P and tends
to the limiting position.
Now we can state the definition.
Definition. The tangent to the graph of a function at its given point is the straight
line occupying the limiting position of the secant when the points of intersection
merge.
y
Therefore f / ( x0 )  lim  tan  .
x0 x
That is the geometrical meaning of the derivative of a function is that it is equal to
the slope of the tangent.
Theorem.
The value of the derivative f / ( x0 ) is equal to the slope of the tangent to the graph of
the function y  f (x) at the point with abscissa x0:
f / ( x0 )  tan  ,

where  is an angle formed by a tangent and the positive direction of x  axis .

Equation of a tangent
We have if a function y  f (x) has a derivative f / ( x0 ) at a given point x0 then there
exist a tangent drawn to the graph of the function at the point P( x0 ; f ( x0 )) with the
gradient f / ( x0 ) .
We now look at how to find the equation of this tangent. The tangent is a straight
line with the gradient f / ( x0 ) , so its equation is given by y  f / ( x0 ) x  b . Since the
tangent passes through the point P we have
f ( x0 )  f / ( x0 ) x0  b

hence
b  f ( x0 )  f / ( x0 ) x0

It follows that the equation of a tangent has the form

y  f / ( x0 ) x  f ( x0 )  f / ( x0 ) x0

y  f ( x0 )  f / ( x0 )(x  x0 )

Normal. Equation of a normal

The normal to a curve at a given point is a straight line which crosses the curve at
that point and is perpendicular to the tangent at that point.
Since the tangent and normal are perpendicular to each other, if the gradient of the
1
tangent is f / ( x0 ) , then the gradient of the normal is  /
. Therefore the
f ( x0 )

equation of the normal has the form

1
y  f ( x0 )  /
( x  x0 )
f ( x0 )

or
1
y  f ( x0 )   /
( x  x0 )
f ( x0 )

Basic differentiation rules

Differentiation and Arithmetic Operations
Suppose the functions u  u (x) and v  v(x) have derivatives u / ( x ) and v / ( x) .
Theorem 1. The derivative of a sum of two functions is equal to the sum of the
derivatives of that functions.
(u  v) /  u /  v /

Proof. Let y be a function represented as the sum of given functions u and v :

y  uv

If y ( x)  u ( x)  v( x) then y ( x  x)  u ( x  x)  v( x  x) .

If the independent variable x is given an increment x the functions u , v gain the
corresponding increments u , v and the function y the increment y . It is clear
that
y  y( x  x)  y( x)  [u( x  x)  v( x  x)]  [u ( x)  v( x)] 
[u ( x  x)  u( x)]  [v( x  x)  v( x)]  u  v

that is the increment of a sum is equal to the sum of the increments:

y  (u  v)  u  v

On dividing the latter relation by x we get

y u v
 
x x x
Now, passing to the limit as x  0 we obtain
 y  u v  u v
y /  lim  lim    lim  lim  u /  v/
x0 x x0 x x  x0 x x0 x

Thus, y /  (u /  v / )  u /  v / , which is what we had to prove.
Similarly can be proved the following
Theorem 2. The derivative of a difference is equal to the difference of the
derivatives
(u /  v / )  u /  v /

Now let us prove the following.

Theorem 3. The derivative of the product of two functions is equal to the sum of
the product of the derivative of the first function by the second function and the
product of the derivative of the second function by the first function.
(uv) /  u / v  v / u

Proof. Let a function y be represented as a product of two functions u and v: y=uv

Then
y  y ( x  x)  y ( x)  u ( x  x)v( x  x)  u ( x)v( x) 
 [u ( x  x)  u ( x)]v( x  x)  [v( x  x)  v( x)]u ( x) 
u  v( x  x)  v  u ( x)  u  [v( x)  v]  v  u ( x)  u  v( x)  v  u ( x)  u  v

That is
y  u  v  v  u  u  v

On dividing the latter relation by x we get

y u v u v
 v  u   x
x x x x x

and passing to the limit as x  0 we obtain

y u v u v
y /  lim  lim  v  lim  u  lim   x 
x 0 x x  0 x  x  0 x x  0 x x
u v u v
 v  lim  u  lim  lim  lim  lim x 
x 0 x x 0 x x 0 x x 0 x x 0

 v u/  u  v/  u/  v/  0  u/  v  v/ u
Thus, y /  (uv) /  u / v  v / u
which is what we wished to prove.
Corollary. A constant factor can be taken outside the derivative sign, i,e.
y /  (Cu ) /  C  u

where C – constant.
Theorem 4. The derivative of the quotient of two functions is equal to the fraction
whose denominator is equal to the square of the divisor and the numerator is equal
to the difference between the product of the derivative of the dividend by the
division and the product of the dividend by the derivative of the divisor

 u  u v u v
/ /
 
v v2

u
Proof. Let y  . Then
v
u ( x  x) u ( x) u ( x)  u u ( x) u  u u u  v  u  v
y  y ( x  x)  y ( x)       
v( x  x) v( x) v( x)  v v( x) v  v v v  (v  v)

u  v  u  v
That is, y 
v  (v  v)

On dividing the latter relation by x we get

u v
v u
y
 x x
x  v 
v v   x 
 x 

and passing to the limit as x  0 we obtain

u v
v u 
y x  u v  uv  u v  uv
/ / / /
y /  lim  lim x
x0 x x0  v  v(v  v /  0) v2
v v   x 
 x 

u v  uv
/

Thus, y    
/ /
/ u
v v2

which is what we had to prove.

Derivative of a Composite Function
Let y  f (u ) and u  g (x) , then the composite function of g and f is
( fog)( x)  f ( g ( x))
Theorem. The derivative of a composite function y  f ( g ( x)) can be found by the
formula
y /  [ f ( g ( x))]/  f / ( g ( x))  g / ( x)

or y /  f / (u )  u /  f / (u )  g / ( x)
that is, the derivative of a composite function y  f ( g ( x)) is equal to the product of
the derivative of the given function f / (u ) with respect to the intermediate argument
u by the derivative of this argument u /  g / ( x) with respect to the independent
variable x.
Proof. If x receives an increment x then the intermediate variable u receives an
increment u and therefore y receives an increment y too. We have
y y u
  (1)
x  u x
u u
Now let x  0 . Then  u x/ , i.e. lim  g / ( x)
x x0 x

u
and hence u   x  u x/  0  0 i.e. u  0 .
x
y y
Therefore  yu/ , i.e. lim  f / ( x)
u u  0  u

Now, passing to the limit as x  0 in formula (1) and taking into account that if
x  0 then u also tends to zero, u  0 , we obtain
y y u
y /  lim  lim  lim  f / (u )  g / ( x)
x0 u u 0 u u 0 x

which is what we wished to prove

Derivative of an Inverse Function

Suppose the function y  f (x) has the inverse x  f 1 ( y ) . We will prove that if the
derivative of the original function y  f (x) is known it is easy to determine the
derivative of the inverse function x  f 1 ( y ) .
Theorem. The derivative of the inverse function can be found by the formula
 df 1 ( y ) 1 1
 or x y/ 
dy df ( x) yx/
dx
Proof. Indeed, we have
x 1

y y
x
and passing to the limit as x  0 and taking into account that if x  0 then y also
tends to zero, y  0 , we obtain
x 1 1
x y/  lim   /
y 0 y x y x
lim
x0 y

which is what we had to prove.

Derivative of a Function Specified Parametrically

Now we will study a method of how to find the derivative of a function specified
by parametric equations.
Let y as function of x be specified parametrically by equations
x   (t ) , y   (t ) .

Theorem. The derivative of a function specified parametrically x   (t ) , y   (t )

can be found by the formula
 / (t ) yt/
y x/  or y /

 / (t ) x
xt/

Proof. Indeed, according to the differentiation rule for composite functions we

have
yx/   / (t )  t x/

And the derivative t x/ is found with the aid of the differentiation rule for an inverse
function
1 1
t x/   /
xt  (t )
/

Therefore, finally, we have

 1  / (t ) yt/
yx/   / (t )   
 / t   / t  xt/

which is what we wished to prove.

Derivatives of inverse trigonometric functions
 
1. y  arcsin x, x  sin y ,  y , 1  x  1
2 2
Applying the differentiation rule for an inverse function, we get
1 1
y x/  (arcsin x) /  /

x y cos y

 
Since  y , then cos y  0 and hence, cos y  1  sin2 y  1  x 2 . Thus, finally,
2 2
we get
1
(arcsin x) / 
1  x2
2. y  arccosx, x  cos y , 0  y   , 1  x  1

Similarly, we have

1 1
y x/  (arccos x) /  /

x y sin y

Since 0  y   , then sin y  0 and therefore,

sin y  1  cos2 y  1  x 2 . Thus, we obtain

1
(arccosx) /  
1  x2
 
3. y  arctan x, x  tan y ,   y ,    x  
2 2
Using the formulas for the derivatives of an inverse function and of the tangent we
obtain

1 1 1 1
y x/  (arctan x) /    cos2 y  
x y/ 1 1  tan y 1  x 2
2

cos2 y
Thus,
1
(arctan x) / 
1  x2
4. y  arc cot x, x  cot y , 0  y   ,    x  

Similarly,
1 1 1 1
y x/  (arc cot x) /     sin 2 y   
x y/ 1 1  cot y 1  x 2
2

sin 2 y

Thus,
1
(arc cot x) /  
1  x2

Derivative of a logarithmic Function

We have y  loga x, x  a y , a  0, a  1, x  0,    y  
Applying the differentiation rule for on inverse function and the formula for the
derivative of an exponential function we get
1 1 1
y x/  (loga x) /  /
 y 
x y a ln a x ln a

Thus,
1
(loga x) / 
x ln a
In particular, when a  e we have
1
(ln x ) / 
x

Derivative of a Power Function

According to the formula of the derivative of a composite function we have

  
/
 /
( x n ) /  (e ln x ) n  e n ln x  e n ln x  n
1
x
1
 x n  n  nxn1
x
Thus,
( x n ) /  nxn1

Table of Derivatives
 Function Derivative Function Derivative
y  f (x) y /  f / ( x) y  f (x) y /  f / ( x)

C 0 ax a x ln a

x 1 ex ex

x2 2x loga x 1
x ln a

xn nxn-1 ln x 1
x
1 1 arcsin x 1

x x2 1  x2

x 1 arccos x 1

2 x 1  x2

sin x cos x

cos x  sin x arctan x 1

1  x2

tan x 1 arccot x 1

cos2 x 1  x2

cot x 1

sin 2 x

Differential
Suppose a function y  f (x) has the derivative at some point x
y
lim  f / ( x)
x0 x
This implies
y
 f / ( x)  
x

Where    (x) is an infinitesimal as x  0 , i.e lim  (x)  0

x0

If follows that
y  f / ( x)x    x
Definition. A linear function f / ( x)x )( f / ( x) -constant) is called the differential of
the function y  f (x) and is denoted dy or df (x) .
In other words
dy  df ( x)  y /  x  f / ( x)  x  f / ( x)dx (dx  x)

Thus, the differential of the function is the product

dy  f / ( x)dx

It is clear, that the existence of the differential at a point x is equivalent to the

existence of the derivative f / ( x) at that point x.

Geometrical Meaning of the Differential

Let us discuss the geometrical meaning of the differential of a function y  f (x) .

Since f / ( x)  tan  the differential dy  f / ( x)dx is equal to the length of the line
segment, BC.

y
C
y
y  f (x) A  dy
C B
dy
A 


x  dx dx

o x x  dx x o x  dx x
x

Thus, the geometrical meaning of the differential of a function: the differential is

equal to the increment of the ordinate of the tangent.

Differentiability
Definition. A function y  f (x) is said to be differentiable at a point x if it
possesses the differential (or the derivative) at that point.
In short, a function which has the differential (or the derivative) is called
differentiable.
Theorem. If a function y  f (x) is differentiable at some point, it is continuous at
that point.
Proof. Let y  f (x) be differentiable at a point x. It means that
y  f ( x  x)  f ( x)  f / ( x)  x    x , where lim   0
x0

Now, passing to the limit as x  0 , we get

lim y  lim f ( x  x)  f ( x)  f / ( x)  lim x  lim   lim x  0
x0 x0 x0 x0 x0

This yields
lim y  0 or lim f ( x  x)  f ( x)
x0 x0

That is, the function y  f (x) is continuous at x. The theorem is proved.

Thus, to put it briefly, if a function y  f (x) is differentiable it is continuous. It
should be noted that the converse statement is not true since there are examples of
continuous functions which do not possess derivatives at some points.
For instance, the function y  x is continuous throughout Ox but has no derivative
at the point x=0

Properties of Differential
The differential of a function is obtained by multiplying its derivative by the
differential of the argument and therefore each property of the derivative
obviously implies the corresponding property of the differential.
In short,
d (u  v)  du  dv

d (u  v)  du  dv

d (uv)  vdu  udv

d (Cu )  Cdu

 u  vdu  udv
d  
v v2
 TEST
1) Find derivative y / , if y  e  x (cos x  sin x)
A)  2e  x sin x B) e  x sin x C) e x cos x D) e  x E) cos x

1  x2
2) Find derivative y / , if y  arccos
1  x2
2 1 1  x2 D) sin x 2 E) cos x 2
A)
1 x2
B)
1  x2
C) y 
1  x2

 2  tgx
3) Find y /   , if y  ln
3 2  tgx

A) 1 B) 2 C) 14 D) 15 E) 16
x
4) Find derivative y / , if y  ln
1  x2

A)
1 B) C)
1 D) 0 1 x 2
x 1  x2 E)
1 x
x(1  x 2 )


5) Find f /   , if f ( x)  sin7 x  sin5 x
2
A) 0 B) 1 C) 2 D) 3 E) 4

6) Find derivative y / , if y  2 x sin x  x 2 cos x

A) 2xcos x B) 2xsin x C) (2 x  x 2 ) sin x D) x 2 cos x E) (2  x) sin x

1 x6
7) Find derivative y / , if y  ln
12 x  6
1 1 1 1 1
A) x B) C) D) E)
12 x  16
2
x  36
2
x  16
2
x6
8) Find tangent to the graph of the function f ( x)  x , if x0  4
A) y  x B) y 
x C) y  x  4 D) y   1
x
E) y 
x
4 4 5
9) Find the normal to the graph of the function f ( x)  ln 2 x, if x0  e
e
A) y  1  ( x  e) B) y  x  e C) y  1  x  e
2
e
D) y  2  e  5 E) y  5  ( x  5)
2

10) Find the differential of the function y  cos x 4  5 x3  2

A) dy  (15x 2  sin x 4 )dx B) dy  (4 cos x 3  15x 2 )dx C) dy  15x 2
D) dy  (15x 2  4 x 3 sin x 4 )dx E) dy  (cos x 4  15x 2 )dx
Lecture 7
Basic theorems of differential calculus
Fermat`s Theorem. Let y  f (x) be a continuous function in a closed
interval [ x1 , x2 ] assuming its greatest (or least) value at an interior point  of that

interval: x1    x2 .Then, if the derivative f / ( x) of function f (x) at the point 

exists it is necessarily equal to zero: f / ( )  0

Proof. For definiteness, let the function f (x) assume its greatest value at the
point  ;this means that f ( x)  f   (1) for all x[ x1 , x2 ]
The value of the derivative at the point  is
f (  x)  f ( )
f    lim
x 0
x
Where the increment x can be positive or negative since  is an interior
point of the interval.
Let us consider the ratio
f (  x)  f ( )
x
According to condition (1), its numerator cannot be
f (  x)  f    0 .Therefore, this ratio is nonpositive for x  0 :
positive:
 f (  x)  f ( )
 0 if x  0
x
The limit of the ratio, as x  0 , x  0 , is the right- hand derivative of
f (x) at the point  and also nonpositive: f r ( ) .But if x  0 the ratio is
/

nonnegative:
f (  x)  f ( )
 0 if x  0
x
Consequently, for the left- hand derivative of f (x) we have f l ( ) .
/

By the hypothesis, the function f (x) has the derivative f / ( x) at the point  ,
and therefore the right- hand and the left- hand derivative must coincide.This
being only possible if f r ( ) = f l ( )  0 ,it follows that f ( )  0 ,which is what
/ / /

we wished to prove.

Rolle’s Theorem. If a function f (x) is continuous in a closed interval [a, b] ,

is differentiable at all its interior points and assumes equal values at the end points
of the interval: f (a)  f (b) , then there is at least one point x  c  (a, b) , i.e. lying
inside that interval at which f / (c)  0 .
Proof. It is known that a function continuous in a closed interval attains its
greatest and least values on that interval. Suppose M  max f ( x) , m  min f ( x) , then
m  f ( x)  M for all x  [a, b] . There are two possible cases here: m  M and m  M .
(i.e. m  M ). If m  M then f (x) is constant and hence f / ( x)  0 for all x  (a, b) . As
a point c in this case we can take any point of the interval ( a, b ) .
If m  M then from the condition f (a)  f (b) it follows that at least one of the
values m or M is not accepted on the ends. For definiteness, let the function
f (x) assumes its greatest value M inside the interval ( a, b ) at some point c  (a, b) ,
i.e. f (c)  M . It follows that f (c  x)  f (c)  0 for any x , and hence
f (c  x)  f (c)
 0 for x  0
x
 f (c  x)  f (c)
 0 for x  0
x
Now, passing to the limit as x  0 in the last two inequalities, we get
f (c  x)  f (c)
lim  f / (c)  0 for x  0
x  0 x
f (c  x)  f (c)
lim  f / (c)  0 for x  0
x  0 x

It follows that f / (c)  0 . Thus, inside the interval (a,b) there is a point c at
which the derivative is equal to zero.
Now let us discuss the geometrical meaning of the Rolle’s theorem.

y y

o a c b x o a b x

Rolle’s theorem is interpreted geometrically as follows: on the graph of a

function continuous in a closed interval and differentiable inside the interval,
assuming equal values at the end points of the interval, there is at least one point at
which the tangent is parallel to the x-axis.
Lagrange’s Theorem. If a function f (x ) is continuous in a closed interval
[a, b] and is differentiable inside the interval ( a, b ) there is at least one point
x  c  (a, b) such that

f (b)  f (a)
 f / (c ) .
(b  a)

Proof. Let consider the auxiliary function

F ( x )  f ( x )  x
 And determine the number  so that F (a )  F (b) .
We have f (a)  a  f (b)  b
f (b)  f (a )
Then   .
ba

It is clear that the function F (x ) satisfies all the conditions of Rolle’s

theorem. Indeed, F (x ) is continuous in [a, b] and is differentiable in ( a, b ) , and
assumes equal values at the end points of the interval: F (a )  F (b) .
Therefore, by Rolle’s theorem, there exist a point c  (a, b) such that F / (c)  0 .
But F / ( x)  f / ( x)   , and hence F / (c)  f / (c)    0, i.e. f / (c)   .
Thus, finally, we get
f (b)  f (a)
 f / (c )
ba

which is what we wished to prove.

Not let us discuss the geometrical meaning of the Lagrange’s theorem. Let
A  (a, f (a )) and B  (b, f (b)) be the ends of the graph of a function f. let AB be a
chord connecting the points A and B.

y y B
B

A A

o a c b x o a b x

Lagrange’s theorem is interpreted geometrically as follows: On the graph of a

function continuous in a closed interval and differentiable inside the interval, there
is at least one point at which the tangent is parallel to the chord AB.
It follows Lagrange’s theorem that
f (b)  f (a)  f / (c)(b  a), a  c  b .

This relation is known as the formula of finite increments.

 Cauchy’s Theorem. If f (x ) and g(x) are continuous functions in a closed
interval [a, b] , differentiable inside the interval (a,b) and if g / ( x)  0 for all points
x  (a, b) there exist at least one point x  c  (a, b) such that
f (b)  f (a) f / (c)
 .
g (b)  g (a) g / (c)

Before to prove the theorem, it should be noted that g (b)  g (a)  0 since, if
otherwise, there would exist, according to Rolle’s theorem, c  (a, b) at which
g / (c)  0 and hence contradicts the hypothesis.

Proof. Let us consider the auxiliary function

F ( x )  f ( x )  g ( x )

and determine the number  so that F (a )  F (b) , that is f (a)  g (a)  f (b)  g (b) .
f (b)  f (a)
It follows that  
g (b)  g (a)

It is obviously, the auxiliary function F (x ) satisfies all the conditions of

Rolle’s theorem. Consequently, by Rolle’s theorem, there is a point c  (a, b) such
that F / (c)  0 , but F / ( x)  f / ( x)  g / ( x) , and hence F / (c)  f / (c)  g / (c)  0 . It
follows that
f (b)  f (a) f / (c)
 
g (b)  g (a) g / (c)

Which is what we had to prove.

It should be noted that, in the particular case, when g ( x )  x . Cauchy’s
theorem gives same result as Lagrange’s theorem.
Examples:
1.Does Rolle`s theorem hold for the function f x   x 2  6 x  100 ,if
a  1, b  5 ?,At what value c?
Solution. Since function f (x) is continuous and differentiable for all values x
and its values at the ends of interval [1,5] are equal: f 1  f 5  95 ,then Rolle’s
theorem is fulfilled on this interval. f x  2x  6
f x  2 x  6  0  c  3
2. Does Rolle`s theorem hold for the function f  x   3 8 x  x 2 ,if a  0, b  8 ?,At
what value c?
Solution. Since function f  x   3 8 x  x 2 is continuous for all values x and
8  2x
has the derivative f / ( x)  for x  0, x  8 i.e differentiable
33 8 x  x 2 
2

in 0,8 . Besides, f 0  f 8  0 . then Rolle’s theorem is fulfilled on this interval.
8  2x
f / ( x)  0 xc4
3 8 x  x
3

2 2

3.On the arc AB of the curve y  2 x  x 2 , find a point M at which the tangent is
parallel to the chord AB, if A1;1, B3;3 .
Solution. Function y  2 x  x 2 is continuous and differentiable for all values x.
By Lagrange`s theorem, between two values a  1, b  3 there exists a value
x   satisfying the equality yb  ya   b  a yc  ,where y  2  2 x
y3  y1  3  1yc
2  3  3   2  1  1   3  1  2  2 
2 2

 4  41       2, y2  0
Thus, point M has coordinates 2;0 
LECTURE 8

L’Hospital - Bernoulli’s Rule

Suppose we need to find the limit of the form

f ( x)
lim
xa g ( x)

as x tends to some point.

In many cases, finding the limit is reduced to a formal substitution of the
corresponding value instead of argument in the formula that defines the considered
function. However, in many cases, this process leads to expressions of the form
0 
, , 0  ,   , 1 , 0 0 ,  0 . These expressions are called indeterminates because
0 

they cannot give us information about the limit.

0
Indeterminate forms of the type
0

Theorem. Let f (x) and g (x) be two functions such that lim f ( x)  0 and
x a

lim g ( x)  0 .
xa

If f (x) and g (x) are differentiable at the point a (i.e. their derivatives
f / (a ) and g / (a) exist) and g / (a)  0 then

f ( x) f / (a)
lim 
x a g ( x) g / (a)

Proof. Since f (x) and g (x) are differentiable at the point a then they are
continuous at that point. It means that lim f ( x)  f (a) and lim g ( x)  g (a) .
xa xa

Consequently, f (a )  g (a )  0 and therefore

f ( x)  f (a )
f ( x) f ( x)  f (a ) xa
 
g ( x) g ( x)  g (a ) g ( x )  g (a)
xa
Now, passing to the limit as x  a and using the theorem on the limit of a
quotient we get
f ( x)  f (a )
lim
f ( x ) x a xa f / (a)
lim   / .
x a g ( x ) g ( x)  g (a ) g (a )
lim
x a xa
which is what we had to prove.

L’Hospital - Bernoulli’s theorem. Let f (x ) and g (x) be two functions such

that
lim
x a
f ( x)  0 and lim g ( x)  0
xa

f / ( x) f ( x)
If the limit lim /
xa g ( x)
exists then the limit lim
xa g ( x )
also exist and
 f ( x) f / ( x)
lim  lim /
x a g ( x ) x a g ( x )

Examples.
sin x  0  (sin x) / cos x
1. lim     lim  lim  cos0  1.
 0  x 0 ( x )
x 0 / x 0
x 1

x 2  4 x  12  0  ( x 2  4 x  12) / 2x  4
2. lim     lim  lim  4.
x 2 x  6 x  8  0  x2 ( x  6 x  8)
2 2 / x 2 2 x  6

1
2x 
x 2  1  ln x  0  ( x 2  1  ln x) / x  3.
3. lim     lim  lim
x 1 e e
x
 0  x1 (e  e)
x / x 1 e x
e

1
ln(1  x)  0  (ln(1  x)) /
4. lim     lim  lim 1  x  1.
  1
x 0 x0 / x0
x 0 ( x )

It sometimes happens that using L’Hospital - Bernoulli’s rule we get a ratio of

0
derivatives which is again an indeterminate form of the . In such cases it is
0

possible to apply the rule again and so on.

For example
e 3 x  3x  1  0  3e 3 x  3  0  9e 3 x 9
5. lim     lim     lim   0.18
 0  x0 5 sin10x  0  x0 50 cos10x 50
x 0 2
sin 5 x

Here we have applied L’Hospital - Bernoulli’s rule two times.


Indeterminate forms of the type


L’Hospital - Bernoulli’s Theorem. Let f (x ) and g (x) be two functions such

that
lim f ( x)   and lim g ( x)  
x a xa

f / ( x) f ( x)
If the limit lim /
exists then the limit lim also exist and
xa g ( x) xa g ( x )

f ( x) f / ( x)
lim  lim /
x a g ( x ) x a g ( x )

Examples.
 1

1. lim 2     lim 2 /  lim x  lim 2  0.
/
ln x (ln x) 1
x  x
   x  ( x ) x  2 x x  2 x

1

2. lim n     lim xn1  lim n  0. for n  0.
ln x 1
x  x
 x  nx x  nx

xn    nx n1    n(n  1) x n2    n(n  1)(n  2)...1

3. xlim  
  x x
lim  
  x
lim     ...  lim 0
  
x x
 e e e x ex

If n is positive integer.
Here we have applied L’Hospital - Bernoulli’s rule n times. Thus, L’Hospital -
Bernoulli’s rule can be used repeatedly as many times necessary.
L’Hospital - Bernoulli’s rule can be applied to some indeterminate forms of other
types such as
0  ,   , 1 ,  0 , o 0

0 
after they are transformed into forms of the type or .
0 

Indeterminate forms of the type 0   .

Let f (x ) and g (x) be two functions such that
lim f ( x)  0 and lim g ( x)   .
x a x a

Suppose we need to find the limit lim f ( x) g ( x) which corresponds to the case
xa

0   . This indeterminate form can be reduced to any of the following

f ( x)  0 
lim f ( x) g ( x)  lim  
x a x a 1 0
g ( x)

g ( x)   
lim f ( x) g ( x)  lim  
x a xa 1 
f ( x)

and then it is easy to apply the corresponding L’Hospital - Bernoulli’s rule.

Examples.
 1
/
ln x (ln x ) 1
1. lim x 2 ln x  lim  lim /
 lim x   lim x 2  0
x 0 x 0 1 x 0 x 0 2 2 x 0
 1   3
x 2   x
x 
2

x ( x) / 1 1 1
2. lim x  cotx  lim  lim  lim  lim cos2 x 
x 0 x 0 tan x x0 (tan x) / x0   x 0 
cos x
2

Indeterminate forms of the type 

Let f (x ) and g (x) be two functions such that

lim f ( x)   and lim g ( x)   .
x a x a

Suppose we want to find the limit lim f ( x)  g ( x) which corresponds the case
x a

. We can transform this limit as follows

  1 1
 1  
g ( x) f ( x)  0 
lim f ( x)  g ( x)  lim
1
   lim  .
x a xa
 1 1  x a 1

1 0
 f ( x) g ( x)  f ( x) g ( x)

and hence, it is possible to apply L’Hospital - Bernoulli’s rule.

Example.
1 1  ex 1 x (e x  1  x ) / ex 1
lim  x   lim  lim  lim 

x 0 x e  1  x0 x(e x  1) x0 ( x(e x  1)) / x 0 e x  1  xe x

(e x  1) / ex 1
 lim  lim 
x 0 (e x  1  xe x ) / x 0 e x ( 2  x ) 2

 x 1  x ln x  x  1 ( x ln x  x  1) / ln x
lim    lim  lim  lim 
x 1 x  1
 ln x  x1 ( x  1) ln x x 1 (( x  1) ln x ) / x 1 x 1
ln x 
x
1
(ln x) / 1
 lim  lim x 
x 1
/ x 1 1
x 1
 1 2
 ln x    2
 x  x x

Indeterminate forms of the type 1

Let f (x ) and g (x) be two functions such that

lim f ( x)  1 and lim g ( x)   .
x a x a
 Suppose we want to find the limit lim f ( x)g ( x ) which corresponds the case 1 .
x a

In this case, after obvious transformations, we obtain

( f ( x ) 1) g ( x )
 f ( x ) 1 
1

lim f ( x)  lim1  ( f ( x)  1)
lim ( f ( x ) 1) g ( x )
 e x a
g ( x)

x a x a
 

Where the lim( f ( x)  1)  g ( x) corresponds to the case 0   that was discussed above.
x a

Thus, in the case 1 we can apply the formula

lim f ( x)
lim ( f ( x )1) g ( x )
 e x a
g ( x)
xa

Examples.
x  1  1
 1  lim  1 2 1  x lim
1. lim  1  2   e x   x   e x  x  e 0  1
x 
 x 
1
ln x
lim lim x
x 0 1 x 0 1
lim 1 x 1ln x  2
1  x ln x  (1 )  e
lim xln x lim (  x )
2. lim x 0
 e x 0 e x
e x
 e x 0  e 0  1.
x 0

lim x 1
1 1
lim( 1)
3. lim x1 x  (1 )  e x 1 1 x
 e x 1  e1.
x 1

Indeterminate forms of the type 0 0

Let f (x ) and g (x) be two functions such that

lim f ( x)  0 and lim g ( x)  0 .

x a x a

Suppose we want to find the limit lim f ( x) g ( x)

.
x a

In this case, we have an indeterminate form of the type 0 0 . We can transform this
limit as follows

lim f ( x)
lim g ( x ) ln f ( x )
 lim e g ( x ) ln f ( x )  e x a
g ( x)
x a x a

where the limit lim g ( x) ln f ( x) corresponds to the case 0  

xa

Examples.
1
ln x lim x
lim
x  0 1 x  0 1
lim x ln x   lim x
x e e e e  e0  1
x
1. xlim
 0
x x  0 x2 x  0
 1
ln x x
lim lim
x 0 1 x 0 cos x  sin x 
lim sin x ln x   lim  tan x 
2. lim x
x 0
sin x
e x  0
e sin x
e sin 2 x
e x 0  x 
 e0  1

Derivatives and Differentials of Higher Orders

I. Derivatives of higher orders. Let y  f (x) . Then the derivative
y /  f / ( x) is called the derivative of the first order or the first derivative of the

function f (x ) . In its turn, f / ( x) is also a function of x and therefore it is possible

to take its derivative which is called the derivative of the second order or the
second derivative of the original function:
y //  ( y / ) /  f // ( x) .

In the same way we define the derivative of the third order (the third
derivative):
y ///  ( y // ) /  f /// ( x)

In the general case, the derivative of the n-th order is defined as

y ( n )  ( y ( n1) ) /  f /( n ) ( x) .

The derivative of the n-th order of a linear combination of functions u  x  and

v x  is readily expressed in terms of the corresponding derivatives of u  x  and
v x  .Namely, if C1 and C2 are constants, then

C u  C v    C u    C v  
1 2
n
1
n
2
n

The rule for finding the derivative of the n-th order of a product of functions is
essentially more complicated. Let u and v be function of x and y  uv

It is required to express y  n  in terms of functions u , v and their derivatives.

Differentiating we find, in succession,
y  uv  uv
y  uv  2uv  uv
y  uv  3uv  3uv  uv
Leibniz`s Rule. For any n there holds the formula
 nn  1 n2 
y n   uv  u n v  nu n1v  u v  ...  nuv n1  uv n 
n

2!
This formula can be formally obtained if we take Newton`s binomial formula
for y  u  v  and then replace the powers of u and v by the derivatives of the
n

corresponding orders of u and v( and put u 0   u, v 0   v )

II. Differentials of higher orders. Let y  f (x) . Then

dy  f / ( x) dx

The differential dy is called the first differential (the first-order differential).

The differential of the second order (the second-order differential) of a function is
the first differential of the first differential of the function. In is denoted by d 2 y
d 2 y  d (dy)  d ( f / ( x)dx)  d ( f / ( x)) dx  ( f / ( x)) / dx  dx  f // ( x)(dx) 2  f // ( x)dx2

Similarly, the third differential or the differential of the third order d 3 y of a

function y  f (x) is the differential of its second differential as a function of x.
d 3 y  d (d 2 y )  d ( f // ( x)dx 2 )  d ( f // ( x)) dx 2  ( f // ( x)) / dx  dx 2  f /// ( x)dx 3

In the general case, the differential of the n th order of a function y  f (x) is

defined as
d n y  d (d n 1 y )  f ( n ) ( x)dxn .

TEST
tgx
1) Calculate the limit by L’Hospital –Bernoulli’s rule lim
x tg 3x
2

A) 1 B) 2 C) 3 D) 4 E) 5

1  cos ax
2) Calculate the limit by L’Hospital –Bernoulli’s rule lim
x 0 1  cosbx

A) a 2 a2 C) b 2 D)
a
E)
b
B)
b2 b a
 e ax  e bx
3) Calculate the limit by L’Hospital –Bernoulli’s rule lim
x 0 sin x
A) b  a B) a  b C) a D) b E) 0
 1x 
4) Calculate the limit by L’Hospital –Bernoulli’s rule lim x e  1
x 
 

A) 1 B) 2 C) 3 D) 4 E) 5
e ax  1
5) Calculate the limit by L’Hospital –Bernoulli’s rule lim
x 0 sin x
A) a B) 1 C) a2 D) 2 E) a3

6) Calculate the limit by L’Hospital –Bernoulli’s rule lim

x 0
x sin x

A) 0 B)1 C) 2 D) 3 E) 4

7) Calculate the limit by L’Hospital –Bernoulli’s rule lim

x 0
xx

A) 2 B) 3 C) 1 D) 4 E)
5

x
8) Calculate the limit by L’Hospital –Bernoulli’s rule lim (  x)tg
x  2

A) 2 B) 3 C) 4 D) 5 E) 6
9) Calculate the limit by L’Hospital –Bernoulli’s rule lim
x 0
(arcsin x) tgx

A) 0 B) 4 C) 3 D) 2 E) 1

x  arctgx
10) Calculate the limit by L’Hospital –Bernoulli’s rule lim
x 0 x3

A)
1
B)
1
C)
1
D)
1 E)1
2 3 4 5

Lecture 9
Taylor’s Formula
The replacement of a function by a simple expression of the type of a polynomial
is extremely convenient in many problems of mathematics. The existence of such a
polynomial is extremely important since it makes it possible to replace,
approximately, the given function by a polynomial.
The computation of the values of this function then reduces to the computation of
the values of the polynomial, which can be achieved with the aid of the simplest
arithmetical operations. The evaluation of the accuracy of these approximations is
also extremely important for practice.
If the function f (x ) possesses the derivative f / (a ) at a point a, then
y
lim  f / (a)
x  0  x

If follows that
y
 f / (a)  
x

or
y  f / (a)x    x

That is f ( x)  f (a)  f / (a)  ( x  a)    ( x  a)

And hence f ( x)  f (a)  f / (a)  ( x  a)    ( x  a)
where lim  0 (i.e.  is an infinitesimal as x  a ). In other words, there exists
x a

the linear function (i.e. the polynomial of the first degree) .

P1 ( x)  f (a)  f / (a)( x  a)

such that
f ( x) P1 ( x)   ( x)

Where
 ( x)   ( x  a) and lim ( x)  0
x a

Now dropping the infinitesimal  (x) we get the following approximate

formula
f ( x)  P1 ( x)
with an accuracy  (x) which is an infinitesimal of higher order than  ( x)  x  a ,
 ( x)
since lim  0 . It turns out that such approximate formulas can be made more
xa  ( x)

accurate if we apply derivatives of higher order. This problem is solved by means

of Taylor’s formula named after Brook Taylor (1685-1731), an English
mathematician.
Theorem. If the function f (x ) possesses the (n  1)th derivative f n1 ( x) (i.e.n+1
times differentiable) at all the points of an interval containing the point a, then it
can be represented as the sum of the polynomial of the n th degree and remainder
term R n :
f / (a) f // (a) f ( n ) (a)
f ( x)  f (a)  ( x  a)  ( x  a)  ... 
2
( x  a ) n  Rn , (1)
1! 2! n!

where Rn   ( x)( x  a) n , lim ( x)  0 (Peano’s form)

x a

or
f ( n 1) ( )
Rn  ( x  a) n 1 ,   a   ( x  a), 0    1 (Lagrange’s form)
(n  1)!

Formula (1) is called Taylor’s formula for f(x) with the remainder Rn (x) .
Proof. Let us first suppose that we are given a polynomial Pn (x ) of degree n. A
polynomial is usually considered as expanded into powers of x
Pn ( x) a0 a1x  a2 x 2  ...  an x n

But it is quite easy to expand it in powers of x-a where a is an arbitrary number.

Pn ( x) c0 c1 ( x  a)  c2 ( x  a)2  ...  cn ( x  a) n (2)
The coefficients c0 , c1 ,...., cn here can be found in the following way. First we put
x  a and obtain Pn (a) c 0 . Then we differentiate the formula (2):
Pn/ ( x) c12c2 ( x  a)  ...  ncn ( x  a)n1

If now we put here x  a then we get Pn/ (a) c1 . Let us differentiate once again:
Pn// ( x)  2c 2 ...  n  (n  1)cn ( x  a)n2

which implies Pn// (a)  2c2  2!c2

Further, in a similar way we derive Pn/// (a)  3!c3 and so on. Generally, Pn( k ) (a)  k!ck ,
which yields
P ( k ) (a)
ck 
k!
Thus, formula ( 2 ) can be rewritten in the following form:
Pn/ (a ) P // (a) P ( n ) (a)
Pn ( x)  Pn (a)  ( x  a)  n ( x  a) 2  ...  n ( x  a) n
1! 2! n!

If now we take the function f (x ) in place of a polynomial Pn (x) then we obtain the
polynomial of the form
f / ( a) f // (a) f ( n ) (a)
Pn ( x)  f (a)  ( x  a)  ( x  a)  ... 
2
( x  a) n
1! 2! n!
the derivations of which coincide with the derivatives of the corresponding order
of the function f (x ) at a point a:
Pn (a)  f (a), Pn/ (a)  f / (a),..., Pn( n) (a)  f ( n) (a) .

Let us denote the difference between f (x ) and Pn (x) by Rn (x)

Rn ( x)  f ( x)  Pn ( x) .

It is easy see that

Rn (a)  Rn/ (a)  Rn// (a)  ...  Rn( n) (a)  0

Therefore, applying the L’Hospital-Bernoulli’s rule n times to the limit

Rn ( x) Rn/ ( x) Rn// ( x) Rn( n ) ( x) Rn( n ) (a)
lim  lim  lim  ...  lim  0
x a ( x  a ) n xa n( x  a ) n 1 xa n( n  1)( x  a ) n 2 x a n! n!

Rn ( x)
finally we get   ( x) , where lim ( x)  0
( x  a) n xa

that is Rn ( x)   ( x)  ( x  a)n  0((x  a)n ) .

Definition. The polynomial of the form
f / ( a) f // (a) f ( n ) (a)
Pn ( x)  f (a)  ( x  a)  ( x  a)  ... 
2
( x  a) n
1! 2! n!
is called Taylor’s polynomials of the n th degree.
Thus, the theorem proved above shows that a Taylor polynomial differs from
f (x ) by a term Rn (x) (the remainder) which is an infinitesimal of the highest order
in comparison with all the polynomials of the same degree as x  a . In other
words, the Taylor polynomial is a good approximation for the function f (x ) as
x  a.

f ( x)  Pn ( x) .

Maclaurin’s Formula
On particular, when a  0 we obtain the formula
f / (0) f // (0) f ( n ) (0)
f ( x)  f (0)  x x  ... 
2
x n  Rn ( x)
1! 2! n!
Which is sometimes called Maclaurin’s formula, where the remainder Rn (x) has
the following forms:
1) Peano’s form: Rn ( x)  0( x n ) ,
0( x n )   ( x)  x n , lim ( x)  0
x0

f ( n1) (x) n1

2) Lagrange’s form: Rn ( x)  x , 0    1.
(n  1)!

Expanding functions by Maclaurin’s Formula

x x 2 x3 xn
ex  1     ...   0( x n ) ,
1! 2! 3! n!

x3 x5 x7 x 2 n1
sin x  x     ...  (1) n
 0( x 2 n2 ) ,
3! 5! 7! (2n  1)!

x2 x4 x6 x 2n
cos x  1     ...  (1) n  0( x 2 n1 ) ,
2! 4! 6! (2n)!

 (  1)  (  1)(  2)  (  1)...(  n  1)
(1  x)  1  x  x2  x 3 ...  x n  0( x n ) ,
2! 3! n!

x 2 x3 xn
ln(1  x)  x    ...  (1) n1  0( x n ) .
2 3 n

Where 0( x n ) is an infinitesimal of higher order than x n i.e.

0( x n )   ( x) x n , lim ( x)  0 .
x0

Example 1.
Find the Maclaurin series for cos2 x
Solution.
 1  cos 2 x
We use the trigonometric identity cos 2 x 
2

 1 x
n 2n

As the Maclaurin series for cos x is  ,we can write:

n 0 2n !
cos 2 x  

 1 2 x 
n 2n



 1 2 2 n x 2 n
n

n 0 2n ! n 0 2n !
Therefore

1  cos 2 x  1  

 1 2 2 n x 2 n
n

2

 1 2 2 n x 2 n
n

n 0 2n ! n 1 2n!

cos x 
2 1  cos 2 x
1 

 1 2 2 n1 x 2 n
n

2 n 1 2n !
Example 2.
Obtain the Taylor series for f x   3x 2  6 x  5 about a  1
Solution.

Compute the derivatives:

f x   6 x  6 , f x  6 , f x   0
As you can see, f n  x   0 for all n  3 .Then for a  1 , we get
f 1  2, f 1  0, f 1  6
Hence, the Taylor expansion for the given function is

1 x  1 6 x  1
n 2

f x    f n 
2  2  3 x  1
2

n 0 n! 2!
Test

1.Find the Maclaurin series for 2x

2. Find the Maclaurin series for sin 2 x

3. Find the Maclaurin series for 3x

4. Find the Maclaurin series for e 2 x
1
5. Obtain the Taylor series for f x   about a  2
x

6. Find the Maclaurin series for sin 3x

7. Obtain the Taylor series for f x   x about a  4
8. Obtain the Taylor series for f x   x 3  2 x  4 about a  2
9. Obtain the Taylor series for f x   2 x 3  x 2  3x  8 about a  1
10. Find the Maclaurin series for 1 x

Lecture 10
Application of differential calculus to investigation of behaviour of functions
Monotonicity
A function f (x ) is called increasing in an interval (a,b) if for any two points
x1 , x2  (a, b)

x1  x2  f ( x1 )  f ( x2 )

In other words, a function is increasing if the values of the functions increase as the
argument increases.

y y

f(x2) f(x2)

f(x1) f(x1)

o x2 x o x1 x2 x
x1

A function f (x ) is called decreasing in an interval (a,b) if for any two points

x1 , x2  (a, b)

x1  x2  f ( x1 )  f ( x2 )

In other words, a function is decreasing if its values decrease as the argument

increases.

y y
f(x2)
f(x2)
f(x1) f(x1)

o x1 x2 x o x1 x2 x

Both increasing and decreasing functions are called monotonic.

 Conditions for Monotonicity
Between the character (nature) of the monotonicity of differentiable function and
the sigh of its derivative there is a connection stated in the following theorem.
Theorem. Let f (x ) be differentiable function in an interval (a,b). Then
f / ( x)  0  f  increases  f / ( x)  0

f / ( x)  0  f  decreases  f / ( x)  0
Proof. Let us prove the first statement, the second is proved similarly.
Suppose f / ( x) is positive in an interval (a,b) , i.e. f / ( x)  0 for all x  (a, b) . Now let
us take two points x1 , x2  (a, b) and assume that x1  x2 . According to the Lagrange’s
theorem we can write
f ( x2 )  f ( x1 )  f / ( )( x2  x1 ) , x1    x2
It is clear that
f / ( x)  0 for x  (a, b)  f / ( )  0  f / ( )  ( x2  x1 )  0  f ( x2 )  f ( x1 )  0 ,

Thus, we have for any two points x1 , x2  (a, b)

x1  x2  f ( x1 )  f ( x2 )

That is f (x) is increasing in (a,b).

The left part of the first statement has proved. Now let us prove the right part
of the first statement. In other words we have to prove that
f  increases  f / ( x)  0

Since f (x ) is differentiable function in (a,b), that is f (x ) possesses the derivative

f / ( x) at any point x  (a, b) , we have

f ( x  x)  f ( x)
f / ( x)  lim
x0 x
It is easy to see that for x  0 we have
x  x  x  f ( x  x)  f ( x)  f ( x  x)  f ( x)  0 
f ( x  x)  f ( x) f ( x  x)  f ( x)
  0  lim  0  f / ( x)  0
x x  0 x
and for x  0 we have
 x  x  x  f ( x  x)  f ( x)  f ( x  x)  f ( x)  0 
f ( x  x)  f ( x) f ( x  x)  f ( x)
  0  lim  0  f / ( x)  0
x x  0 x
Thus, for any case f / ( x)  0 which is what we had to prove.
Example. Find the intervals of increasing and decreasing of the function
f ( x)  x 4  8 x 2  3 .

Solution. We find f / ( x)  4 x3  16x .

It is easy to verify that
f / ( x)  0 for x  (2,0)  (2, ) and

f / ( x)  0 for x  (,2)  (0,2)

Thus, f (x) increases in  2,0  2, and

f (x ) decreases in  ,2  0,2.

Points of Extremum. Extrema of Functions

Definition. A point x0 is called a point of local maximum of the function f (x ) if
the value f ( x0 ) is the greatest value of the function f (x ) in a neighbourhood of the
point x0 .
In other words , it means that there exist a  - neighbourhood   0 of the point x0 ,
U  ( x0 )  ( x0   , x0   ) such that for all x  U  ( x0 ) f ( x)  f ( x 0 ) .

y y y
max
f(x0) max
f(x0) max f(x0)

x0   o x0   x0 x0   o x0   x0 x0   x
o x0   x0 x x

We also say that f ( x0 ) is a local maximum of the function f (x ) .

local
y max
max
max

o  
Here,  ,  ,  are points of local maximum of the function f (x ) , and
f ( ), f (  ), f ( ) are local maximum of the function f (x ) . The point of local
minimum and the local minimum of a function are defined in a similar way.
Definition. A point x0 is called a point of local minimum of the function f (x ) , if
the value f ( x0 ) is the least value of the function f (x ) in a neighbourhood of the
point x0 .

Lecture 11
Convexity and Concavity of Curve. Asymptotes of the function graph.

Let a function y  f (x) have a graph of the shape shown below

y

convex

o x

We see that the graph lies below the tangent drawn at any point.
Definition. The graph of the function y  f (x) is said to be convex (convex up or
concave down) on the interval (a,b) if it lies below the tangent drawn at any point
of that interval.
Now let us suppose that the graph of a function y  f (x) has the shape shown
below

concave

o x
In this case we see that the graph lies above the tangent drawn at any point.
Definition. The graph of the function y  f (x) is said to be concave (concave up or
convex down) on the interval (a,b) if it lies above the tangent drawn at any point
of that interval.
Let us consider in detail the graph of the function y  sin x (the sinusoid) on the
interval [ ,  ]

 o  x
Considering the arcs of the sinusoid corresponding to the intervals [  ,0] and
[0,  ] we readily observe that the graph is concave (or convex down) for x  [ ,0] ,

that is it lies above the tangent drawn at any point of that interval, while the graph
is convex (or convex up) for x  [0,  ] that is it lies below the tangent drawn at any
point of that interval. However, it should be noted, that the arc of sinusoid
corresponding to the whole interval [ ,  ] is neither convex nor concave.
Theorem. (Sufficient condition for the convexity (concavity) of the graph of a
function).
If f // ( x)  0 on the open interval (a, b) then the graph of the function is convex on
that interval.
If f // ( x)  0 on the open interval (a, b) then the graph of the function is concave on
that interval.
Proof. Let f // ( x)  0 at all points x  (a, b) . Let us take an arbitrary point
x0  (a, b) and draw the tangent to the graph of the function y  f (x) at the point

( x0 , f ( x0 )) . The equation of the tangent has the form

y  f ( x0 )  f / ( x0 )(x  x0 )  T ( x)
We have to prove that all the points of the graph of the function y  f (x) lie below
its every tangent y  T (x) . To prove it, we have to show that the ordinate of any
point of the graph of the function y  f (x) is less than the ordinate of the tangent
y  T (x) at the same value of x. To this end, we consider the difference

f ( x)  T ( x)  f ( x)  f ( x0 )  f / ( x0 )(x  x0 ) .

Applying the Lagrange’s theorem to the difference f ( x)  f ( x0 ) we get

f ( x)  T ( x)  f / (c)  ( x  x0 )  f / ( x0 )( x  x0 ) 
[ f / (c)  f / ( x0 )]  ( x  x0 )

where the point c lies between x0 and x.

Now applying again the Lagrange’s theorem to the difference f / (c)  f / ( x0 ) we
obtain
f ( x)  T ( x)  f // ( )(c  x0 )  ( x  x0 )

and f // ( )  0 since f // ( x)  0 for x  (a, b) .

There are two possible cases here: x  x0 or x  x0 . It is clear that

x  x0  x0  c  x  (c  x0 )(x  x0 )  0  f // ( )(c  x0 )(x  x0 )  0 since f // ( )  0 ,

x  x0  x  c  x0  (c  x0 )(x  x0 )  0  f // ( )(c  x0 )(x  x0 )  0 since f // ( )  0 .

Therefore in any case f ( x)  T ( x)  0 , t.e. f ( x)  T ( x) ,which is what we had to

prove.
The case when f // ( x)  0 is considered similarly.
An important characteristic of a curve are the points separating its convex and
concave arcs.
Definition. The point ( x0 , f ( x0 )) of the graph of the function separating its convex
part from the concave part is called a point of inflexion.
At a point of inflection the tangent intersects the curve. In the vicinity of such a
point the curve lies on both sides of its tangent drawn through that point.

inflection point
This does not require the point to be a stationary point. Geometrically, the portion
of the graph around a point of inflexion can interpreted as “S-bend” which is
equivalent to the curve switching from convex to concave or vice versa.
These considerations lead to a new definition.
Definition. Let y  f (x) be differentiable at a point x0 and let y  T (x) be the
equation of a tangent to the graph of the function at the point ( x0 , f ( x0 )) .
If the difference f ( x)  T ( x) changes sing as x passes through the point x0 is a point
of inflexion.
Theorem. (The first sufficient condition for the inflexion point) Let f (x ) be
differentiable at a point x0 and possesses the second derivative f // ( x) in a
neighbourhood of x0 except possible the point x0 itself.
If the second derivative f // ( x) changes sign as x passes through the point x0 then x0
is a point of inflexion of the function f (x ) .
Theorem. (The second sufficient condition for the inflexion point) If f // ( x)  0 and
f /// ( x0 )  0 then x0 is a point of inflexion.

Theorem. (Necessary condition for a point of inflexion). If x0 is the abscissa of the

inflexion point of the graph of the function y  f (x) , then the second derivative
f // ( x0 ) is equal to zero or does not exist.

Thus, briefly, at a point of inflexion f // ( x0 )  0 . This proposition provides a

necessary condition for the point of inflexion.
Although a point of inflexion must have zero y

second derivative, f // ( x0 )  0 , the converse is not

y  x4

o x
true, that is by far not every root of the equation f // ( x)  0 is the abscissa of a point
of inflexion. For example, we can see this from the function f ( x)  x 4 .
The first and the second derivatives of the function f ( x)  x 4 are f / ( x)  4 x 3 and
f // ( x)  12 x 2 . Although f // (0)  0 , the point (0,0) is the vertex of the parabola
y  x 4 and does not serve as its point of inflexion.

Examples.
1.Find the intervals of convexity and concavity of the graph of the function
y  x 5  5x  8 .

Solution. We have y /  5 x 4  5 , y //  20x 3 .

It is clear that
if x  0 , then y //  0 and the curve is concave
if x  0 , then y //  0 and the curve is convex.
Thus we see that the curve is convex on the interval (  ,0 ) and concave on the
interval ( 0,  ).
2. Find the coordinates of the point of inflexion on the curve y  x 3 .
Solution. We have y /  3x 2 , y //  6 x . Now we equate the second derivative to zero:
6 x  0, i.e. x  0 . y
Thus, the point (0,0) is an inflexion point.
3. The curve y  x 4  6 x 2  7 x  2 has two points of y  x3
inflexion. Find their coordinates.
Solution. We have y /  4 x 3  12x  7 , y //  12x 2  12 . o x
At a point of inflexion y //  0 : 12 x 2  12  0 i.e.

x  1.
Thus, the points (-1,-10) and (1,4) are the points of
inflexion.

Asymptotes
When studying the behaviour of a function or when investigating the graph of a
function as its argument tends to plus or minus infinity we deal with the branches
of its graph receding to infinity.
Definition. A straight line L is called an asymptote to the curve y  f (x) if the
distance between the point P( x, y ) of the curve and the line L tends to zero as the
distance from this point to the origin increases indefinitely.
A graph of y  f (x) may have vertical, horizontal and inclined asymptotes.
Definition. A straight line x=a is called a vertical asymptote of the curve y  f (x)
if lim f ( x)   or lim f ( x)   .
x a x a

There may be any number of vertical asymptotes, even an infinite number.

Examples. y
1 1
1. y  , lim  
x2 x 0 x2
The straight line x=0 is a vertical asymptote of the
1 o x
curve y  .
x2
2. y  tan x y
This function has an infinite number
of asymptotes:

y  k , kZ
2
Definition. A straight line y=b is  3
 o  3 x

2 2 2 2
called a horizontal asymptote of the
curve y  f (x) if there exist a limit
lim f ( x)  b or lim f ( x)  b .
x  x 

There cannot be more than two horizontal asymptotes.

Examples.
1. y  arctan x

y

2

o x



2

 
lim arctan x  y is a horizontal asymptote,
x   2 2
 
lim arctan x    y  is a horizontal asymptote.
x   2 2
 
Thus, this function has two asymptotes: y   and y  .
2 2

2. y  arccot x


2

o x

lim arc cot x  0  y  0 is a horizontal asymptote,

x 

lim arc cot x    y   is a horizontal asymptote.

x 

Therefore, this function has two asymptotes: y

y  0 and y   .

1
3. y 
x
1 o x
lim  0  y  0 is a horizontal asymptote,
x  x
 1 1
lim  , lim    x  0 is a vertical asymptote.
x  0 x x  0 x
Thus, this function has one vertical and one horizontal asymptotes.
Definition. A straight line y  kx  b is called an inclined asymptote of the curve
y  f (x) if there exist limits

f ( x)
k  lim , b  lim [ f ( x)  kx]
x   x x  

or
f ( x)
k  lim , b  lim [ f ( x)  kx] .
x   x x  

There cannot be more than two inclined asymptotes corresponding to x   and

x  -

These asymptotes are found as follows: Let a straight line y  kx  b be an

asymptote of the graph of y  f (x) as x   .

y
y  kx  b

y  f (x )

o x

Then the difference d between the asymptote y  kx  b and the curve y  f (x) is
equal to
d  (kx  b)  f ( x)

and tends to zero as x   .

It follows that
f ( x)  kx  b  d

and, hence
f ( x) b d
k  k
x x x
that is
 f ( x)
k  lim
x x

Besides ,
d  (kx  b)  f ( x)  f ( x)  kx  b  d  b

That is
b  lim f ( x)  kx .
x 

Each of these limits must exist and be finite, otherwise there will be no asymptote
as x   . If these finite limits do exist then the asymptote also exists since it is
seen from the last equality that the value  f ( x)  kx  b tends to zero as x  , i.e.
lim f ( x)  (kx  b)  0
x 

Examples.
x3
1. Find the asymptotes of the curve y 
x2  1
Solution.
The graph of this function has two vertical asymptotes:
x  1 and x  1 , since

x3 x3
lim   and lim  
x  1 x 2  1 x 1 x 2  1

To determine its inclined asymptote we compute the limits

f ( x) x3 x2
k  lim  lim  lim 2 1
x  x x  x( x 2  1) x  x  1

 x3 
b  lim f ( x)  kx  lim f ( x)  x   lim  2
x
 x   lim 2 0
x  x  1
x  x 
  x  x  1

Since k=1 and b=0 the asymptote has the form y=x. Thus, the bisector of the first
quadrant y=x is the only inclined asymptote the graph possesses.

TEST
1) Find the coordinates of the point of inflexion on the curve
y  ( x  3) 3  3( x  3) 2  x  5
 A) (3;-2) B) (2;-1) C) (4;3) D) (1;0) E) (5;20)

2) Find the coordinates of the point of inflexion on the curve y  ( x  4) 5  2 x  3

A) (4;5) B) (3;2) C) (2;33) D)(5;8) E)(6;39)

3) Find the coordinates of the point of inflexion on the curve y  x 3  3x 2  4 x  1

A) (0;1) B) (1;-1) C)(2;11) D)(-2;11) E)(-1;5)

4) Find the interval of convex of the graph of the function y  ( x  2) 3  7 x  5

A) (;2) B) (2;) C) (-2;2) D) (0;) E) (;0)

5) Find the interval of concave of the graph of the function y  2 x 3  6 x 2  5

A) (; ) B) (;1) C) (1; ) D) (1;1) E) (0;)

6) Find the interval in which y  2( x  3) 3  6 x  4 concave

A) (0;3) B) (;3) C) (3;) D) (-3;3) E) (3;)

7) Find the interval in which y  2 x 4  6 x 2  12 concave

A) (;) B) (0;) C) (;0) D) (2;) E) (;2)

8) Find the interval in which y   x 4  6 x 2  15 convex

A) (;4) B) (4;) C) (;0) D) (;) E) (0;)

9) Find the interval in which y  2 x 3  5 x 2  7 convex

C) (0;)
A)   ;  D)   ;  E)  ; 
5 B) (;0) 5 5 5
 6  6 6 6 

10) Find the interval in which y  x 3  2 x 2  11 concave

B) (;)
C)  ;  D)   ;  E)   ; 
A) (;0) 2 2 2 2
3   3  3 3
 Lecture 12
Indefinite Integral
Antiderivative
When differentiating a function we solve the problem of finding of the derivative
of a given function. Now we want to solve the inverse problem: given the
derivative of a function, to find this function.
Definition. The function F (x ) is called an antiderivative or a primitive of the given
function f (x ) in a given interval if F / ( x)  f ( x) or dF ( x)  f ( x)dx .
Let us consider some examples.
x5
1. The function F ( x)  is an antiderivative of the function f ( x)  x 4 in the
5
interval (,) , since
/
 x5 
F ( x)      x 5    5 x 4  x 4  f ( x)
/ 1 / 1
 5  5 5

2. The function F ( x)  sin 2 x is a antiderivative of the function f ( x)  sin 2 x in the

interval (,) , since

F / ( x)  sin 2 x   2  sin x  (sin x) /  2  sin x  cos x  sin 2 x  f ( x)

/

x5
It is easy to see, from these two examples, that the derivatives of  C and
5
sin 2 x  C , where C is an arbitrary constant, are also equal respectively to x 4 and

sin 2 x . These examples indicate that a given function has infinitely many
antiderivatives because the constant C can take on arbitrary values.
Thus we see that the problem of finding an antiderivative has infinitely many
solutions.
Properties of an antiderivative
1. If the function f (x ) possesses an antiderivative F (x ) , then it possesses infinitely
many antiderivatives, all the antiderivatives being contained in the expression
F ( x )  C , where C is an arbitrary constant.

2. Any two primitives differ by a constant.

Proof. Let F1 and F2 be two antiderivatives of f. It means that F1/  f and F2/  f .
Let us consider the difference F1  F2 and find its derivative. We have
( F1  F2 ) /  F1/  F2/  f  f  0

It follows that F1  F2  C , where C is a constant.

3. If the function f possesses an antiderivatives F and the function g possesses an
antiderivatives G, then the sum f  g possesses an antiderivative F  G .
Proof. Indeed since F /  f and G /  g we have
( F  G) /  F /  G /  f  g

4. If the function f possesses an antiderivatives F, then the function kf , where k is a

constant, possesses an antiderivative kF .
Proof. Indeed we have
(kF ) /  k ( F ) /  k  f

5. If the function f (x ) possesses a primitive F (x ) , then the function f (t ) possesses

a primitive F (t ) .
6. If the function f (x ) possesses a primitive F (x ) , then the function f (ax  b)
1
possesses a primitive F (ax  b) , where a and b are constant and also a  0 .
a

Proof. According to the rule for finding the derivative of a composite function we
have
/
1  1 /
 F (ax  b)   F (ax  b)  a  f (ax  b) .
a  a

Indefinite Integral and its Properties

Definition. The collection of all antiderivatives of a function f (x ) is called the
indefinite integral and is denoted by the symbol  f ( x)dx .
  f ( x)dx  F ( x)  C
Here  is the integral sign, f (x ) is the integrand, f ( x )dx is the element of

integration, and x is the variable of integration.

The process of finding an indefinite integral is called integration of a function.

Properties of an Indefinite Integral (rules of integration)

1.  ( f ( x)dx) /  f ( x)

Proof. By the definition of the indefinite integral we have

(  f ( x)dx) /  ( F ( x)  C ) /  F / ( x)  C /  F / ( x)  f ( x)

2. d  ( f ( x)dx)  f ( x)dx

Proof. According to the definitions of the differential and the indefinite integral
we have
d (  f ( x)dx)  (  f ( x)dx) /  f ( x)dx

3.  f / ( x)dx  f ( x)  C or  df ( x)  f ( x)  C
Thus, the signs of integration and differentiation mutually cancel out. The result of
computing an indefinite integral can always be verified by finding the derivative of
the result.
4. A constant factor in the integrand can be taken outside the sign of the integral:

 Cf ( x)dx  C  f ( x)dx , C=constant

5. The integral of a sum of a finite number of functions is equal to the sum of the
integrals of these functions:

  f ( x)  f
1 2 ( x)dx   f1 ( x)dx   f 2 ( x)dx ,

f 1  f 2  ...  f n dx   f1 dx   f 2 dx  ...   f n dx

6. On the invariance of integration formulas.

If  f ( x)dx  F ( x)  C and u   (x) any function possessing a continuous derivative,

then  f (u)du  F (u)  C .

Proof. From  f ( x)dx  F ( x)  C it follows that F / ( x)  f ( x) .
For the composite function F (u )  F ( ( x)) its differential is
dF (u )  F / (u )du  f (u )du .

Therefore  f (u )du   dF (u )  F (u )  C
Table of Basic Integrals

1.  dx  x  C 10.  e x dx  e x  C

x  1 ax
2.  x dx 

 C for   1 11.  a dx 
x
C
 1 ln a
1 dx
3.  x dx  ln x  C 12.  1 x2
 arcsin x  C

4.  sin xdx   cos x  C dx x

13.  a2  x2
 arcsin  C
a

5.  cos xdx  sin x  C dx

14. 1 x 2
 arctan x  C

1 dx 1 x
6.  cos 2
x
dx  tan x  C 15. a 2
x 2
 arctan  C
a a
1 dx 1 xa
7.  sin 2
dx   cot x  C 16. x  ln C
x 2
 a 2 2a x  a
8.  tan xdx   ln cos x  C dx
17.   ln x  x 2  a 2  C
x a2 2

9.  cot xdx  ln sin x  C

Basic Methods of Integration

The basic methods of integration are the following methods:
1. The decomposition method
2. Integration by parts.
3. Integration by change of variable (by substitution).

The Decomposition Method

A given integral can often be represented in the form of a sum of tabular integrals
(i.e. the integrals from the table). Then we perform termwise integration and thus
obtain the answer. Let us consider several examples.

 a 0  a1 x  a 2 x 2  ...  a n x n dx  a0  dx  a1  xdx a 2  x 2 dx ...  a n  x n dx 

1. 1 1 1
 a0 x  a1 x 2  a 2 x 3  ...  a n x n 1  C
2 3 n 1

 3x 
 2 sin x  5e x dx   3x 2 dx   2 sin xdx   5e x dx 
2
2.

x3
3 x dx  2 sin xdx  5 e dx  3 
2
 2  ( cos x)  5e x  C 
x

3
 x  2 cos x  5e  C
3 x

sin 2 x 1  cos2 x
 tan xdx     cos2 x dx 
2
dx
cos2 x
3.
 1  1
   1dx   dx   dx  tan x  x  C
 cos x 
2
cos2 x

Integration by parts
The method of integration by parts is implied by formula for differentiating a
product of two functions. Let u  u (x) and v  v(x) be functions of x possessing
continuous derivatives. We have
d (uv)  udv  vdu
whence
udv  d (uv)  vdu
Integrating both sides of the latter equality, we get

 udv   d (uv)   vdu

that is,

 u dv  uv   vd u (*)

This is the formula of integration by parts which can be rewritten as follows

 u ( x )v ( x)dx  u ( x)v( x)   v( x)u / ( x)dx

/
We do not write down the arbitrary constant appearing in the integration of d (uv)
and include it into the arbitrary constant in the second integral on the right-hand
side of equality.
Let us consider some examples.
1. Let us find  xe x dx

We put x  u, e x dx  dv, whence

du  dx , and v   e x dx  e x .

Now using the formula of integration by parts (*) we get

 xe dx  xe   e dx  xe  ex  C
x x x x

2.Let us find  x cos xdx .

Here we put x  u , cos xdx  dv , whence

du  dx and v   cos xdx  sin x

By formula (*) we obtain

 x cos xdx  x  sin x   sin xdx  x  sin x  cos x  C

3. Take the integral  x 2 cos xdx .

Here we put x 2  u, cos xdx  dv, whence

du  2 xdx and v  sin x .
According to the formula (*) we have

x cos xdx  x 2  sin x  2 x sin xdx .

2

The last integral is again found with the aid of integration by parts. Finally we
obtain

x cos xdx  x 2 sin x  2 x cos x  2 sin x  C .

2

4. Let us find the integral  x ln xdx .

We put ln x  u, xdx  dv
Whence
1 x2
du  dx, and v   xdx  .
x 2
Then by the formula (*) we get
x2 1 x2 x2
 x ln xdx  2
ln x   xdx  ln x 
2 2 4
C.

Integration by change of variable

A change of variable in an indefinite integral is performed by means of
substitutions of two types:
I. x   (t ) , where  (t ) is a monotonic, continuously differentiable function of the
new variable t. In this case, the integration is carried out by the formula

 f ( x)dx   f [ (t )]   / (t )dt ;
.

II. u   (x) , where u is a new variable. With the aid of this substitution, the
integration is performed by the formula.

 f [ ( x)]  ( x)dx   f (u)du

/.

Let us consider several examples.

1. Find the integral  (5x  2) 8 dx
1
Solution. Putting 5x  2  t , we get 5dx  dt , i..e. dx  dt , which yields
5
1 8 1 1 1
 (5 x  2) dx   t dt   t 9  C  (5 x  2) 9  C .
8

5 5 9 45

2. Find the integral  sin 3 x cos xdx .

Solution. Putting sin x  u we get du  d (sin x)  cos xdx and hence

u4 1 4
 sin x cos xdx   u du  4  C  4 sin x  C .
3 3

arctan x
3. Find the integral  1 x2
dx .

1
Solution. Putting arctan x  u we get du  dx and hence
1 x2
arctan x u2 1
 1  x 2 dx   udu  2  C  2 arctan x  C
2

A rational function
Definition. A rational function (rational fraction) is a fraction of the from
P( x)
Q( x)

where P(x) and Q (x ) are polynomials.

Definition. A rational fraction is said to be proper if the degree of the numerator
P (x) is less than the degree of the denominator; otherwise the fraction is said to be

improper.
Let us consider the integral of the form
P( x)
 Q( x)dx
P( x)
If is an improper, then it can be represented in the form of a sum of a
Q ( x)

polynomial and a proper fraction:

P( x) M ( x)
 L( x ) 
Q( x ) Q( x)

Consequently.
P( x) M ( x)
 Q( x)dx   L( x)dx  Q( x) dx
The integration of the polynomial L( x)  a0  a1 x  ...  an x n does not present any
difficulties, and hence the problem reduces to integrating the proper fraction in
which the degree of the numerator M (x) is less that of the denominator Q (x ) . The
term partial (elementary) fractions is used for proper fractions of the following
types:
A
I. ;
xa
A
II. , where n is a positive integer;
( x  a) n

Ax  B p2
III. , where  q  0 , that is the quadratic trinomial x 2  px  q does not
x 2  px  q 4

posses real roots;

Ax  B p2
IV. , where n is a positive integer, and  q  0.
( x 2  px  q) n 4
In all the four cases it is assumed that A, B, p, q, a are real numbers. The fractions
we have enumerated will be called, respectively, partial fractions of types I, II, III
and IV.
M ( x)
It turns out that the proper rational fraction can be represented as a
Q( x)

decomposition into partial fractions.

Thus we need to know how to find the integrals of the partial fractions.
Let us consider the integrals of the partial fractions of the first three types. We
have
A
I.  x  a dx  A  ln x  a  C
Indeed, putting x  a  t we find dx  dt and hence
A A 1
 x  a dx   t dt  A t dt  A ln t  C  A ln x  a  C
A A 1
II.  ( x  a) n
dx   
n  1 ( x  a) n 1
C

Here we do the same: x  a  t  dx  dt and we get

A A A 1 A 1
 ( x  a) n
dx  
t n
dt  A t n dt    n1  C  
n 1 t

n  1 ( x  a) n1
C

dx 2 2x  p
III. x 2
 px  q

4q  p 2
arctan
4q  p 2
C

Indeed, for this special case of the partial fraction of type III we obtain
2
 p p2
x 2  px  q   x    q  or x 2  px  q  t 2  a 2 ,
 2 4

p 4q  p 2 p2
where t  x  , a  (here  q  0 ), whence
2 2 4
dx dt 1 t 2 2x  p
x 2
 px  q
 2
t a 2
 arctan  C 
a a 4q  p 2
arctan
4q  p 2

dx
Example. Find the integral x 2
 8 x  32
Solution, we have
dx dx d ( x  4) 1 x4
x 2
 8x  32

( x  4)  16
2
  arctan
( x  4)  16 4
2
4
C
We will show now how to integrate, in a general form, partial fractions of type III.
Ax  B p2
It is required to find  x 2  px  q dx , where
4
 q  0 . Let us isolate the derivative

of the denominator from the numerator of the fraction. To do that, we will

represent the numerator in the form
A Ap
Ax  B  (2 x  p)  B 
2 2

Then
A Ap
(2 x  p)  B 
Ax  B A 2x  p  Ap  dx
 x 2  px  q dx   x 2  px  q dx  2  x 2  px  qdx   B  2  x 2  px  q
2 2

In the first integral the numerator is the derivative of the denominator; therefore,
2x  p d ( x 2  px  q) du
 x 2  px  q   x 2  px  q   u  ln u  C  ln( x  px  q)  C (here
2
dx

x 2  px  q  0 )

The second integral, as has been indicated, can be found by the formula
dx 2 2x  p
x 2
 px  q

4q  p 2
arctan
4q  p 2
C

Thus, finally we have

Ax  B A 2 B  Ap 2x  p
x 2
 px  q
dx  ln( x 2  px  q) 
2 4q  p 2
arctan
4q  p 2
C

(3 x  1)dx
Example. Find the integral  4x 2
 4 x  17
Solution. We have
3 3
(8 x  4)  1 
(3x  1)dx 3 8x  4 1 dx
 4 x 2  4 x  17   4 x 2  4 x  17 dx  8  4 x 2  4 x  17dx  2  4 x 2  4 x  17 
8 2

3 d (4 x 2  4 x  17) 1 1 dx 3 du 1 dx 3 1 dt
8  4 x  4 x  17
2
  
2 4 2
 
17 8 u 8 
 
1 
2
 ln u    2
8 8 t  22

x x x    4
4  2
3 1 1 t 3 1 2x  1
 ln(4 x 2  4 x  17)    arctan  C  ln(4 x 2  4 x  17)   arctan C .
8 8 2 2 8 16 4
Now let us show how to integrate the partial fractions of type IV. It is required to
find the integral
Ax  B p2
 ( x 2  px  q) n dx , where
4
 q  0.

Let us isolate the derivative of the denominator from the numerator of the fraction:
A Ap
(2 x  p)  B 
Ax  B A 2x  p  Ap  dx
 ( x 2  px  q) n dx   ( x 2  px  q) n dx  2  ( x 2  px  q) n dx   B  2  ( x 2  px  q) n
2 2

The first integral on the right-hand side is easily found with the aid of the
substitution x 2  px  q  t whence (2 x  p )dx  dt . We have
2x  p dt 1 1
 (x 2
 px  q) n
dx   n  
t (n  1)t n 1
C  
(n  1)( x  px  q) n 1
2
C

The second integral on the right-hand side must be transformed as follows:

2x  p dx
 (x 2
 px  q) n

 2 n
p 
2
p 
 x     q   
 2  4  

p p2
Putting now x   t , dx  dt and introducing the designation q   a 2 we get
2 4
dx dt
 (x 2
 px  q) n


t 2  a2 
n

Thus we see that integration of an elementary fraction of type IV can be carried out
with the aid of the recurrence formula.
Thus the integral of every rational fraction (rational function) can be reduced to
integrals of partial fractions of the indicated types. Besides, the integral of a
rational fraction is always expressible in terms of elementary functions such as
rational functions, the logarithmic function and the arctangent.

In the conclusion we formulate the main stages of integration the rational fractions.
P( x)
Before integrating the rational fraction , we must perform the following
Q( x)

algebraic transformations and calculations:

I. If we are given an improper rational fraction, we have to isolate its entire rational
function (a polynomial), that is, represent in the form
P( x) M ( x)
 L( x ) 
Q( x ) Q( x)

M ( x)
Where L(x) is a polynomial and is a proper rational fraction.
Q( x)

II. Factor the denominator of the fraction into linear and quadratic multipliers:
p2
Q( x)  ( x  a) n ...( x 2  px  q) k ... , where q 0.
4
III. Decompose the proper rational fraction into partial fractions:
M ( x) A1 A1 A
  n 1
 ...  n  ...
Q( x) ( x  a) n
( x  a) xa
B x  C1 B x  C2 B x  Ck
...  2 1  2 2 k 1
 ...  2 k  ...
( x  px  q) k
( x  px  q) x  px  q

IV. Compute the undetermined coefficients A1 , A2 ,..., An , B1 , C1 , B2 , C 2 ,...Bk , C k ,... for

which purpose it is necessary to reduce the latter equality to the common
denominator, equate the coefficients in the same degrees of x on the right-hand and
left-hand sides of the identity obtained and solve the system of linear equations
with respect to the sought-for coefficients.
Thus, as a result, integration of a rational fraction will reduce to finding integrals
of a polynomial and of partial rational fractions.
Let us consider several examples.
x3
1. Find the integral x 3
x
dx .

Solution. The given proper rational fraction can be represented as a sum of partial
fractions of type I:
x3 x3 A B C
   
x  x x( x  1)( x  1) x x  1 x  1
3

First of all we must find the unknown coefficients A, B and C . Clearing the
equation of fractions, we obtain.
x  3  A( x 2  1)  Bx ( x  1)  Cx ( x  1) .

Consequently, combining the terms of like degrees, we get

 x  3  ( A  B  C) x 2  (B  C)x  A .

Since it is an identity, the coefficients in like powers of x on both sides must be

equal:
x  3  0  x 2  x  3  ( A  B  C) x 2  (B  C) x  A

Comparing the coefficients in like powers of x, we get the system of equations

A  B  C  0

B  C  1
 A  3

From which we find A  3, B  1 and C  2 .
There is a still simpler method of determined the coefficients A, B and C ; putting in
succession x  0, x  1, x  1 , we obtain  3   A,  2  2 B, and  4  2C , .that is,
again A  3, B  1 and C  2 .
Thus, decomposition of a rational fraction into partial fractions has the form
x3 3 1 2
  
x  x x x 1 x 1
3

and therefore
x3 dx dx dx x3
 x3  x dx  3 x  x 1  x 1
  2  3  ln x  ln x  1  2 ln x  1  C  ln
( x  1)( x  1) 2
C

x2 1
2. Find the integral  ( x  1) 3 ( x  3)dx .
Solution. The factor ( x  1) 3 is associated with the sum of three partial fractions
1 B C D
  , and the factor x  3 , with the partial fraction .
( x  1) 3
( x  1) 2
x 1 x3

x2 1 A B C D
Therefore    
( x  1) ( x  3) ( x  1)
3 3
( x  1) 2
x 1 x  3

Clearing the equation of fractions, we get

x 2  1  A( x  3)  B( x  1)( x  3)  C ( x  1) 2 ( x  3)  D( x  1) 3 .

Consequently,
x 2  1  A( x  3)  B( x 2  2 x  3)  C ( x 3  x 2  5 x  3)  D( x 3  3x 2  3x  1)

Combining the terms of like degrees, we get

 x 2  1  (C  D) x 3  ( B  C  3D) x 2  ( A  2 B  5C  3D) x  3 A  3B  3C  D

Comparing the coefficients in like powers of x, we get the system of equations

C   D
C  D  0 B  4D  1
 B  C  3D  1 

   A  16 D  2
 A  2 B  5C  3D  0 
3 A  3B  3C  D  1 D   5
 32
1 3 5 5
from which we find A  , B  , C  , D .
2 8 32 32

The unknowns A, B, C and D in the decomposition could have been determined in

another way. The real roots of the denominator are the numbers 1 and -3. Putting
1 5
x=1 we get 2=4A, i.e. A= . For x  3 , we have 10  64D , i.e. D   . Whence
2 32
5 3
C and B  .
32 8

Thus, finally decomposition of a rational fraction into partial fractions has the form
x2 1 1 3 5 5
   
( x  1) ( x  3) 2( x  1)
3 3
8( x  1) 2
32( x  1) 32( x  3)

Thus, we get
x2 1 1 dx 3 dx 5 dx 5 dx
 ( x  1) 3 ( x  3) dx  2  ( x  1) 3  8  ( x  1) 2  32  x  1  32  x  3 
1 3 5 x 1
   ln C
4( x  1) 2
8( x  1) 32 x  3

x 3  3x 2  5 x  7
3. Find the integral  x2  2
dx .

Solution. Since the integrand is an improper rational fraction, first of all we have to
isolate its entire rational function i.e. a polynomial:
x 3  3x 2  5 x  7 x2  2
-
x3  2x x3

3x 2  3x  7
-
3x 2 6
 3x  1

Thus, we have
x 3  3x 2  5 x  7 3x  1
 x 3 2
x 2
2
x 2
It follows that
x 3  3x 2  5 x  7  3x  1  3x  1
 x 2
2
dx    x  3  2
 x  2
dx   xdx  3 dx   2
x 2
dx 

1 2 xdx dx 1 3 d ( x 2  2) dx
 x  3x  3 2  2  x 2  3x   2  2 
2 x 2 x 2 2 2 x 2 x 2
1 3 1 x
 x 2  3x  ln( x 2  2)  arctan C .
2 2 2 2

Integration of Trigonometric Functions

I. Integrals of the form  R(sin x, cos x)dx , where R is a rational function.

Integrals of the indicated form can be reduced to integrals of rational functions

x
with the aid of the so-called universal trigonometric substitution t  tan .
2

As a result of this substitution we have

x x x
2  sin  cos 2 tan
sin x  2 2  2  2t ;
x x x 1 t 2
cos 2  sin 2 1  tan 2
2 2 2
x x x
cos2  sin 2 1  tan 2
2  1 t ;
2
cos x  2 2
x x x 1 t 2
cos2  sin 2 1  tan 2
2 2 2
x x
tan  t   arctan t  x  2 arctan t ;
2 2
2dt
dx 
1 t2
Thus, we get
 2t 1  t 2  2dt
 R(sin xç cos x)dx   R 1  t 2 , 1  t 2  
 1 t
2
  R1 (t )dt

where R1 (t ) is a rational function.

1
Example. Find the integral  1  sin x dx .
Solution. The integrand depends rationally on sin x . Using the substitution
x
t  tan , we get
2
1 1 2dt dt 2 2
 1  sin x dx   
2t 1  t 2
 2
(1  t ) 2

1 t
C  
x
C
1 1  tan
1 t 2 2
dx
Example. Find the integral  sin x .
Solution. We have
2dt
dx 1  t 2  dt  ln t  C  ln tan x  C
 sin x  2t  t

2
1 t 2

dx
Example. Find the integral  3 sin x  4 cos x  5 .
Solution. Here the integrand depends rationally on sin x and cos x . Therefore, we
x
apply the substitution t  tan ; then
2
2dt
dx 1 t 2 dt dt 2
 3 sin x  4 cos x  5  2t

1 t 2
 2 2
t  6t  9
 2
(t  3) 2

t 3
C
3  4 5
1 t 2 1 t 2
Returning to the old variable, we get
dx 2
 3sin x  4 cos x  5   x
C
tan  3
2
x
The universal substitution t  tan often leads to very complicated expressions
2
containing rational fractions. In certain particular cases it is better to use some
other substitutions such as t  sin x , t  cos x or t  tan x , which we are going to
consider here.
I. If R(sin x, cos x) is an even function with respect to sin x and cos x , that is, if
R( sin x, cos x) = R(sin x, cos x) , then the integral can be rationalized by means of
the substitution t  tan x .
 1
Example. Find the integral  cos 4
x
dx .

Solution. Using the substitution t  tan x , we get

 cos
1
4
x
dx  
1
2

1
2
cos x cos x
 
dx   tan 2 x  1 d tan x 

t3 1
  (t 2  1)dt   t  C  tan 2 x  tan x  C.
3 3

sin 2 x
Example. Find the integral  cos6 x dx .
Solution. We have
sin 2 x sin 2 x 1 1
 cos6 x  cos2 x  cos2 x  cos2 x dx   tan x  (1  tan x)  d tan x 
 2 2
dx

t3 t5 1 1
  t 2 (1  t 2 )dt   (t 2  t 4 )dt    C  tan 3 x  tan 5 x  C .
3 5 3 5
II. If R(sin x, cos x) is an odd function with respect to sin x , that is, if
R ( sin x, cos x)   R(sin x, cos x) , the integral can be rationalized by means of the

substitution t  cos x .
(sin x  sin 3 x)dx
Example. Find the integral  cos 2 x
Solution. Since the integrand is odd with respect to the sin x , we put t  cos x . If
follows that sin 2 x  1  cos2 x  1  t 2 , cos 2 x  2 cos2 x  1  2t 2  1, dt  d cos x   sin xdx .
Thus we have
(sin x  sin 3 x)dx (1  sin 2 x) sin xdx (2  t 2 )(dt)
 cos 2 x
  cos 2 x
  2t 2  1 
t2  2 1 2t 2  4 1 2t 2  1  3 1 3 dt
 2 dt   2   dt   dt   2 
2t  1 2 2t  1 2 2t  1
2
2 2 2t  1
t 3 d ( 2t ) t 3 2t  1 1 3 2 cos x  1
    
2 2 2 ( 2t )  1 2 2 2
2
ln
2t  1
 C  cos x 
2 2 2
ln
2 cos x  1
C

III. If R(sin x, cos x) is an odd function with respect to cos x , that is, if
R (sin x, cos x)   R(sin x, cos x) , then the integral can be rationalized by means of the

substitution t  sin x .
(cos3 x  cos5 x)dx
Example. Find the integral  sin 2 x  sin 4 x
Solution. Here the integrand is odd with respect to the cos x . Therefore, we use the
substitution t  sin x . Then
cos2 x  1  sin 2 x  1  t 2 , dt  d sin x  cos xdx .

Thus we have
(cos3 x  cos5 x)dx cos2 x(1  cos2 x) cos xdx (1  t 2 )(2  t 2 )
 sin 2 x  sin 4 x   sin 2 x  sin 4 x

t2  t4
dt .

Note that
(1  t 2 )(2  t 2 ) 2  3t 2  t 4 t 2  t 4  2  4t 2
  
t2 t4 t2 t4 t2 t4
2  4t 2 2(1  t 2 )  6t 2 2 6
 1 2 4  1  1 2  .
t t t (1  t )
2 2
t 1 t 2

Therefore,
(1  t 2 )(2  t 2 )  2 6  2
 t 2  t 4 dt   1  t 2  1  t 2 dt  t  t 
2
 6 arctan t  C  sin x   6 arctan(snx)  C.
sin x

II. Integrals of the form  sin m x  cosn xdx

Here we will consider two cases which are of especial significance.

Case 1. At least one of the exponent m and n is an odd positive number. If n is an
odd positive number, then we use the substitution sin x  t ; now if m is odd positive
number, the substitution cos x  t .
Example. Find the integral  sin8 x  cos7 xdx .

Solution. Putting sin x  t , cos x  dt, we obtain.

 sin x  cos xdx   sin 8 x  cos6 x  cos xdx   sin 8 x(1  sin 2 x) 3 d sin x 
8 7

  t (1  t )
8 2 3
dt   t 8 (1  3t 2  3t 4  t 6 )dt   (t 8  3t 10  3t 12  t 14 )dt 
t9 t 11 t 13 t 15 1 3 3 1
  3  3   C  sin 9 x  sin11 x  sin13 x  sin15 x  C.
9 11 13 15 9 11 13 15

Example. Find the integral  sin5 x  cos6 xdx .

Solution. Setting cos x  t ,  sin xdx  dt , we get

 sin x  cos6 xdx   sin 4 x  cos6 x  sin xdx   (1  cos 2 x) 2 cos6 x sin xdx    (1  t 2 )  t 6 dt 
5 2

t7 t 9 t 11
   (1  2t  t )t dt    (t  2t  t )dt    2    C 
2 4 6 6 8 10

7 9 11
1 2 1
  cos7 x  cos9 x  cos11 x  C.
7 9 11

Case 2. Both exponents m and n are even positive numbers. In this case the
integrand must be transformed by mean of the formulas
1
sin x  cos x  sin 2 x ;
2
1 2
sin 2 x  cos2 x  sin 2 x ;
4
1  cos 2 x 1  cos 2 x
sin 2 x  ; cos2 x  .
2 2

Example. Find the integral  sin 2 x  cos2 xdx .

Solution. We have
1 1 1  cos 4 x
 sinx  cos2 xdx   sin 2 2 xdx   dx 
2

4 4 2
1 1 1
  (1  cos 4 x)dx  x  sin 4 x  C
8 8 32

III. Integrals of the form  tan m xdx and  cot m xdx

These integrals can be found with the aid the formula

1 1
tan 2 x  1; or cot 2 x   1,
cos2 x sin 2 x
by means of which the degree of the tangent or cotangent is consecutively lowered
Example. Find the integral  tan 5 xdx .

Solution. We have
 1 
 tan xdx   tan 3 x  1dx   tan 3 xd tan x   tan 3 xdx 
5

 cos x 
2

 1 
4
tan x tan 4 x tan 2 x
   tan x  1dx    ln cos x  C.
 cos x 
2
4 4 2
 IV. Integrals of the form  sin mx cos nxdx,  cos mx cos nxdx,  sin mx sin nxdx

These integrals can be found with the aid of the following trigonometric formulas

sin   cos  
1
sin(   )  sin(   ),
2

cos  cos  
1
cos(   )  cos(   ) ,
2

sin   sin  
1
cos(   )  cos(   )
2
by means of which make it possible to represent a product of trigonometric
functions as their sum.
Example. Find the integral  sin 3x  cos7 xdx .

Solution. We have

 sin 3x  cos7 xdx  2  sin10x  sin(4 x)dx  2  sin10xdx 

1 1

1 1 1

2  sin 4 xdx   cos10x  cos 4 x  C
20 8

Lecture 13
Definite Integral
There are various problems of geometry and physics leading to the notion of the
definite integral: determining the area of a plane figure, computing the work of a
variable force, finding the distance travelled by a body with a given velocity, and
many other problems.
Definition. Suppose a function f (x ) is defined on a closed interval [a,b]. Let us
divide the interval [a,b] into n arbitrary parts by points a  x0  x1  x2  ...  xn  b ,
choose an arbitrary point  k in each subinterval xk 1 , xk  and find the length of
every such subinterval: xk  xk  xk 1 .
The integral sum for the function f (x ) on the closed interval [ a, b] is the sum of
the form
 n

 f (
k 1
k )xk  f (1 )x1  f ( 2 )x2  ...  f ( n )xn

Each term (summand) of the integral sum is equal to the area of the rectangle with
the base xk and the altitude f ( k ) and which is obviously equal to
f ( k )x k (Fig4.1).

y  f (x)
f ( k )

a  x0 x1 x2 xk 1 xk xn1 xn  b
o x
Fig 4.1
The integral sum is the area of the entire step figure consisted of these rectangles.
The definite integral of the function f (x ) on the closed interval [ a, b] is the limit of
the integral sum under the condition that the length of the greatest subinterval tends
to zero:
b n

 f ( x)dx  lim
max xk 0
 f (
k 1
k )xk
a

The numbers a and b, are called respectively, the lower limit and the upper limit
of integration. As in the case of the indefinite integral, the function f (x ) is called
the integrand, the expression f ( x )dx is called the element of integration, the
interval [ a, b] is called the interval of integration, and the variable x is the
variable of integration.
Definition. A function f (x ) for which the definite integral exists is called
integrable.
 b
The process of forming the definite integral shows that the symbol  f ( x)dx is
a
a

certain number. Its value only depends on the integrand f (x ) and on the numbers
a and b determining the interval of integration.

Geometrical interpretation of the definite integral

Let us suppose that f ( x)  0 on the interval [ a, b] .

Definition. A figure bounded by the y
y=f(x)
graph of a function y  f (x) , by the
straight lines y  0, x  a and x  b is called
a curvilinear trapezoid.
x=a x=b
If f ( x)  0 on [ a, b] , then, in geometrical
b
interpretation, the definite integral  f ( x)dx
a o a y=0 b x
is the area of a curvilinear trapezoid
(Fig.4.2): Fig.4.2
b
A   f ( x)dx
a

Definition of Darboux sums.

Let f be a bounded function defined on a closed bounded interval [a, b].
A partition P of [a, b] is any finite selection of points
a  x0  x1  ...  xn1  xn  b .
P creates a subdivision of [a, b] into subintervals xk 1 , xk .The norm of P is the
maximum length of a subinterval, P  max xk  xk 1 . We say that P′ is a
k

refinement of P if P ⊂ P′ . Let mk and M k be the infimum and the supremum of f

over xk 1 , xk .
mk  inf f x 
M k  sup f x
Consider the lower and upper Darboux sums of f corresponding to a partition P ∶
L f , P    mk xk  xk 1 
n

k 1
U  f , P    M k xk  xk 1 
n

k 1

The geometric meaning of the lower and upper Darboux sums is that they are equal
to the planar areas of stepped figures consisting of rectangles whose base widths
are xk and with respective heights mk and M k ( see Fig.) if f  0 . These figures
approximate, from the inside and outside, the curvilinear trapezium formed by the
graph of f, the abscissa axis and the rectilinear segments x  a, x  b ( which may
degenerate into points).

Properties of Darboux sums.

● Lower and upper bounds If m ≤ f ≤ M on [a, b],
then m(b − a) ≤ L(f, P) ≤ U(f, P) ≤ M(b − a).
● Monotonicity with respect to the partition If Q ⊃ P is a refinement of P, then
L(f, P) ≤ L(f, Q) and U(f, P) ≥ U(f, Q).
● Inequality of lower and upper sums.
For any two partitions P, P′ , L(f, P) ≤ U(f, P′).

Lower and upper Darboux integrals.

The lower integral of f over [a, b] is the quantity L(f) = supP L(f, P).
The upper integral of f over [a, b] is the quantity U(f) = infP U(f, P).
Note that L(f) ≤ U(f).
Riemann integrability.
One says that f is integrable in the sense of Riemann if L(f) = U(f).
b

The common value is denoted by  f x dx

a

Criterion for Riemann integrability.

A function f is Riemann integrable on [a, b] if and only if for every ε > 0 there is a
partition P of [a, b] such that U(f, P) − L(f, P) < ε.
 Basic Properties of a Definite Integral

1. A definite integral does not depend on the notation of the variable of integration,
i.e.
b b b


a
f ( x)dx   f (t )dt   f (u )du  ...
a a

b
2.  dx  b  a .
a

It is clear that any integral sum for the function f ( x)  1 is equal to b  a :

n n n

 f ( k )xk  1 xk   xk  x1  x2  ...  xn 

k 1 k 1 k 1

 x1  x0  x2  x1  ...  xn  xn1  xn  x0  b  a

3. If the upper and the lower limits of the definite integral are interchanged the
integral only changer its sign:
b a

 f ( x)dx   f ( x)dx .
a b

4. If the limits of integration coincide then the integral is equal to zero, i.e
a

 f ( x)dx  0
a

5.For any a, b, c there holds the formula

b c b

 f ( x)dx   f ( x)dx   f ( x)dx

a a c
(*)

Proof. Let a  c  b .
Since the limit of the integral sum does not depend on the way the interval [ a, b] is
divided into parts, we can divide the interval so that the point C is always a point
of division. Then the integral sum can be represented in the form
n

 f (
k 1
k
)xk   / f ( k )xk   // f ( k )xk
k k

where the first sum on the right-hand side includes all the terms corresponding to
the points of division of the interval [ a, c ] and the second sum includes all the terms
corresponding to the points of division of the interval [c, b] . The first and the
second sums are, respectively, integral sums for the function f (x ) corresponding to
the intervals [ a, c ] and [c, b] .
If the number of the points of division increases indefinitely and the length of the
greatest subinterval tends to zero for the whole interval [ a, b] the same obviously
takes place for the intervals [ a, c ] and [c, b] ; therefore, passing to the limit as max
xk  0 , we get
n
lim
max xk 0
 f (
k 1
k )xk  lim
max xk 0

k
/
f ( k )xk  lim
max xk 0

k
//
f ( k )xk

The first limit on the right-hand side is equal to the integral from a to c and the
second to the integral from c to b, and thus the required equality is obtained.
b c b

 f ( x)dx   f ( x)dx   f ( x)dx

a a c

Equality (*) also holds in the case when the point c lies outside the interval [ a, b] to
the right of it, i.e. a  b  c , or to the left of it, i.e. c  a  b .
Indeed, let a  b  c . As has been proved, since b lies between a and c we have
c b c


a
f ( x)dx   f ( x)dx   f ( x)dx
a b

whence
b c c

 f ( x)dx   f ( x)dx   f ( x)dx ,

a a b

Now, interchanging the limits of integration in the second integral on the right-
hand side we obtain
b c b


a
f ( x)dx   f ( x)dx   f ( x)dx
a c

which is what we had to prove.

It can similarly be proved that equality (*) is valid for c  a  b .
6. The integral of a sum of a finite number of functions is equal to the sum of the
integrals of these functions:
b b b

  f ( x)  g ( x)dx   f ( x)dx   g ( x)dx ,

a a ac
 b b b

  f1 ( x)  ...  f n ( x)dx   f1 ( x)dx  ...   f n ( x)dx .

a a ac

Proof. According to the definition of the integral, we have

 
b n n n

  f ( x)  g ( x)dx  maxlim
a
x 0
  f ( k )  g ( k )xk  maxlim
k
k 1
x 0 
 f ( k )xk   g ( k )xk  
k
 k 1 k 1 
n n b b
lim
max xk 0
 f (
k 1
k )xk  lim
max xk 0
 g (
k 1
k )xk   f ( x)dx   g ( x)dx
a ac

7. A constant factor in the integrand can be taken outside the integral sign:
b b

 Cf ( x)dx  C  f ( x)dx
a a

where C is constant.
Proof. According to the definition of the integral, we have
b n n b

 Cf ( x)dx  maxlim
x 0
 Cf ( k )xk C maxlim
k k 1
x 0
 f ( k )xk  C  f ( x)dx . k k 1
a a

8. If f ( x)  0 on [ a, b] then  f ( x)dx  0 .
a

Proof. Let f ( x)  0 on [ a, b] . Then in the integral sum all the summands are
nonnegative, and hence
n

 f (
k 1
k )xk  0 .

It follows that
b n

 f ( x)dx  lim
max xk 0
 f (
k 1
k )x k  0
a

since the limit of a nonnegative quantity cannot be negative.

b b

9. If f ( x)  g ( x) on [ a, b] then a
f ( x)dx   g ( x)dx .
a

Proof. We have
b
f ( x)  g ( x)  f ( x)  g ( x)  0    f ( x)  g ( x)dx  0 
a
b b b b
  f ( x)dx   g ( x)dx  0  f ( x)dx   g ( x)dx .
a a a a
10. If m and M are, respectively, the least and the greatest values of the function
f (x ) in the interval [ a , b ] ( a  b ) then
b
m(b  a )   f ( x)dx M (b  a ) .
a

Proof. We have m  f ( x)  M for any x  [a, b] .

According to the previous property we can integrate the inequality and hence
b b b

 mdx   f ( x)dx  Md x
a a a

b b b
m  dx   f ( x)dx M  d x
a a a

b
m(b  a)   f ( x)dx M  (b  a) .
a

The geometrical meaning of the inequalities we have just proved is the following:
The area of a curvilinear trapezoid is
y
greater than the area of the rectangle with
altitude equal to m and is less than the
area of the rectangle with altitude equal to
M
M (Fig.4.3)
11. If at every point x  [ a, b] the
m
inequalities
o a b x
 ( x)  f ( x)   ( x)

are fulfilled then Fig.4.3

b b b

  ( x)dx   f ( x)dx   ( x)dx .

a a a

This means that an inequality between functions implies an inequality of the same
sense between their definite integrals, or, briefly speaking, that it is allowable to
integrate inequalities.

Estimate of a Definite Integral

There holds the following theorem:
Theorem. The absolute value of the definite integral does not exceed the integral
of the absolute value of the integrand function
b b

 f ( x)dx   f x  dx .
a a

Proof. For any x we have

 f ( x)  f ( x)  f ( x)

It follows that
b b b
  f ( x) dx   f ( x)dx   f ( x) dx
a a a

that is
b b

 f ( x)dx  
a a
f ( x) dx.

Mean Value Theorem

The definite integral possesses the following important property:
Theorem. Let f (x ) be a continuous function on [ a, b] .
Then there exists at least one point x    (a, b) such that
b

 f ( x)dx  f ( )(b  a) .
a

Proof. If the function f (x ) is constant, i.e. f ( x)  C the formula is obvious

b b

 Cdx  C  dx  C (b  a)
a a

Let us suppose f (x ) is nonconstant, then according to the property 10 of the

definite integral we have
b
1
b  a a
m f ( x)dx  M

and, hence,
b
1
b  a a
f ( x)dx  
where  is a number lying between the least value m and the greatest value M of
the function f (x ) on [ a, b] .
Since f (x ) is continuous on [ a, b] according to well-known property of continuous
function the function f (x ) assumes all the intermediate values between m and M.
Therefore, there exist at least one point   (a, b) at which f ( )  
i.e.
b
1
b  a a
f ( x)dx  f ( )

In geometrical interpretation the mean value theorem shows that the definite
integral of a continuous function is equal to the area of some rectangle with
altitude f ( ) (Fig.4.4).

 

o a  b x

Lecture 14
Integral with Variable Upper Limit
Let us consider the function of the form
x
F ( x)   f (t )dt
a

which is called the integral with variable upper limit. Here the lower limit is fixed
(constant), and the upper limit x is variable.
For our further aims it is important to study some properties of this function F (x ) .
 Continuity of the Integral with Variable Upper Limit

Theorem. If f (x ) is integrable on [ a, b] , then the function F (x ) is continuous on

[a, b] .

Proof. Let us take an arbitrary x  [a, b] and give the argument x an increment x .
Let us consider the difference F ( x  x)  F ( x) . We have
x  x x

F ( x  x)  F ( x)  
a
f (t )dt   f (t )dt 
a
x x  x x x  x
.
  f (t )dt   f (t )dt   f (t )dt   f (t )dt
a x a x

Now estimating the expression F ( x  x)  F ( x) we get

x  x x  x x  x
F ( x  x)  F ( x)   f (t )dt  
x x
f (t ) dt  x
f (t ) dt

Since the function f (x ) is integrable on [ a, b] then the function is bounded on [ a, b] ,

that is
f (t )  C for any t  [a, b] , C=constant

Taking it into account, we obtain

x  x x  x

F ( x  x)  F ( x)   Cdt
x
C  dt  C  x .
x

It follows that F ( x  x)  F ( x)  0 as x  0 , and, hence lim F ( x  x)  F ( x) which

x0

means that the function F (x ) is continuous at x.

Differentiability of the Integral with Variable Upper Limit

Theorem. If a function f(x) is continuous on [ a, b] then the derivative of the
integral with respect to its upper limit is equal to the integral:
/
x 
F ( x)    f (t )dt   f ( x) .
/

a 

In other words: an integral with variable upper limit is an antiderivative of its

integrand.
 F
Proof. We have to prove that lim  f ( x) .
x 0 x

For the further it should be noted that

x  x x  x x  x
1 1
 dt  x 
x
x  dt  1  f ( x) 
x
x  f ( x)dt .
x

We have
F F ( x  x)  F ( x) 1 x  x
x

x
  f t dt .
x x
F
Let us consider the difference  f (x) .
x

F 1 x x 1  xx  1 xx 1 x x
x
 f ( x)   f (t )dt  f ( x)  x  x f (t )dt   x x f ( x)dt  x x  f (t )  f ( x)dt
x x

Since the function f (x ) is continuous, it is clear that

x  0  t  x  f (t )  f ( x)  f (t )  f ( x)  0 . It follows that for any  0

f (t )  f ( x)   and hence

F 1 xx x  x

 f ( x)    f (t )  f ( x)dt  1   f (t )  f ( x) dt 
1
   x  
x x x x x x

F
or  f (x)  
x
F
which means that lim  f ( x), i.e. F / ( x)  f ( x)
x 0 x

Theorem. Any function continuous on [a,b] possesses an antiderivative.

The Newton-Leibniz Theorem

Theorem. The definite integral is equal to the difference of the values of any
antiderivative of the integrand corresponding to the upper and lower of integral.
b

 f ( x)dx  (b)  (a) , where ( x)  f ( x) .

a

In other words:
The definite integral is equal to the increment of any antiderivative of the integrand
gained on the interval of integration.
This result is one of the most important theorem in mathematics. It is called the
Newton-Leibniz theorem, and the equality
b

 f ( x)dx  (b)  (a)

a

is referred to as the Newton-Leibniz formula.

Proof. Let us consider the function
x
F ( x)   f (t )dt
a

and let  (x) be arbitrary antiderivative of the function f (x ) . Since any two
antiderivatives differ by a constant we get
F ( x)  ( x)  C

or
x

 f (t )dt  ( x)  C
a

Here putting x  a we obtain

x

 f (t )dt  ( x)  C
a
or 0   (a)  C , C   ( a )

and hence
x

 f (t )dt  ( x)  (a) .
a

Now setting x  b we finally get

x

 f (t )dt  (b)  (a)

a

which is what we had to prove.

Methods of Evaluating Definite Integral

1.Newton-Leibniz formula:
x

 f ( x)dx  F ( x)  F (b)  F (a ) ,
b
a
a
where F (x ) is an antiderivative of f (x ) , i.e. F / ( x)  f ( x) .
2. Integration by parts:
b b

 udv  uv   vdu ,
b
a
a a

where u  u (x) are continuously differentiable functions on the interval [a,b].

Proof. We have
d (uv)  vdu  udv

Integrating both sides of this equality from a to b we obtain

b b b

 d (uv)  vdu   udv

a a a

whence
b b b

 udv  d (uv)   vdu

a a a

or
b b

 udv  uv   vdu .
b
a
a a

3. Change of variable in the definite integral:

b 


a
f ( x)dx   f [ (t )]   / (t )dt

(1)

where x   (t ) and  / (t ) are continuous on interval [  ,  ],  ( )  a,  (  )  b,

and f [ (t )] is a function continuous on [  ,  ].
Proof. Let us suppose F (x ) is an antiderivative of f (x) . Then

 f ( x)dx  F ( x)  C , (2)

 f [ (t )]   (t )dt  F [ (t )]  C .
/
(3)

From (2) we get:

b

 f ( x)dx  F ( x)  F (b)  F (a)

b
a
a
From (3) we obtain:

(t )dt  F  t    F [ (  )   ( )]  F (b)  F (a)

 f [ (t )]  
/

Comparing the equalities we get the formula (1).

4. If f (x ) is an odd function, i.e. f ( x)   f ( x) , then
a

a
 f ( x)dx  0 .
Proof. We have
a 0 a

a
 f ( x)dx   f ( x)dx   f ( x)dx .
a 0

Now, changing the variable of integration in the first integral on the right-hand side
by means of the substitution x  t we get
a a a a a


a
f ( x)dx   f (t )dt   f ( x)dx    f (t )dt   f ( x)dx  0 .
0 0 0 0

5. If f (x ) is an even function, i.e. f ( x)  f ( x) , then

a a


a
f ( x)dx  2 f ( x)dx
0

Proof. We do the same here

a 0 a a a a a a

 f ( x)dx   f ( x)dx   f ( x)dx   f (t )dt   f ( x)dx   f (t )dt   f ( x)dx  2   f ( x)dx .

a a 0 0 0 0 0 0

Let us consider several examples.

2
1.Evaluate the integral  x 2 dx .
1

Solution. We have
2
2 3 (1) 3 8 1
2
x3
      3
2
x dx
1
3 1
3 3 3 3

2
2. Evaluate the integral  cos xdx .
0

Solution. We have
 

2

 cos xdx  sin x 02  sin
0
2
 sin 0  1  0  1 .


4
dx
3. Evaluate the integral  cos
0
2
x
.

Solution. We have


4
dx 
0 cos2 x  tan x 04  tan 4  tan 0  1  0  0 .
1
4. Calculate the integral  xe  x dx .
0

Solution. We use here the method of integration by parts.

We put u  x, dv  e  x dx , whence du  dx, v  e  x .
Then
e2
1 1

 xe dx   xe   e  x dx  e 1  e  x
x x 1 1
 2e 1  1  .
0
0
0
0 e

2
5. Calculate the integral  e x cos xdx .
0

Solution. We use here method of integration by parts.

We set u  e x , dv  cos xdx , whence du  e x dx, v  sin x . Then
  
2  2  2

 e cos xdx  e  sin x 02   e sin xdx  e 2   e sin xdx

x x x x

0 0 0

Applying here the same method, we put e x  u, sin xdx  dv , whence

du  e x dx, v   cos x . Then we obtain
  
2   2  2

e cos xdx  e  e  cos   e cos xdx  e  1   e x cos xdx .

x 2 x 2 x 2
0
0 0 0

It follows that
 
2
e 2 1
0 
x
e cos xdx .
2
 ln 2 x
e

6. Calculate the integral 1 x dx .

Solution. We use here the method for change of variable.
dx
We put ln x  t , then  dt .
x

It is clear that if x  1, then t  0 ; if x  e, then t  1 .

Consequently,
e 1 1
ln 2 x t3 1 1
0 x dx  0 t dt  3  0  .
2

0
3 3
8
xdx
7. Evaluate the integral 
3 1 x
.

Solution. We set 1  x  t ;then x  t 2  1, dx  2tdt .

If x  3 , then t  2 ; if x  8 ,then t  3 . Therefore,
3
(t 2  1)  2tdt
8 3 3
xdx t3
   2 (t 2  1)dt  2   2t 2 
3

3 1 x 2 t 2
3 2

2 2 32 2
 (27  8)  2(3  2)   19  2   10
3 3 3 3

8. Calculate the integral  sin 3 x cos2 dx .


Solution. Here the integrand is an odd function.



 sin x cos2 dx  0
3

LECTURE 15
Improper Integrals
Up to now, when speaking of definite integrals we assumed that the interval of
integration was finite and closed and that the integrand was continuous.
Improper integrals are:
(1) integrals with infinite limits of integration;
(2) integrals of unbounded functions.

Improper Integrals with Infinite Limits

We introduce the following definition

 
Definition. The improper integral  f ( x)dx
a
of the function f (x ) over the interval
b
[a;) is the limit of the integral  f ( x)dx as b   provided
a
that this limit exists:
 b


a
f ( x)dx  lim  f ( x)dx
b  
a

If the limit exists and is finite, the improper integral is called convergent; now if
the limit does not exist or is equal to infinity, it is called divergent.
Similarly,
b b

 f ( x)dx  lim  f ( x)dx


a 
a

and
 b



f ( x)dx  lim
a    f ( x)dx
b   a

Examples.


e
x
1.Compute the improper integral dx .
0

Solution. We have
 b b

 e dx  lim  e dx  lim(e ) 
x x x
b   b   0
0 0
.
 1
 blim (  e b
 e 0
)  lim  1  b 
1
  b  
 e 


e
x
Consequently, the improper integral dx is equal to 1 and hence, it is
0

convergent.
1
dx
2. Compute the improper integral x

2
.
Solution. We find
1
1
dx 1
dx  1  1
 x 2  alim
  
a x
2
 lim
a  
    alim
 xa  
1 

 1
a
that is, the improper integral converges.

dx
3. Compute the improper integral 1 x

2
.
Solution. The integrand is even, therefore
 
dx dx
1  x 2  2 0 1  x 2 .
Then we have
  b
dx dx
0 1  x 2 b 0 1  x 2  blim
 
b
lim  
arctan x 0


 blim
 
(arctanb  arctan 0)  blim
 
arctan b 
2

dx
Thus, finally we get 1 x

2
 ,

That is, the improper integral is convergent.


4. Compute the improper integral  cos xdx (or establish its divergence).
0

Solution. We have
b b
lim  cos xdx  lim sin x  lim (sin b  sin 0)  lim sin b .
b b  0 b  b 
0

The limit lim sin b does not exist since sin b oscillates between -1 and 1 as b   .
b 

Consequently, the improper integral is divergent.

Comparison Tests for Convergence or Divergence of Improper Integrals with

Infinite Limits

To investigate the convergence of improper integrals, we use one of the

comparison tests. To this end we will prove the following theorems (comparison
tests).
Theorem 1. Let the inequalities
0  f ( x)   ( x)
hold for all the values of x  a .
 
Then, if the integral   ( x)dx is convergent the integral
a
 f ( x)dx
a
is convergent as

well.

Proof. Let us suppose that the integral   ( x)dx
a
is convergent and is equal to M. It

is clear that
b 

  ( x)dx    ( x)dx  M for any b

a a

and, hence
b

  ( x)dx  M
a
for any b.

According to the property on integrating inequalities, we get

b b
f ( x)   ( x)   f ( x)dx    ( x)dx M .
a a
 b
It follows that I (b)   f ( x)dx is a bounded function, and therefore the integral
a


 f ( x)dx is convergent. The theorem has been proved.

a

Theorem 2. Let the inequalities

0  f ( x)   ( x)
hold for all the values of x  a .
 
Then, if the integral 
a
f ( x)dx is divergent the integral   ( x)dx is also divergent.
a

Proof. Now suppose that the integral  f ( x)dx is divergent.
a
b
Then the increasing function  f ( x)dx
a
tends to   as b   . But since
b b

  ( x)dx  f ( x)dx ,
a a
b 
the function   ( x)dx also tends to   as b   and hence, the integral   ( x)dx is
a a

divergent, which is what we had to prove.


dx
Example 1. Test the integral x
a
p
for convergence.

Solution. If p=1 we have by definition

 A
dx dx
a x  Alim
   x
A
 lim ln x  lim (ln A  ln a )  
A  a A 
a

dx
since ln A   as A   . Consequently, at p=1 the integral 
a
x
is divergent.

If p  1 we have by definition
A

dx A
dx A
p  1 
 x p A  x p A  x dx  Alim
 lim  lim  x  p 1  
   p  1
a a a  a
 1 1 
 lim  A  p 1  a  p 1  
A   p  1  p 1
 
1 1 1 1
 lim p 1   p 1
 p  1 A A p 1 a
1
Assume that p  1 , then p  1  0 and Alim  0.
  A p 1
It follows that at p  1 the integral converges and

dx 1 1
x
a
p
  p 1 .
p 1 a
1
Assume now that p  1 , then p  1  0 and Alim   . This means that at p  1
  A p 1
the integral diverges.
Therefore, finally we get
 1 1

dx   p 1 at p  1
a x p   p  1 a
  at p  1


Thus, we have

dx converges at p  1
x
a
p

diverges at p  1.

dx
Example 2. Test the integral 1 x
1
8
for convergence.
Solution. The integrand
dx
f ( x) 
1  x8
is smaller over the interval of integration than  ( x)  18 since
x
1 1
 8,
1 x 8
x

dx
and the integral x
1
8
is convergent (see the previous example).
Therefore, using the comparison test (theorem 1), it follows that, the given integral
converges too.

dx
Example 3. Test the integral  1 1 x  3 1 x2
for convergence.

Solution. It is clear that the inequality

1 1 1
 
1  x  3 1  x2
1 2 7

x x
2 3
x6

dx 7
holds for all x  1, and the integral 
1
7
is convergent since p
6
 1.
6
x
Consequently, applying the comparison test (theorem 1), we get that the given
integral is also convergent.

Example 4. Test the integral  x dx for convergence.
11 x

Solution. It is obvious that the inequality

x x 1
 
1 x x  x 2 x

dx 1
holds for all x>1 and the integral 
1 x
is divergent because p 
2
 1 . Therefore,

according to the comparison test (theorem 2), the given integral is also divergent.
 
Theorem 3. If the integral 
a
f ( x) dx of the absolute value of the function f (x ) is


convergent the integral  f ( x)dx
a
is also convergent.

In this case the integral  f ( x)dx
a
is said to be absolutely convergent. Therefore,

theorem 3 means that the absolutely convergent integral is also convergent.

 
cos x sin x
Example 5. The integrals  dx and 1 x dx are absolutely convergent
0 1 x
2 2
0

because the absolute values of the integrands do not exceed the positive function

1 dx
1 x2
and the integral 1 x
0
2
is convergent. Indeed, we have

cos x cos x 1
 
1 x 2
1 x 2
1 x2
and, similarly,

sin x 1

1 x 2
1 x2
By definition

dx dx
b

0 1  x 2  blim   lim arctan x 0  lim arctan b 
b

  1  x 2 b   b   2
a

and hence, the integral is convergent.

Now, applying the comparison test (theorem 1), we get that the integrals
 
cos x sin x
1 x
0
2
dx and 1 x
0
2
dx

are convergent, and it means that the integrals

 
cos x sin x
0 1  x 2 dx and 1 x
0
2
dx

are absolutely convergent.

It is easy to prove the following theorem.
 
Theorem 4. If the integral 

f ( x) dx is convergent, the integrals  f ( x) cos xdx

and


 f ( x) sin xdx are absolutely convergent.



Proof. Indeed, it clear that the absolute values of the integrands obviously do not
exceed f (x)
f ( x) cos x  f ( x)  cos x  f ( x)
f ( x) sin x  f ( x)  sin x  f ( x) .
Therefore, according to the comparison test (theorem 1) we get that both given
integrals are absolutely convergent, which is what we had to prove.
 
It should be noted, the fact that the integral  f ( x) dx is divergent does not enable
a

us to judge upon the convergence of the integral  f ( x)dx because it can be either
a

divergent or convergent.

Definition. The improper integral  f ( x)dx is said to be conditionally (not
a

absolutely) convergent if the integral  f ( x)dx is convergent and the integral
a



a
f ( x) dx is divergent.

sin x
Example 6. Let us consider the Dirichlet integral 
0
x
dx . It can be shown that

the Dirichlet integral is conditionally convergent because this integral is

convergent and its numerical value is found by means of some special techniques:

sin x 

0
x
dx 
2
while the integral of the absolute value of the integrand is divergent:

sin x

0
x
dx   .

Thus, the Dirichlet integral is conditionally divergent.

Improper Integrals of Unbounded Functions

Let y  f (x) be a continuous function for all x  [a, b) (i.e., a  x  b ) having an

infinite discontinuity at x  b . It is clear that the ordinary definition of the definite
integral is inapplicable here because the function f (x ) is unbounded in [a,b]. Let us
first take the ordinary definite integral
b 
I ( )   f ( x)dx
a
where   0

and then make  tend to zero: lim I ( ) . Then I ( ) either tends to a finite limit or has
 0

no finite limit (in the latter case it either tends to infinity or has no limit at all).
We introduce the following definition
b
Definition. The improper integral  f ( x)dx of a function
a
f (x ) continuous for
b 

a  x  b and unbounded as x  b is the limit of the integral  f ( x)dx

a
as

  0 (  0) provided this limit exists:

b b 

 f ( x)dx  lim
a
  f ( x)dx ,
0
a
 0
If the limit exists we say that the improper integral is convergent. If this limit does
not exist the equality becomes meaningless, and the improper integral written on
the left is said to be divergent.
Similary, if the function f (x ) has an infinite discontinuity only at x  a of the
interval [a,b] and is continuous for all x  (a, b] (i.e., a  x  b ) then, we put by
definition
b b


a
 0 
f ( x)dx  lim f ( x)dx, where   0 .
a 

provided this limit exists.

Definition. If the function f (x ) possesses an infinite discontinuity at a point
c  (a, b) and is continuous at a  x  c and c  x  b , then, by definition
b c  b


a
f ( x)dx  lim
 0 
a
f ( x)dx  lim
 0  f ( x)dx .
c
b
The improper integral  f ( x)dx (where
a
f ( c )  , a  c  b ) is said to be convergent,

if both limits exist on the right-hand side of the equality, and divergent, if either of
them does not exist.
Now let us consider some examples.
1
dx
Example 7. Compute the improper integral 
0
x
.
1
Solution. The integrand f ( x)  at the point x  0 is unbounded and, therefore, we
x
have
1 1 1
dx dx
0 x  0 0 x  0   lim
 lim  lim ln x
 0
(ln1  ln  )   ,

that is, the improper integral diverges.

Examples. Calculate the following improper integrals:
a a a
dx dx
1. 
0
 lim   lim 2 x  lim(2 a  2  )  2 a .
x  0  x  0   0

1 1
dx dx
 1  x  0  1  x  lim
 lim
 0
 ln(1  x)10 
2. 0 0

 lim ln   ln1   lim ln   

 0  0

That is, the integral is divergent

1
dx
3. 
1 1  x2
. In this example the integrand has infinite discontinuities at both end
points of the interval of integration. We have
 1
1 1
dx dx
  lim  lim arcsin x 
 0  0
1 1 x2  0 1 1 x2  0 1

 limarcsin(1   )  arcsin(1   ) 
 0
 0

 lim arcsin(1   ) lim arcsin(1   ) 

 0  0

  
 arcsin1  arcsin(1)      .
2  2

TEST
1. Compute the following improper integrals with infinite limits (or establish their
divergence)

dx
1. x
1
4

A)
1
B)
1
C)
1
D)
1 E) diverges
3 2 4 5


e
 ax
2. dx (a  0)
0

A) 6 B)
1
C)
1 D) diverges E)
1
5 2a a


2 xdx
3. x

2
1

A) 2 B) diverges C)
1 D)0,6 E) 7
2

dx
4. x

2
 2x  2

A)  B) 2 C) diverges D) 3 E) 4
 

5.  xe x dx
2

A) 0,8 B) diverges C) 0,5 D) 0,6 E) 0,7


ln x
6. 
2
x
dx

A) diverges B)
1
C)
1
D)
1
E)
1
2 3 4 5

7.Compute the following improper integrals of unbounded functions (or establish

their divergence)
1
dx

0 1 x2

A)  B)

C)
 D) 2 E)0
2 3

2
8.  2 dx
0 x  4x  3

A)0 B) 1 C) diverges D) 5 E) 6

2
xdx
9. 
1 x 1

A) 1 B)
8
C) 5
1 D) 2 E) 0
3 3
1
10.  x ln xdx
0

A)  1 B) C)diverges 1 1
D)  E) 
1 4 5

2

Answers:
1)A 2)E 3)B 4)A 5)C
6)A 7)B 8)C 9)B 10)D