Assiut University

Faculty of Science

Department of Mathematics

Course Title: Mathematics (II)

Course Code: MATH121

Course hours per week:

Lecture Tutorial / Total
Practical
3 2 5

For
Faculty of Computers & Information
By
Professor
OSAMA RASHED SAYED

-0-
 Table of Contents
CHAPTER I: MATRICES
Matrices 6
Exercise Set (1.1) 19
Elementary Row Operations 21
Exercise Set (1.2) 26
Inverse matrix 27
Exercise Set (1.3) 37
Special matrices 36
Exercise Set (1.4) 46
CHAPTER II: SYSTEMS OF LINEAR EQUATIONS
Introduction of systems of linear equations 47
Exercise Set (2.1) 67
Homogeneous system of linear equations 70
Exercise Set (2.2) 76
CHAPTER III: DETERMINATES
Determinants 78
Exercise Set (3.1) 84
Cramer' Rule 85
Exercise Set (3.2) 87
CHAPTER VII: LINEAR SPACE
Euclidean n-space 89
Linear Space 91
Exercise Set (4.2) 103
Subspaces 106
Exercise Set (4.3) 113
Linear combination and spanning sets 115
Exercise Set (4.4) 125
Linear Independence 127
Exercise Set (4.5) 135
Basis and Dimension 136
Exercise Set (4.6) 150

-1-
CHAPTER 8: INNER PRODUCT SPACES
Inner product 154
Exercise Set (5.1) 162
Orthogonally 164
Exercise Set (5.2) 173
Orthonormal Basis 176
Exercise Set (5.3) 187
EigenValues and EigenVectors 188
Exercise Set (5.4) 200
CHAPTER VI: PARTIAL FRACTIONS & SEQUENCES
Partial Fractions Decomposition 202
Exercise Set (6.1) 210
Sequence 211
Exercise Set (6.2) 223
Monotone Sequences 227
Exercise Set (6.3) 236
CHAPTER IV: INFINITE SERIES
Infinite Series 239
Exercise Set (7.1) 250
Convergence; The Integral Test 253
Exercise Set (7.2) 262
Additional Convergence Tests 264
Exercise Set (7.3) 269
Applying the Comparison Test 272
Exercise Set (7.4) 281
Alternating Series; Conditional Convergence 284
Exercise Set (7.5) 297
Power Series 300
Exercise Set (7.6) 309
Taylor and Maclaurin Series 311
Exercise Set (7.7) 323

-2-
CHAPTER V: DIFFERENTIAL EQUATION
Introduction 326
Definitions and Terminology 326
Initial-Value and Boundary-Value Problems 339
Solutions of First Order Differential Equations 347
Exercise Set (8.1) 360
MISCELLANEOUS PROBLEMS 363
REFERENCES

-3-
CHAPTER (I)
MATRICES

-4-
 CHAPTER (I)
MATRICES
1. MATRICES

Idea of matrices was introduced by Arthur Caylet in 19th century.

Rectangular arrays of real numbers arise in many
contexts other than as augmented matrices for systems of
linear equations. In this section, we shall consider such
arrays as objects and develop some of their properties for
use in our later work.
Definition 1.
A matrix is a rectangular array of numbers or square grid
of numbers arranged into rows and columns. The
numbers in the arrays are called the entries or elements.
Example 1.
Some examples of matrices are.
1 2 −√2 𝜋 𝑒 1
[ ]
[ 3 0] , 2 1 0 − 3 , [ 3 1⁄2 0] , [ 3] , [ 4 ]
−1 4 0 0 0

●If 𝑎𝑖𝑗 , where 1 ≤ 𝑖 ≤ 𝑚 and 1 ≤ 𝑗 ≤ 𝑛 is the elements

of the matrix 𝐴, then we write

-5-
●We say that 𝐴 is a matrix of size (order) 𝑚 × 𝑛 (or m by
n matrix) where 𝑚 is the number of the rows of 𝐴 and n
is the number of the columns of 𝐴. So, in Example 1 the
matrices are of size 3 × 2 , 1 × 4, 3 × 3, 2 × 1, and
1 × 1, respectively.
● We will use R to denote for row and C for column.
●We will use capital letters A, B, C, … to denote for
matrices.
●We write 𝐴 = [𝑎𝑖𝑗 ] or 𝐴 = [𝑎𝑖𝑗 ] if the size is
𝑚×𝑛

known.
●We say that 𝐴 is a square matrix of size n, if 𝑛 = 𝑚.
●An 1 × 𝑛 matrix is called a row matrix (or, row vector)
and written as [𝑎1 𝑎2 … 𝑎𝑛 ] and 𝑛 × 1 matrix is called a

-6-
 𝑎1
𝑎2
column matrix (or, vertical vector) and written as [ ⋮ ].
𝑎𝑛
In Example 1 the 2 × 1 matrix is a column matrix, the
1 × 4 matrix is a row matrix, and 1 × 1 matrix is both a
row matrix and a column matrix.
Definition 2.
Let 𝐴 = [𝑎𝑖𝑗 ] and 𝐵 = [𝑏𝑖𝑗 ] be two matrices of the same
size. Then 𝐴 = 𝐵 if and only if 𝑎𝑖𝑗 = 𝑏𝑖𝑗 for all 𝑖 and 𝑗.
Example 2.
𝑎 𝑏 1 1 3
The matrices 𝐴 = [ ] and 𝐵 = [ ] are not
𝑐 𝑑 −2 0 5
equal because they are not of the same size.
5 −1
But, 𝐶 = [ ] is equal to 𝐴 if and only if 𝑎 = 5,
7 0
𝑏 = −1, 𝑐 = 7 and 𝑑 = 0.◄
Definition 3.
Let 𝐴 = [𝑎𝑖𝑗 ] and 𝐵 = [𝑏𝑖𝑗 ] be two matrices of size
𝑚 × 𝑛. Then 𝐴 + 𝐵 is of size 𝑚 × 𝑛 and is defined as
𝐴 + 𝐵 = [𝑎𝑖𝑗 + 𝑏𝑖𝑗 ].

-7-
Example 3.
1 3 −3 −3 −2 0
[−1 4 ]+[ 0 4 ] = [−1 8].
0 −5 5 6 5 1
Example 4.
𝑎 3 4 8 1 11
Let [ ]+[ ]=[ ]. Find 𝑎, 𝑏 and 𝑐.
1 𝑏 𝑐 −1 5 −4
Solution:
𝑎 3 4 8 𝑎+4 11
Since [ ]+[ ]=[ ],
1 𝑏 𝑐 −1 1+𝑐 𝑏−1
𝑎 + 4 = 1, 1 + 𝑐 = 5, 𝑏 − 1 = −4. Therefore, we
obtain 𝑎 = 𝑏 = −3 and 𝑐 = 4.◄
Definition 4.
A matrix whose elements are zeros is called a zero-
matrix and is denoted by 0 (or [0]).
Definition 5.
Let 𝐴 = [𝑎𝑖𝑗 ] be a matrix and 𝑘 ∈ ℝ. The matrix 𝑘𝐴 is
defined as 𝑘𝐴 = [𝑘𝑎𝑖𝑗 ].
Notes:
●If 𝑘 = −1 , we write – 𝐴 instead of (−1)𝐴.
●𝐴 − 𝐵 = 𝐴 + (−𝐵).

-8-
Theorem 1.
Let [𝑎𝑖𝑗 ] , 𝐵 = [𝑏𝑖𝑗 ] and 𝐶 = [𝑐𝑖𝑗 ] be three 𝑚 × 𝑛
matrices; and 𝑟, 𝑠 ∈ ℝ . Then
(1) 𝐴 + 𝐵 = 𝐵 + 𝐴;
(2) 𝐴 + (𝐵 + 𝐶 ) = (𝐴 + 𝐵) + 𝐶;
(3) 𝐴 + 𝟎 = 𝐴;
(4) 𝐴 + (−𝐴) = 𝟎;
(5) 𝑟(𝐴 + 𝐵 ) = 𝑟𝐴 + 𝑟𝐵;
(6) (𝑟 + 𝑠)𝐴 = 𝑟𝐴 + 𝑠𝐴;
(7) (𝑟𝑠)𝐴 = 𝑟(𝑠𝐴);
(8) 1𝐴 = 𝐴.

Proof.
(1) Note that:
𝐴 + 𝐵 = [𝑎𝑖𝑗 ] + [𝑏𝑖𝑗 ] = [𝑎𝑖𝑗 + 𝑏𝑖𝑗 ]
= [𝑏𝑖𝑗 + 𝑎𝑖𝑗 ] = 𝐵 + 𝐴.
Therefore 𝐴 + 𝐵 = 𝐵 + 𝐴.
(2) 𝐴 + (𝐵 + 𝐶 ) = [𝑎𝑖𝑗 ] + ([𝑏𝑖𝑗 ] + [𝑐𝑖𝑗 ])
= [𝑎𝑖𝑗 ] + [𝑏𝑖𝑗 + 𝑐𝑖𝑗 ]
= [𝑎𝑖𝑗 + (𝑏𝑖𝑗 + 𝑐𝑖𝑗 )]

-9-
 = [(𝑎𝑖𝑗 + 𝑏𝑖𝑗 ) + 𝑐𝑖𝑗 ]
= [𝑎𝑖𝑗 + 𝑏𝑖𝑗 ] + [𝑐𝑖𝑗 ]
= ([𝑎𝑖𝑗 ] + [𝑏𝑖𝑗 ]) + [𝑐𝑖𝑗 ]
= (𝐴 + 𝐵) + 𝐶.
Therefore 𝐴 + (𝐵 + 𝐶 ) = (𝐴 + 𝐵) + 𝐶.
(3) 𝐴 + 𝟎 = [𝑎𝑖𝑗 ] + [0] = [𝑎𝑖𝑗 + 0] = [𝑎𝑖𝑗 ] = 𝐴.
Therefore 𝐴 + 𝟎 = 𝐴.
(4) 𝐴 + (−𝐴) = [𝑎𝑖𝑗 ] + [−𝑎𝑖𝑗 ] = [𝑎𝑖𝑗 − 𝑎𝑖𝑗 ]
= [0] = 𝟎.
Therefore 𝐴 + (−𝐴) = 𝟎.
(5) 𝑟(𝐴 + 𝐵) = 𝑟([𝑎𝑖𝑗 ] + [𝑏𝑖𝑗 ]) = 𝑟[𝑎𝑖𝑗 + 𝑏𝑖𝑗 ]
= [𝑟(𝑎𝑖𝑗 + 𝑏𝑖𝑗 )] = [𝑟𝑎𝑖𝑗 + 𝑟𝑏𝑖𝑗 ]
= [𝑟𝑎𝑖𝑗 ] + [𝑟𝑏𝑖𝑗 ] = 𝑟[𝑎𝑖𝑗 ] + 𝑟[𝑏𝑖𝑗 ]
= 𝑟𝐴 + 𝑟𝐵.
Therefore 𝑟(𝐴 + 𝐵) = 𝑟𝐴 + 𝑟𝐵.
(6) (𝑟 + 𝑠)𝐴 = (𝑟 + 𝑠)[𝑎𝑖𝑗 ] = [(𝑟 + 𝑠)𝑎𝑖𝑗 ]
= [𝑟𝑎𝑖𝑗 + 𝑠𝑎𝑖𝑗 ] = [𝑟𝑎𝑖𝑗 ] + [𝑠𝑎𝑖𝑗 ]
= 𝑟[𝑎𝑖𝑗 ] + 𝑠[𝑎𝑖𝑗 ] = 𝑟𝐴 + 𝑠𝐴.
Therefore (𝑟 + 𝑠)𝐴 = 𝑟𝐴 + 𝑠𝐴.

- 10 -
(7) (𝑟𝑠)𝐴 = (𝑟𝑠)[𝑎𝑖𝑗 ] = [((𝑟𝑠)𝑎𝑖𝑗 )]

= [(𝑟(𝑠𝑎𝑖𝑗 ))] = 𝑟[𝑠𝑎𝑖𝑗 ] = 𝑟(𝑠[𝑎𝑖𝑗 ])

= 𝑟(𝑠𝐴).
Therefore (𝑟𝑠)𝐴 = 𝑟(𝑠𝐴).
(8) Obvious. ◄
Example 5.
0 −1 −1 0
Let 𝐴 = [2 5 ] and 𝐵 = [ 5 −1]. Then
1 3 3 2
0 −1 −1 0 1 −3
3𝐴 − 𝐵 = 3 [2 5 ] − [ 5 −1] = [1 16 ].
1 3 3 2 0 7
Example 6.
Solve the following matrix equation.
0 −2 2 −1
𝐴 + 3[ ]=[ ].
1 −1 4 0
Solution:
Note that the size of the matrix 𝐴 should be 2 × 2. So, we
assume that.
𝑎 𝑏 2 −1 0 −6 2 5
𝐴=[ ]=[ ]−[ ]=[ ].◄
𝑐 𝑑 4 0 3 −3 1 3

- 11 -
Definition 6.
If 𝐴 = [𝑎𝑖𝑗 ] is any 𝑚 × 𝑛 matrix, then the transpose of
𝐴, denoted by 𝐴𝑇 , is the 𝑛 × 𝑚 matrix obtained from 𝐴 by
interchanging the rows and columns of 𝐴, 𝑖. 𝑒. 𝐴𝑇 = [𝑎𝑗𝑖 ].

Theorem 2.
Let 𝐴 = [𝑎𝑖𝑗 ] and 𝐵 = [𝑏𝑖𝑗 ] be an 𝑚 × 𝑛 matrices. The
following statements are true.
(1) (𝐴𝑇 )𝑇 = 𝐴.
(2) (𝑘𝐴)𝑇 = 𝑘𝐴𝑇 ,where𝑘 ∈ ℝ.
(3) (𝐴 + 𝐵)𝑇 = 𝐴𝑇 + 𝐵𝑇 .

𝑇
Proof. (𝐴𝑇 )𝑇 = [𝑎𝑗𝑖 ] = [𝑎𝑖𝑗 ] = 𝐴;
𝑇
(1) (𝑘𝐴)𝑇 = [𝑘𝑎𝑖𝑗 ] = [𝑘𝑎𝑗𝑖 ] = 𝑘[𝑎𝑗𝑖 ] = 𝑘𝐴𝑇 ;
𝑇
(2) (𝐴 + 𝐵)𝑇 = [𝑎𝑖𝑗 + 𝑏𝑖𝑗 ] = [𝑎𝑗𝑖 + 𝑏𝑗𝑖 ]
= [𝑎𝑗𝑖 ] + [𝑏𝑗𝑖 ] = 𝐴𝑇 + 𝐵𝑇 .
♣ We noted that sum of matrices is nearly like the
addition of real numbers. But the product of matrices is

- 12 -
not like the product of real numbers. Before we define the
product of matrices, we define the Euclidean product.
Definition 7.
𝑏1
𝑏
Let 𝐴 = [𝑎1 𝑎2 … 𝑎𝑛 ] be a row vector and 𝐵 = [ 2 ] be a
:
𝑏𝑛
column vector. We define the Euclidean product of 𝐴 and
B as follows:
𝐴𝐵 = 𝑎1 𝑏1 + 𝑎2 𝑏2 + ⋯ + 𝑎𝑛 𝑏𝑛 .
Definition 8.
Let 𝐴 = [𝑎𝑖𝑗 ] be an 𝑚 × 𝑛 matrix and 𝐵 = [𝑏𝑖𝑗 ] be an
𝑛 × 𝑝 matrix.
The product of 𝐴 and 𝐵, written as 𝐴𝐵, is the 𝑚 × 𝑝
matrix 𝐶 = [𝑐𝑖𝑗 ] whose 𝑖𝑗 - component is
𝑛

𝑐𝑖𝑗 = 𝑎𝑖1 𝑏1𝑗 + 𝑎𝑖2 𝑏2𝑗 + ⋯ + 𝑎𝑖𝑛 𝑏𝑛𝑗 = ∑ 𝑎𝑖𝑘 𝑏𝑘𝑗

𝑘=1

Note that 𝐴𝐵 is defined if and only if 𝐴 has the same

number of columns as 𝐵 has rows.
If 𝐴 is an 𝑛 × 𝑛 matrix, then 𝐴𝐴 is defined and denoted
by 𝐴2 .

- 13 -
Example 6.

Example 7.
1 2
0 −1 2
Let 𝐴 = [−1 0] and 𝐵 = [ ].
−4 1 0
0 3
Compute 𝐴𝐵 and 𝐵𝐴, if it is possible?
Solution:
Since 𝐴 is of size 3 × 2 and 𝐵 is of size 2 × 3, then 𝐴𝐵
is of size 3 × 3 and 𝐵𝐴 is of size 2 × 2. Now
1 2 −8 1 2
0 −1 2
𝐴𝐵 = [−1 0] [ ] = [ 0 1 −2].
−4 1 0
0 3 −12 3 0
1 2
0 −1 2 1 6
Also 𝐵𝐴 = [ ] [−1 0] = [ ].◄
−4 1 0 −5 −8
0 3

- 14 -
Notes.
(1) 𝐴𝐵 ≠ 𝐵𝐴, 𝑖. 𝑒., The product of matrices is not
commutative.
(2) It is possible that that 𝐴 ≠ 𝟎 and 𝐵 ≠ 𝟎 such that
𝐴𝐵 = 𝟎.

1 0 0 0
For example, 𝐴 = [ ],𝐵 = [ ] and
0 0 0 1
0 0
𝐴𝐵 = [ ].
0 0
Theorem 3.
Let 𝐴 = [𝑎𝑖𝑗 ], 𝐵 = [𝑏 𝑖𝑗 ] and 𝐶 = [𝑐𝑖𝑗 ] be matrices with
suitable size such that the following operations are
defined and 𝑘 ∈ ℝ. Then:
(1) 𝐴(𝐵𝐶 ) = (𝐴𝐵 )𝐶
(2) 𝐴(𝐵 + 𝐶 ) = 𝐴𝐵 + 𝐴𝐶
(3) (𝐵 + 𝐶 )𝐴 = 𝐵𝐴 + 𝐶𝐴
(4) 𝑘 (𝐴𝐵) = (𝑘𝐴)𝐵 = 𝐴(𝑘𝐵 )
(5) (𝐴𝐵 )𝑇 = 𝐵 𝑇 𝐴𝑇

Proof.
(1) Let 𝐴, 𝐵, 𝐶 have the size 𝑚 × 𝑟, 𝑟 × 𝑠, 𝑠 × 𝑛,
respectively.

- 15 -
 Therefore
𝑟 𝑠 𝑟

(𝐴𝐵)𝐶 = [∑ 𝑎𝑖𝑘 𝑏𝑘𝑗 ] 𝐶 = [∑ (∑ 𝑎𝑖𝑘 𝑏𝑘𝑡 ) 𝐶𝑡𝑗 ]

𝑘=1 𝑡=1 𝑡=1
𝑠 𝑟

= [∑ ∑ 𝑎𝑖𝑘 𝑏𝑘𝑡 𝑐𝑡𝑗 ]

𝑡=1 𝑘=1
𝑟 𝑠

= [∑ 𝑎𝑖𝑘 (∑ 𝑏𝑘𝑡 𝑐𝑡𝑗 )]

𝑘=1 𝑡=1
𝑠

= 𝐴 [∑ 𝑏𝑘𝑡 𝑐𝑡𝑗 ] = 𝐴(𝐵𝐶 ).

𝑡=1

(2) Let 𝐴, 𝐵, 𝐶 have the size 𝑚 × 𝑛, 𝑛 × 𝑝, 𝑛 × 𝑝 ,

respectively. Then

𝐴(𝐵 + 𝐶 ) = 𝐴[𝑏𝑖𝑗 + 𝑐𝑖𝑗 ]

𝑛

= [∑ 𝑎𝑖𝑘 (𝑏𝑘𝑗 + 𝑐𝑘𝑗 )]

𝑘=1
𝑛 𝑛

= [∑ 𝑎𝑖𝑘 𝑏𝑘𝑗 + ∑ 𝑎𝑖𝑘 𝑐𝑘𝑗 ]

𝑘=1 𝑘=1
𝑛 𝑛

= [∑ 𝑎𝑖𝑘 𝑏𝑘𝑗 ] + [∑ 𝑎𝑖𝑘 𝑐𝑘𝑗 ]

𝑘=1 𝑘=1

= 𝐴𝐵 + 𝐴𝐶.
- 16 -
(3) Similar to (2)
(4) Let 𝐴 be of size 𝑚 × 𝑛 and 𝐵 of size 𝑛 × 𝑝. Therefore
𝑛 𝑛

𝑘(𝐴𝐵) = 𝑘 [∑ 𝑎𝑖𝑟 𝑏𝑟𝑗 ] = [∑(𝑘𝑎𝑖𝑟 )𝑏𝑟𝑗 ] = (𝑘𝐴)𝐵

𝑟=1 𝑟=1

Similarly, 𝑘(𝐴𝐵 ) = 𝐴(𝑘𝐵).

(5) Let A be of size 𝑚 × 𝑛 and 𝐵 of size 𝑛 × 𝑝. Then
𝐴𝑇 = [𝑎𝑗𝑖 ] and 𝐵 𝑇 = [𝑏𝑗𝑖 ]. Therefore
𝐵𝑇 𝐴𝑇 = [∑𝑛𝑘=1 𝑏𝑘𝑗 𝑎𝑗𝑘 ] = [∑𝑛𝑘=1 𝑎𝑗𝑘 𝑏𝑘𝑗 ] = (𝐴𝐵)𝑇 .◄
♣An 𝑛 × 𝑛 matrix 𝐴 = [𝑎𝑖𝑗 ] is called diagonal if 𝑎𝑖𝑗 = 0
whenever 𝑖 ≠ 𝑗.
The diagonal matrix is called identity matrix if 𝑎𝑖𝑗 = 1
for 𝑖 = 1,2, … . , 𝑛. We denote 𝐼𝑛 for the identity matrix of
size 𝑛 × 𝑛 or 𝐼 if the size is known. For every 𝑛 × 𝑛
matrix, we have 𝐴𝐼𝑛 = 𝐼𝑛 𝐴 = 𝐴.
Definition 9.
Let 𝐴 = [𝑎𝑖𝑗 ] be an 𝑛 × 𝑛 matrix. The trace of 𝐴 is
denoted by 𝑡𝑟(𝐴) and is defined as follows:
𝑛

𝑡𝑟 (𝐴) = ∑ 𝑎𝑖𝑖
𝑖=1

- 17 -
Example 8.
1 3 −2
Let 𝐴 = [0 5 7 ].
1 2 −3
Then 𝑡𝑟(𝐴) = 1 + 5 + (−3) = 3.◄
Theorem 4.
Let 𝐴 and 𝐵 be an 𝑛 × 𝑛 matrices and 𝑘 ∈ ℝ. Then the
following statements are true.
(1) 𝑡𝑟(𝐴 + 𝐵 ) = 𝑡𝑟(𝐴) + 𝑡𝑟(𝐵);
(2) 𝑡𝑟(𝐴𝑇 ) = 𝑡𝑟(𝐴);
(3) 𝑡𝑟(𝐴𝐵 ) = 𝑡𝑟(𝐵𝐴);
(4) 𝑡𝑟(𝑘𝐴) = 𝑘 𝑡𝑟(𝐴);
2
(5) 𝑡𝑟(𝐴𝐴𝑇 ) = ∑𝑖≠𝑗 𝑎𝑖𝑗 .
Proof. Left as exercise.

- 18 -
 Exercise Set (1.1)
3 0
4 −1 1 4 2
Let 𝐴 = [−1 2] , 𝐵 = [ ],𝐶 =[ ]
0 2 3 1 5
1 1
1 5 2 6 1 3
𝐷 = [−1 0 1]and 𝐸 = [−1 1 2]
3 2 4 4 1 3
In Exercise (1) to (30) Find:
(1) 𝐷 + 𝐸 (2) 𝐷 − 𝐸
(3) 5𝐴 (4) −7𝐶
(5) 2𝐵 − 𝐶 (6) 4𝐸 − 2𝐷
(7) −3(𝐷 + 2𝐸) (8) 𝐶 − 𝐶 𝑇
(9) 𝑡𝑟(𝐷) (10) 𝑡𝑟(𝐷 − 3𝐸 )
(11) 4𝑡𝑟(7𝐵) (12) 𝑡𝑟(𝐴)
(13) 2𝐴𝑇 + 𝐶 (14) 𝐷𝑇 − 𝐸 𝑇
(15) (𝐷 − 𝐸 )𝑇 (16) 𝐵𝑇 + 5𝐶 𝑇
1 1
(17) 𝐶𝑇 𝐴 (18) 𝐶 𝑇 − 𝐶 𝑇
4 2

(19) 3𝐸 𝑇 − 2𝐷𝑇 (20) 𝐴𝐵

(21) 𝐵𝐴 (22) 𝐴(𝐵𝐶 )
(23) (𝐶 𝑇 𝐵)𝐴𝑇 (24) 𝑡𝑟(𝐶 𝑇 𝐴𝑇 + 2𝐸 𝑇 )
(25) 4𝐵𝐶 + 2𝐵 (26) 𝐵𝑇 (𝐶𝐶 𝑇 − 𝐴𝑇 𝐴)
(27) 𝐷𝑇 𝐸 𝑇 − (𝐸𝐷 )𝑇 (28) 3𝐴 − 𝐵𝑇 𝐶 𝑇

- 19 -
(29) 𝑡𝑟(𝐷𝐸 + 𝐸𝐷) (30) 𝑡𝑟(𝐷𝐸 − 𝐸𝐷)
(31) Find 𝐴 of 𝐵 when 2𝐴 − 3𝐵 = 6(𝐴 + 3𝐵)
(32) Let 𝐵 be an 𝑚 × 𝑛 matrix such that 𝐵 + 𝐴 = 𝐴 for
every 𝑚 × 𝑛 matrix 𝐴. Prove that 𝐵 = 𝟎.
(33) Find all 2 × 2 matrices such that 𝐴2 = 𝐼.
(34) Prove that there is no an 2 × 2 matrices 𝐴 and 𝐵
such that 𝐴𝐵 − 𝐵𝐴 = 𝐼.

- 20 -
2-ELEMENTARY ROW OPERATIONS
The following operations, called elementary row
operations, on the rows of a matrix:
1- Interchange two rows;
2- Multiply a row by a non-zero constant;
3- Replace one row by the sum of itself and a constant
multiple of another row.

Definition 1. Let A and B be two matrices. We say that A

and B are row equivalent, written 𝐴~𝐵 if we can obtain
one of them from the another by a finite number of
elementary row operations.
Notes.
●It is obvious that if 𝐴~𝐵, then the size of A is same as
B
●For ease of exposition we shall denote the elementary
row operations by a shorthand notation.
● 𝑅𝑖𝑗 will indicate that the 𝑖𝑡ℎ and 𝑗𝑡ℎ rows have been
interchanged.
● 𝑘𝑅𝑖 will indicate that the 𝑖𝑡ℎ row is multiplied by a
nonzero constant 𝑘.

- 21 -
●𝑘𝑅𝑖𝑗 will indicate that 𝑘 times the 𝑖𝑡ℎ row is added to
the 𝑗𝑡ℎ row and the sum becomes the new 𝑗𝑡ℎ row.
●If we obtained B from A by elementary row operation,
then we can obtain B from A by the inverse of the
elementary row operation as follows:
1
(a') 𝑅𝑗𝑖 ; (b') 𝑅𝑖 ; (c') −𝑘𝑅𝑖𝑗 .
𝑘

Definition2. We say that a matrix is in a row-echelon

form if it has the following properties: -
(1) The first entry of the first column (leading entry) is
nonzero (1, if it is possible) and all other entries of this
column are zeros.
(2) The first column that has a nonzero entry that is
not in the first has a nonzero entry in the second row,
and every entry in this column below the second row
equals zero.
(3) For 𝑘 ≥ 3, continue in the manner described (2).
That is, the first column has a nonzero 𝑘 𝑡ℎ entry has
every entry below the 𝑘 𝑡ℎ entry to zero. Moreover, this

- 22 -
 column occurs to the right of the first column that
contains a nonzero (𝑘 − 1)𝑡ℎ entry.

13 5 7
1 4 5
01 1 0
In particular, [0 1 4] and [ ] are in row-
00 1 3
0 0 0
00 0 1
2 3 5
echelon form, while [4 0 1] is not.
0 0 0
♣ A matrix in row-echelon form that satisfy the following
two properties is said to be in reduced row-echelon
form:
● If a row does not consist entirely of zeros then the
leading entry of that row, is 1.
● Any column containing the leading entry of some row
consists entirely of zeros except for that single leading
entry.
For example,
1 0 2
1 0 0 1 4 7 0
0 1 0
[0 1 0], [ ] and [0 0 0 1]
0 0 0
0 0 1 0 0 0 0
0 0 0
are in reduced row-echelon form,
while,
- 23 -
 1 0 1
1 1 0 1 2 6 6
0 0 0
[0 1 3], [ ] and [0 0 2 1]are not.
0 1 0
0 0 1 0 0 0 0
0 0 0
♣ A systematic process that transforms any matrix to a
matrix in row-echelon form is as follows:
(1) Interchanging rows (if necessary) to replace a
nonzero number in the first entry of the first row.
(2) Use the second elementary row operation as many
times necessary to replace a zero as the first entry in
every row below the first.
(3) Hold the first row fixed, and interchange other
rows (if necessary) to give the second row a nonzero
entry as for as left as possible.
(4) Hold the first two rows fixed; use the second
elementary row operation to replace a zero in every row
after the second row directly beneath the first nonzero
entry in the second row.
(5) Continue to hold the first two rows fixed and
interchange other rows (if necessary) to give the third
row a nonzero entry as for left as possible.

- 24 -
 (6) Hold the first three rows fixed; use the second
elementary row operation to replace a zero in every row
after the third row directly beneath the first nonzero
entry in the third row.
(7) Continue in this fashion row-echelon forms
obtained.

22 2 4
Example 1. Transform the matrix [7 3 4 4] to a matrix
61 2 1
in row- echelon form.
Solution:
2 2 2 4 1 𝑅1 1 1 1 2 −7𝑅12 ,−6𝑅13 1 1 1 2
2
[7 3 4 4 ] → [ 7 3 4 4] → [0 −4 −3 −10]
61 2 1 61 2 1 0 −5 −4 −11
1
− 𝑅2 1 1 1 2 5𝑅23 1 1 1 2
4
→ [0 1 3/4 5/2 ] → [0 1 3/4 5/2]
0−5 −4 −11 0 0 −1/4 3/2

−4𝑅3 1 1 1 2
→ [0 1 3/4 5/2]
0 0 1 −6

Therefore row-echelon form of the matrix is: -

- 25 -
 1 1 1 2
[0 1 3/4 5/2]
0 0 1 −6

Example 3.
Put the matrix
5 2 2
𝐴 = [7 3 4]
6 1 2
in reduced row echolen-form.
Solution:
1
𝑅 1 2/5 2/5 −7𝑅12,−6𝑅13 1 2/5 2/5
5 1
𝐴→ [7 3 4 ]→ [0 1/5 6/5 ]
6 1 2 0 −7/5 −2/5

7
5𝑅21 2/5 2/5 𝑅23 1 2/5 2/5
5
→ [0 1 6 ]→ [0 1 6 ]
0 −7/5 −2/5 0 0 8
1
𝑅 1 2/5 2/5 −2 𝑅21 1 0 −2
8 3 5
→ [0 1 6 ]→ [0 1 6 ]
0 0 1 0 0 1

2𝑅31, −6𝑅32 1 0 0
→ [0 1 0]
0 0 1

- 26 -
 Exercise Set (1.2)
Transform the following matrices to the reduced row-
echelon form: -

1 2 1 1 1 3 1 7
1 −1 4 2 (2) [3 −2 4 1]
(1) [ ]
2 1 3 1 4 1 9 6
−3 1 1 4
1 2 1 0 2 4
1 3 5 (4) [1 −3 5 ]
(3) [ ]
2 5 6 3 −1 −1
3 7 7
1 3 −2 0 2 0 1 1
2 6 −5 −2 4 −3 (6) [2 −1]
(5) [ ]
0 0 5 10 0 15 1 2
2 6 0 8 4 18
1 3 −2 −1 0
6 3
4 −5 (8) [ 2 1 1
(7) [6 ] 0 0 1 1 0
]
2 −3 2
2 6 4 2 −2
1 3 1 1
(9) [−1 −2 1 1]
3 7 −1 1

- 27 -
3. Inverse of Matrix
Definition 1.
Let A be an 𝑛 × 𝑛 matrix. The matrix B is an inverse
matrix of A if B is an 𝑛 × 𝑛 matrix and 𝐴𝐵 = 𝐵𝐴 = 𝐼. In
this case, we say that A is invertible or nonsingular.
Example 1.
7 3
Prove that 𝐵 = [ ] is an inverse of
2 1
1 −3
𝐴=[ ].
−2 7
Solution. Since
1 −3 7 3 1 0
𝐴𝐵 = [ ][ ]=[ ]=𝐼
−2 7 2 1 0 1
And
7 3 1 −3 1 0
𝐵𝐴 = [ ][ ]=[ ] = 𝐼.
2 1 −2 7 0 1
Then B is the inverse of A.◄
0 1
Example 1. Prove that the matrix 𝐴 = [ ] has no
0 1
inverse (singular).
𝑎 𝑏
Solution. Let 𝐵 = [ ] be the inverse of A. Then
𝑐 𝑑
0 1 𝑎 𝑏 1 0
𝐴𝐵 = [ ][ ]=[ ].
0 1 𝑐 𝑑 0 1

- 28 -
Therefore 𝑐 = 0 and 𝑐 = 1, impossible. Then A has no
inverse. ◄
Theorem 1.
The inverse of a square matrix A is unique.
Proof.
Let B and C are inverses of A, then 𝐴𝐶 = 𝐶𝐴 = 𝐼 and
𝐴𝐵 = 𝐵𝐴 = 𝐼. Therefore
𝐶 = 𝐶𝐼 = 𝐶 (𝐴𝐵 ) = (𝐶𝐴)𝐵 = 𝐼𝐵 = 𝐵.◄
Note.
We will write 𝐴−1 for the inverse of 𝐴
Definition 2.
Let A be an 𝑛 × 𝑛 matrix and k be positive integer. We
define 𝐴𝑘 inductively as follows: 𝐴𝑘+1 = 𝐴𝑘 𝐴 and 𝐴0 =
𝐼 , 𝑘 ≥ 0. Also, 𝐴−𝑘 = (𝐴−1 )𝑘 .
Theorem 2.
(1) 𝐼 −1 = 𝐼;
(2) (𝐴−1 )−1 = 𝐴
(3) If both A and B have an inverse, then AB has an
inverse and (𝐴𝐵 )−1 = 𝐵−1 𝐴−1 ;
(4) If 𝐴1 , 𝐴2 , … . , 𝐴𝑘 have an inverse, then 𝐴1 𝐴2 … 𝐴𝑘
−1 −1
has an inverse (𝐴1 𝐴2 … 𝐴𝑘 )−1 = 𝐴−1
𝑛 … 𝐴2 𝐴1 .

- 29 -
 (5) If A has an inverse, then 𝐴𝑘 has an inverse for all
𝑘 ≥ 1 and (𝐴𝑘 )−1 = (𝐴−1 )𝑘
(6) If A has an inverse and 0 ≠ 𝑟 ∈ ℝ, then 𝑟𝐴 has an
1
inverse and (𝑟𝐴)−1 = 𝐴−1
𝑟

(7) If A has an inverse, then 𝐴𝑇 has an inverse and

(𝐴𝑇 )−1 = (𝐴−1 )𝑇

Proof.
(1) Since 𝐼𝐼 = 𝐼, then 𝐼 −1 = 𝐼.
(2) Since 𝐴−1 exists and 𝐴𝐴−1 = 𝐼 = 𝐴−1 𝐴, by
uniqueness of the inverse of 𝐴, 𝐴 is the inverse of 𝐴−1 .
Therefore (𝐴−1 )−1 = 𝐴.
(3) Since(𝐴𝐵)(𝐵 −1 𝐴−1 ) = 𝐴(𝐵𝐵−1 )𝐴−1
= 𝐴𝐼𝐴−1 = 𝐴𝐴−1 = 𝐼
and
(𝐵−1 𝐴−1 )(𝐴𝐵) = 𝐵−1 (𝐴−1 𝐴)𝐵 = 𝐵−1 𝐼𝐵
= 𝐵𝐵−1 = 𝐼.
Then (𝐴𝐵 )−1 = 𝐵−1 𝐴−1 .
(4) We use induction.

The statement is true for 𝑘 = 1. Also, it is true for 𝑘 = 2.

Now, suppose that the statement is true at 𝑘 = 𝑚, 𝑚 > 2,
- 30 -
𝑖. 𝑒.,
(𝐴1 𝐴2 … 𝐴𝑚 )−1 = 𝐴−1 −1 −1
𝑚 … 𝐴2 𝐴1 .

We prove the statement is true for 𝑘 = 𝑚 + 1.

(𝐴1 𝐴2 … 𝐴𝑚 𝐴𝑚+1 )−1 = 𝐴−1
𝑚+1 (𝐴1 𝐴2 … 𝐴𝑚 )
−1

= 𝐴−1 −1 −1 −1
𝑚+1 𝐴𝑚 … 𝐴2 𝐴1 .

So, the statement is true for all 𝑘.

(5) We put 𝐴1 = 𝐴2 = ⋯ = 𝐴𝑘 = 𝐴 in (4).
(6) Note that
1 1
(𝑟𝐴) ( 𝐴−1 ) = 𝑟 ( ) 𝐴𝐴−1 = 1𝐼 = 𝐼,
𝑟 𝑟
1 1
( 𝐴−1 ) (𝑟𝐴) = ( ) 𝑟𝐴−1 𝐴 = 1𝐼 = 𝐼.
𝑟 𝑟
1
Then (𝑟𝐴)−1 = 𝐴−1
𝑟

(7) 𝐴𝑇 (𝐴−1 )𝑇 = (𝐴−1 𝐴)𝑇 = 𝐼𝑇 = 𝐼,

(𝐴−1 )𝑇 𝐴𝑇 = (𝐴𝐴−1 )𝑇 = 𝐼𝑇 = 𝐼.
Therefore
(𝐴𝑇 )−1 = (𝐴−1 )𝑇 .
Definition 3.

Any matrix can be obtained from 𝐼𝑛 by using only one

elementary row operation is called an elementary matrix.

- 31 -
Example 3.
1 0
The matrix 𝐸 = [ ] is an elementary matrix, because
0 5
we obtained it from 𝐼2 by only one elementary row
operation (5𝑅2 ). Also,
1 0 5
𝐹 = [0 1 0]
0 0 1
is an elementary matrix, because we obtained it from 𝐼3 by
only one elementary row operation (5𝑅31 ).
1 0 2 3
Now let 𝐴 = [2 −1 3 6]. If 𝐵 is the matrix obtained
1 4 4 0
from 𝐴 by the elementary row operation (5𝑅13 ), 𝑖. 𝑒.,
1 0 2 3
𝐵 = [2 −1 3 6 ], then 𝐸𝐴 = 𝐵, where𝐸 is the
6 4 14 15
elementary matrix obtained from 𝐼3 by(5𝑅13 ).
Thus
100 1 0 23 1 0 2 3
𝐸𝐴 = [0 1 0] [2 −1 3 6] [2 −1 3 6 ] = 𝐵.◄
5 0 1 1 4 4 0 6 4 14 15

- 32 -
Theorem 3.
Let 𝐴 be an 𝑛 × 𝑛 matrix and 𝐸 be elementary matrix
obtained from 𝐼𝑚 . Then 𝐸𝐴 is a matrix obtained from 𝐴
by the same elementary row operation which done in 𝐼𝑚
to obtained 𝐸.
Proof.
We prove the theorem for one elementary row operation.
For example, multiplying𝑐𝑅𝑘 . That is, suppose that 𝐴 =
𝑎11 𝑎12 … 𝑎1𝑛
𝑎21 𝑎22 … 𝑎2𝑛
[ ⋮ ⋮ ⋱ ⋮ ]and 𝐸 is the elementary matrix
𝑎𝑚1 𝑎𝑚2 … 𝑎𝑚𝑛
obtained from 𝐼𝑚 . By 𝑐𝑅𝑘 , 𝑖. 𝑒.,
10 … 0 𝑎11 𝑎12 … 𝑎1𝑛
01 … 0 𝑎21 𝑎22 … 𝑎2𝑛
⋮⋮ : ⋮ ⋮ ⋮ ⋮ ⋮
𝐸= . Then𝐸𝐴 = 𝑐𝑎𝑘𝑛 and
00… 𝑐 …0 𝑐𝑎𝑘1 𝑐𝑎𝑘2 …
⋮⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮
[ 00 … 1] [ 𝑎𝑚1 𝑎𝑚2 … 𝑎𝑚𝑛 ]
𝑎11 𝑎12 … 𝑎1𝑛
𝑎21 𝑎22 … 𝑎2𝑛
𝑐𝑅𝐾 ⋮ ⋮ ⋮ ⋮
𝐴 → 𝑐𝑎 𝑐𝑎 …𝑐𝑎 = 𝐸𝐴 .
𝑘1 𝑘2 𝑘𝑛
⋮ ⋮ ⋮ ⋮
[ 𝑎𝑚1 𝑎𝑚2 … 𝑎𝑚𝑛 ]
Similarly, we can prove the other two statements.
- 33 -
Theorem 4.
Elementary matrices are nonsingular, and the inverse of
an elementary matrix is another elementary matrix.
Proof.
Let 𝐸 be the elementary matrix obtained from 𝐼𝑛 by
multiplying its 𝑖 𝑡ℎ row by a nonzero constant c.
Let 𝐹 be the elementary matrix obtained from 𝐼𝑛 by
1
multiplying its 𝑖 𝑡ℎ row by . Then 𝐸𝐹 is the matrix
𝑐
1
obtained from 𝐹 by multiplying its 𝑖𝑡ℎ row by .
𝑐

Therefore 𝐸𝐹 = 𝐼𝑛 . Then the matrix 𝐸 is nonsingular and

𝐹 = 𝐸 −1 .
The proof of the following theorem is left as an exercise.
Theorem 5.
Let 𝐴 be an 𝑛 × 𝑛 matrix the following are equivalent:
(1) 𝐴 has an inverse.
(2) The reduced row-echelon form of 𝐴 is 𝐼𝑛 .
(3) 𝐴 can be written as the product of finite number of
elementary matrices.

♣ There is a relatively easy way to determine whether an

𝑛 × 𝑛 matrix has an inverse and to compute 𝐴−1 whether
- 34 -
it exists. To begin we construct an 𝑛 × 𝑛, called a
partitioned matrix and denoted by [𝐴|𝐼𝑛 ], by 𝐴 and 𝐼𝑛
next to each other:
𝑎11 𝑎12 … 𝑎1𝑛 1 … 0
𝑎 𝑎22 … 𝑎2𝑛 0 … 0
[𝐴|𝐼𝑛 ] = [ 21
: : … : |: : ]
∶
𝑎𝑛1 𝑎𝑛2 … 𝑎𝑛𝑛 0 1
We now attempt to change the left side of this matrix into
𝐼𝑛 using elementary row operation.
If this can be done, the right side of resulting matrix is
𝐴−1 . If the left side of [𝐴|𝐼𝑛 ] cannot changed into 𝐼𝑛 ,
then 𝐴 has no inverse.
Example 4.
1 1 1
−1
Compute 𝐴 , where 𝐴 = [−1 3 2].
2 1 1
Solution:
𝟏 1 1 1 0 0
[𝐴|𝐼𝑛 ] = [−1 3 2 | 0 1 0]
2 1 1 0 0 1
1𝑅12 , −2𝑅13 𝟏 1 1 1 0 0
→ [0 4 3 | 1 1 0]
0 −1 −1 −2 0 1

- 35 -
 𝑅231 1 1 1 0 0
→ [0 −1 −1 |−2 0 1]
0 4 3 1 1 0
−1𝑅2 1 1 1 1 0 0
→ [0 𝟏 1 |2 0 −1]
0 4 3 1 1 0

−4𝑅23 1 1 1 1 0 0
→ [0 𝟏 1 | 2 0 −1]
0 0 −1 −7 1 4
−1𝑅3 1 1 1 1 0 0
→ [0 1 1 |2 0 −1]
0 0 𝟏 7 −1 −4
−1𝑅23 ,−1𝑅31 1 1 0 −6 1 4
→ [0 𝟏 0 |−5 1 3]
0 0 𝟏 7 −1 −4

−1𝑅21 1 0 0 −1 0 1
→ [0 1 0 |−5 1 3 ].
0 0 1 7 −1 −4
Therefore
−1 0 1
−1
𝐴 = [−5 1 3]
7 −1 −4

Example 5.
Find the inverse of the matrix (if it exists)
- 36 -
 1 2 3
𝐴 = [4 5 6]
7 8 9
Solution:
1 2 3 1 0 0
[4 5 6 |0 1 0]
7 8 9 0 0 1
−4𝑅12 ,−7𝑅13 1 2 3 1 0 0
→ [0 −3 −6 |−4 1 0]
0 −6 −12 −7 0 1
1
− 𝑅23 1 2 3 1 0 0
3
→ [0 1 2 |4/3 −1/3 0]
0 −6 −12 −7 −2 1

6𝑅23 1 2 3 1 0 0
→ [0 1 2 |4/3 −1/3 0]
0 0 0 1 −2 1
Since the last row in the left matrix is zero, the matrix 𝐴
has no inverse.

- 37 -
 Exercise Set (1.3)
𝑎
𝑏
(1) Let 𝐴 = [ ] and 𝑎𝑑 − 𝑏𝑐 ≠ 0. Prove that 𝐴 has
𝑐
𝑑
1 𝑎 −𝑏
an inverse and 𝐴−1 = [ ]
𝑎𝑑−𝑏𝑐 −𝑐 𝑎
(2) Use Exercise (1) to find the inverse of the following
matrix (if it exists).
1 2 −1 3 1/2 0 3 6
(a)[ ] (b) [ ](c)[ ] (d) [ ].
1 3 2 5 0 1/3 2 4
(3) Compute the inverse of the following matrices (if it
exists).
1 7 12 1 2 1
(a)[5 −1 2 ] (b) [3 5 0]
2 3 5 1 7 1
1 1 2 1 4 2
(c) [1 1 1] (d) [2 3 3]
1 2 4 4 1 4
−1 4 5 2 1 2 3 0 0
0 0 0 −1 1 1 1 0 0
(e) [ ]
1 −2−2 0 (f) 0 1 5 1 0
0 −1−1 0 0 0 1 2 2
[0 0 0 0 1]

- 38 -
(4) Write the following matrices as a product of
elementary matrices (if it is possible).
102
2 1 5 0
(a)[ ], (b) [ ], (c)[011],
3 1 3 2
216

1 0 −3
(d) [ 0 1 4 ]
−2 2 15
(5) Prove that 𝐸 𝑡 is an elementary matrix.
(6) Let 𝐴 be an invertible matrix and 𝐴𝐵 = 𝐴𝐶. prove
that 𝐵 = 𝐶.
(7) Let 𝐴 be an invertible matrix and 𝐴2 − 3𝐴 + 𝐼 = 0.
Prove that 𝐴−1 = (3𝐼 − 𝐴)
(8) Let 𝐴 and 𝐵 be matrices of size.prove that:-
(a)𝐴(𝐼 + 𝐵𝐴) = (𝐼 + 𝐴𝐵)𝐴
(b) (𝐼 + 𝐵𝐴)𝐵 = 𝐵(𝐼 + 𝐴𝐵).

- 39 -
4. SPECIAL MATRICES
Definition 1.
An 𝑛 × 𝑛 matrix 𝐴 = [𝑎𝑖𝑗 ] is called upper triangular if
𝑎𝑖𝑗 = 0 whenever 𝑖 > 𝑗 and lower triangular if 𝑎𝑖𝑗 = 0
whenever 𝑖 < 𝑗. A matrix is called triangular if it is
either upper or lower triangular.

Definition 2.
Let 𝐴 be a square matrix:
(1) 𝐴 is symmetric if 𝐴𝑇 = 𝐴.
(2) 𝐴is anti(skew)-symmetric if 𝐴𝑇 = −𝐴.
Note.
We considered only the matrices of real numbers.
Sometimes we need to study the matrices of complex
number, i. e., its elements is complex numbers.

- 40 -
Example 1.
1 2 −1 𝑖 𝑖 1+𝑖
Both 𝐴 = [ 2 3 7 ] and 𝐵 = [ 𝑖 2𝑖 2 − 𝑖]
−1 7 5 1+𝑖 2−𝑖 1
1 −1
are symmetric matrices. But 𝐶 = [ ] is not
3 2
symmetric.
Example 2. Both the matrices
0 2 −1 0 𝑖 1+𝑖
𝐴 = [−2 0 7 ] and 𝐵 = [ −𝑖 0 2 − 𝑖]
1 −7 0 −1 − 𝑖 −2 + 𝑖 0
are anti-symmetric.
Theorem 1. Let 𝐴 be a square matrix and 𝑘 ∈ ℝ. Then
(1) 𝐴𝐴𝑇 is symmetric;
(2) 𝐴 + 𝐴𝑇 is symmetric;
(3) 𝐴 − 𝐴𝑇 is anti-symmetric;
(4) If 𝐴 is symmetric, then 𝑘𝐴 is symmetric.
(5) If 𝐴 is anti-symmetric, then 𝑘𝐴 is anti-symmetric.
(6) There exist a unique matrices 𝐵 and 𝐶, where 𝐵 is
symmetric and 𝐶 is anti-symmetric such that
𝐴 = 𝐵 + 𝐶.

- 41 -
Proof.
(1) Since(𝐴𝐴𝑇 )𝑇 = (𝐴𝑇 )𝑇 𝐴𝑇 = 𝐴𝐴𝑇 , 𝐴𝐴𝑇 is
symmetric.
(2) Since (𝐴+𝐴𝑇 )𝑇 = 𝐴𝑇 + (𝐴𝑇 )𝑇 = 𝐴𝑇 + 𝐴 =
𝐴 + 𝐴𝑇 , 𝐴 + 𝐴𝑇 is symmetric.
(3) Since (𝐴−𝐴𝑇 )𝑇 = 𝐴𝑇 + (−𝐴𝑇 )𝑇 = 𝐴𝑇 − (𝐴𝑇 )𝑇 =
𝐴𝑇 − 𝐴 = −(𝐴−𝐴𝑇 ), 𝐴−𝐴𝑇 is anti-symmetric.
(4) Since A is symmetric, 𝐴𝑇 = 𝐴.
Therefore (𝑘𝐴)𝑇 = 𝑘𝐴𝑇 = 𝑘𝐴. Hence 𝑘𝐴 is
symmetric.
(5) Since 𝐴 is anti-symmetric, then 𝐴𝑇 = −𝐴
Therefore (𝑘𝐴)𝑇 = 𝑘𝐴𝑇 = 𝑘(−𝐴) = −𝑘𝐴. Hence
𝑘𝐴 is anti-symmetric.
1
(6) Note that 𝐵 = (𝐴 + 𝐴𝑇 ) is symmetric and 𝐶 =
2
1
(𝐴 − 𝐴𝑇 ) is anti-symmetric.
2
1 1
Then 𝐵 + 𝐶 = (𝐴 + 𝐴𝑇 ) + (𝐴 − 𝐴𝑇 ) = 𝐴.
2 2

To prove the uniqueness, suppose 𝐷 is symmetric

and 𝐸 is anti-symmetric such that 𝐴 = 𝐷 + 𝐸.
Now, 𝐴𝑇 = (𝐷 + 𝐸 )𝑇 = 𝐷𝑇 + 𝐸 𝑇 = 𝐷 − 𝐸.
Furthermore 2𝐷 = 𝐴 + 𝐴𝑇 and 2𝐸 = 𝐴 − 𝐴𝑇 .
- 42 -
 1 1
Then 𝐷 = (𝐴 + 𝐴𝑇 ) = 𝐵 and 𝐸 = (𝐴 − 𝐴𝑇 ) =
2 2

𝐶. Hence 𝐵 and 𝐶 are unique. ◄

Example 3.
1 5 1
If 𝐴 = [ 3 3 −1], then the symmetric matrix 𝐵 and
−1 2 4
the anti-symmetric matrix C, where 𝐴 = 𝐵 + 𝐶 are.
1 4 0
1
𝐵 = (𝐴 + 𝐴𝑇 ) = [4 3
1
2] and
2 1
0 4
2
0 1 1
1 𝑇)
𝐶 = (𝐴 − 𝐴 = [−1 0 −3/2].◄
2
−1 3/2 0

Definition 3.
Let 𝐴 = [𝑎𝑖𝑗 ] be a complex matrix. The 𝐴 = [𝑎𝑖𝑗 ] is
called the conjugate matrix for 𝐴, where 𝑎𝑖𝑗 is conjugate
of the complex number 𝑎𝑖𝑗 .
There following theorem gives the properties of the
conjugate matrix. It easily be proved by using the
properties of the complex numbers. The proof is left to
the reader as exercise.
- 43 -
Theorem 2.
Let 𝐴 = [𝑎𝑖𝑗 ] and 𝐵 = [𝑏𝑖𝑗 ] be a complex matrices.

(1) ̿𝐴 = 𝐴;
(2) (𝑘𝐴) = 𝑘 𝐴, 𝑘 ∈ ℂ;

(3) 𝐴 + 𝐵 = 𝐴 + 𝐵, where 𝐴, 𝐵 have the same size;

(4) 𝐴𝐵 = 𝐴̅ 𝐵̅ , where 𝐴 is of size 𝑚 × 𝑛 and 𝐵 is of
size 𝑛 × 𝑝.

Definition 4.
Let 𝐴 be a complex matrix. The conjugate transpose of 𝐴
is denoted by 𝐴∗ and is defined as follows:
𝑇
𝐴∗ = (𝐴) = 𝐴𝑇 .
Example 4.
0 1+𝑖
𝑖 0 1−𝑖 −𝑖
If 𝐴 = [ ], then 𝐴 = [ ]
2 3 − 2𝑖
−𝑖 2 3 + 2𝑖 𝑖
𝑇
0 2
∗
And 𝐴 = (𝐴) = [1 − 𝑖 3 + 2𝑖 ].
−𝑖 𝑖
Theorem 3.
(1) (𝐴∗ )∗ = 𝐴;
(2) (𝐴 + 𝐵)∗ = 𝐴∗ +𝐵 ∗ ;
(3) (𝑘𝐴)∗ = 𝑘𝐴∗ , 𝑘 ∈ ℂ ;
- 44 -
 (4) (𝐴𝐵)∗ = 𝐵 ∗ 𝐴∗ .

Proof.
𝑇
𝑇 ∗ 𝑇 𝑇 𝑇
(1) (𝐴∗ )∗
= [(𝐴) ] = ((𝐴) ) = [(𝐴̿) ] = 𝐴;
𝑇 𝑇
(2) (𝐴 + 𝐵)∗ = (𝐴 + 𝐵) = (𝐴 + 𝐵) .
𝑡 𝑡
= (𝐴) + (𝐵) = 𝐴∗ + 𝐵 ∗ ;
𝑇 𝑇
(3) (𝑘𝐴)∗ = (𝑘 𝐴) = 𝑘(𝐴) = 𝑘𝐴∗
𝑇 𝑇 𝑇
(4) (𝐴𝐵)∗ = (𝐴𝐵) = (𝐴̅𝐵̅)𝑇 = (𝐵) (𝐴) = 𝐵∗ 𝐴∗ .

Definition 5.
Let 𝐴 be a nonsingular complex matrix. 𝐴 is said to be
unitary if 𝐴−1 = 𝐴∗ .
Example 5.
3 4
𝑖
5 5
Prove that the matrix 𝐴 = [−4 3 ] is unitary.
𝑖
5 5

Solution:

3/5 −4/5
𝐴∗ = 𝐴𝑇 = [ ].
−4/5𝑖 −3/5𝑖

3/5 4/5𝑖 3/5 −4/5 1 0

Also, 𝐴𝐴∗ = [ ][ ]=[ ].
−4/5 3/5𝑖 −4/5𝑖 −3/5𝑖 0 1

- 45 -
Hence 𝐴−1 = 𝐴∗ .

Remark:
If 𝐴 is real matrix, then 𝐴 = 𝐴. Therefore 𝐴∗ = 𝐴𝑇 .
Hence 𝐴 is unitary if and only if 𝐴−1 = 𝐴𝑇 . The real
unitary matrix is called orthogonal matrix.
Example 6.
3/7 2/7 6/7
The matrix 𝐴 = [−6/7 3/7 2/7 ] is orthogonal
2/7 6/7 −3/7
matrix because 𝐴𝑇 𝐴 = 𝐼, i.e., 𝐴−1 = 𝐴𝑇 .
Definition 6.
A square complex matrix 𝐴 is said to be Hermitian
matrix if 𝐴∗ = 𝐴 and anti-Hermitian matrix if 𝐴∗ = −𝐴.
Example 7.
−2 1−𝑖 −1 + 𝑖
The matrix 𝐴 = [ 1 + 𝑖 0 3 ] is
−1 − 𝑖 3 5
𝑖 2 1 + 2𝑖
Hermitian matrix, while 𝐵 = [ −2 5𝑖 −2 + 𝑖]
−1 + 2𝑖 2 + 𝑖 0
is anti-Hermitian matrix.

- 46 -
 Exercise Set (1.4)
(1) Let𝐴 and 𝐵 be symmetric matrices of size 𝑛.
Prove that
(a) 𝐴𝐵 + 𝐵𝐴 is symmetric;
(b) 𝐴𝐵 − 𝐵𝐴 is anti-symmetric;
(c) 𝐴𝐵 is symmetric if and only if 𝐴𝐵 = 𝐵𝐴.

(2) Let 𝐴 and 𝐵 be anti-symmetric matrices of size n.

Prove that 𝐴𝐵 is anti-symmetric if and only if
𝐴𝐵 = −𝐵𝐴.
(3) Prove that if 𝐴 is symmetric, 2𝐴2 − 3𝐴 + 𝐼 is
symmetric.
(4) We say that the matrix A is idempotent, if 𝐴2 = 𝐴.
Prove that if 𝐴𝑇 𝐴 = 𝐴, 𝐴 is symmetric and idempotent.
(5) Let 𝐴 and 𝐵 be matrices of size n.
Prove that:
(a) 𝐴𝑇 𝐵 + 𝐵𝑇 𝐴 is symmetric.
(b) 𝐴𝑇 𝐵 − 𝐵𝑇 𝐴 is anti-symmetric.

(6) Determine whether the following matrices is unitary

and find the inverse of unitary one.

- 47 -
 𝑖𝑒
(a) [ 𝑒 𝑖𝑒
1 𝑒 −𝑖𝑒 ]
√2 𝑖𝑒 −𝑖𝑒 −𝑖𝑒
3 4𝑖
(b) [ ]
−4 3𝑖
1 1 1
(c) [−1 + 𝑖 1 + 𝑖 ]
√2
√2 √2

3+𝑖 1 − 𝑖 √3
[√
1
(d) ]
2√2 1 + 𝑖 √3 1 − √3

(7) Prove that if 𝐴 is unitary, then 𝐴∗ is unitary.

(8) The complex square matrix 𝐴 is normal if 𝐴𝐴∗ =
𝐴∗ 𝐴. Prove that:
1 1+𝑖
(a)[ ] is normal.
1−𝑖 −3
(b) If 𝐴 is Hermition, then 𝐴 is normal.
(c)If 𝐴 is unitary, then 𝐴 is normal.

(9) Find 𝑎, 𝑏 and 𝑐 such that the following matrix 𝐴 is

Hermitian:
−1 𝑎 −𝑖
𝐴 = [3 − 5𝑖 0 𝑐]
𝑏 2 + 4𝑖 2
(10) Let 𝐴 be an invertible matrix. Prove that 𝐴∗ is
invertible and (𝐴∗ )−1 = (𝐴−1 )∗ .

- 48 -
(11) Let 𝐴 be a square matrix. Prove that 𝐴 + 𝐴∗ is
Hermitian.
(12) Which of the following matrices is Hermitian and
which anti- Hermitian:
0
𝑖 𝑖 𝑖 1 1+𝑖
[ ], [ ], [ ],
2 −𝑖 𝑖
𝑖 1 − 𝑖 −3
−2 1 − 𝑖 −1 + 𝑖
[ 1+𝑖 0 3 ]
−1 − 𝑖 3 5

- 49 -
 CHAPTER (II)
SYSTEMS OF LINEAR EQUATIONS

- 50 -
 CHAPTER (II)
SYSTEMS OF LINEAR EQUATIONS
1. Introduction of systems of linear equations
The study of systems of linear equations and their
solutions is one of the major topics in Mathematic (II). In
this section we shall introduce same basic terminology
and discuss a method for solving such systems. We
suppose that the student is familiar with a simple of linear
system and its solution. In general, an arbitrary system of
𝑚 linear equations in 𝑛 unknowns can be written as.
𝑎11 𝑥1 + 𝑎12 𝑥2 + ⋯ + 𝑎1𝑛 𝑥𝑛 = 𝑏1
𝑎21 𝑥1 + 𝑎22 𝑥2 + ⋯ + 𝑎2𝑛 𝑥𝑛 = 𝑏2
⋮⋮⋮⋮
𝑎𝑚1 𝑥1 + 𝑎𝑚2 𝑥2 + ⋯ + 𝑎𝑚𝑛 𝑥𝑛 = 𝑏𝑚
Where 𝑥1 , 𝑥2 , … … , 𝑥𝑛 are the unknowns and the
subscripted a's and b's denote constants.
If we mentally keep track of the location of +'s, the x's,
and the = 's, a system of 𝑚 linear equations in 𝑛
unknowns can be abbreviated by writing only the matrix
equation 𝐴𝑥 = 𝑏, where

- 51 -
 𝑎11 𝑎12 . . 𝑎1𝑛 𝑥1 𝑏1
𝑎21 𝑎22 . . 𝑎2𝑛 𝑥2 𝑏2
𝐴=[ : : .. : ] , 𝑥 = [ : ] 𝑏 = [ ]
:
𝑎𝑚1 𝑎𝑚2 . . 𝑎𝑚𝑛 𝑥𝑛 𝑏𝑚
𝐴 is the coefficients matrix, 𝑏 is the constants matrix 𝑥 is
the unknown's matrix, and [𝐴|𝑏] is the augmented matrix.
Note that the number of the rows in 𝐴 is equal the number
of the equations in the system and the number of the
columns is equal the number of the unknowns.
For example, the augmented matrix for the system of
equations
𝑥1 + 𝑥2 + 2𝑥3 = 9
2𝑥1 + 4𝑥2 − 3𝑥3 = 1
3𝑥1 + 6𝑥2 − 5𝑥3 = 0
is
1 1 2 9
[2 4 −3| 1]
3 6 −5 0
Remark. When constructing an augmented matrix, the
unknown must be written in the same order in each
equation.
The basic method for solving a system of linear equations
is to replace the given system by a new system that has
- 52 -
the same solution set but is easier to solve. This new
system is generally obtained by the elementary row
operations. Now, we introduce two methods using the
augmented matrix to solve a system of linear equations.
(1) Gaussian elimination.
To solve the given system (1) by Gaussian elimination we
put the augmented matrix for the system (1) in row-
echelon form, then the solution set of the system will be
evident by inspection or after a few simple steps.
(2) Gauss-Jordan-elimination:

This method is similar to Gauss-elimination, but we put

the augmented matrix in a reduced row-echelon form.
Note that the reduced row echelon form of a matrix is
unique. No matter how the row operations are varied.
Before introducing examples on Gauss and Gauss-Jordan
we have that following definition.
Definition 1.
A system of linear equations is consistent if it has a
solution. Otherwise is inconsistent.

- 53 -
Example 1.
Solve the following system by Gauss and Gauss-Jordan
elimination.
2𝑥2 + 4𝑥3 = 3
𝑥1 − 3𝑥2 + 5𝑥3 = 1
3𝑥1 − 𝑥2 − 𝑥3 = 1
Solution.
(a) Gaussian-elimination

The augmented matrix for the given system is

0 2 4 3
[𝐴|𝑏] = [1 −3 5 | 1].
3 −1 −1 1
Now
1 −3 5 1
𝑅12
[𝐴|𝑏] → [0 2 4 | 3]
3 −1 −1 1
−3𝑅13 1 −3 5 1
→ [0 2 4 | 3]
0 8 −16 −2
1
( )𝑅2 1 −3 5 1
2
→ [0 1 2 | 3/2]
0 8 −16 −2
(−8)𝑅23 1 −3 5 1
→ [0 1 2 | 3/2 ]
0 0 −32 −14
- 54 -
 (−1/32)𝑅3 1 −3 5 1
→ [0 1 2| 3/2 ]
0 0 1 7/16
The last augmented matrix is on row-echelon form.
Therefore, the equivalent system to the given system is
𝑥1 − 3𝑥2 + 5𝑥3 = 1, 𝑥2 + 2𝑥3 = 3/2, 𝑥3 = 7/16.
Substituting 𝑥3 = 7/16 into the second equation yields
𝑥2 = 5/8.
Finally, substituting 𝑥2 and 𝑥3 into the first equation
yields 𝑥1 = 11/16 .
Hence, the unique solution for the given system is
11 5 7
( , , ).
16 8 16
(b) Gauss – Jordan elimination.
We complete where we end in (a) to put the augmented
matrix in reduced row-echelon form.
1 −3 5 1
[0 1 2| 3/2 ]
0 0 1 7/16

(−2)𝑅32 ,(−5)𝑅31 1 −3 0 −19/16

→ [0 1 0| 5/8 ]
0 0 1 7/16

- 55 -
 3𝑅21 1 0 0 11/16
→ [0 1 0| 5/8 ]
0 0 1 7/16
The corresponding system of equation is
11 5
𝑥1 = , 𝑥2 = , 𝑥3 = 7/16.
16 8
Therefore, the unique solution of the system is (11/16,
5/8, 7/16) which we obtained in (a). ◄
Example 2.
Use Gauss and Gauss-Jordan to solve the following: -
𝑥1 − 3𝑥2 + 2𝑥3 + 𝑥4 = 2
3𝑥1 − 9𝑥2 + 10𝑥3 + 2𝑥4 = 9
2𝑥1 − 6𝑥2 + 4𝑥3 + 2𝑥4 = 4
2𝑥1 − 6𝑥2 + 8𝑥3 + 𝑥4 = 7
Solution.
(a) Gauss-elimination
1 −3 2 1 2
[𝐴|𝑏] = [ 3 −9 10 2 | 9]
2 −6 4 2 4
2 −6 8 1 7
1 −3 2 1 2
−3𝑅12 − 2𝑅13
=[0 0 4 −1 |3]
−2𝑅14 0 0 0 0 0
0 0 4 −1 3

- 56 -
 1 −3 2 1 2
1/4𝑅2 1 −1/4 3/4
→ [0 0 | ]
0 0 0 0 0
0 0 4 −1 3
1 −3 2 1 2
−4𝑅24 0 0 1 −1/4 3/4
→ [ | ]
0 0 0 0 0
0 0 0 0 0
The entire matrix is now in row-echelon form. The
corresponding system of equations is
𝑥1 − 3𝑥2 + 2𝑥3 + 𝑥4 = 2
1
𝑥3 − 𝑥4 = 3/4
4
Since 𝑥1 and 𝑥3 correspond to leading 1's in the
augmented matrix, we call them leading variables. The
no leading variables (in this ease 𝑥2 and 𝑥4 ) are called
free variables.
Solving for the leading variables in terms of the free
variable gives
𝑥1 = 2 + 3𝑥2 − 2𝑥3 − 𝑥4
3 1
𝑥3 = + 𝑥
4 4 4
3 1
Substituting 𝑥3 = + 𝑥4 into first equation yields
4 4

- 57 -
 1 3
𝑥1 = + 3𝑥2 − 𝑥4 ,
2 2
3 1
𝑥3 = + 𝑥
4 4 4
Assign arbitrary values to the free variables, we obtain
𝑥4 = 𝑡 and 𝑥2 = 𝑠.
Therefore, the general solution is given by the formulas
1 3 3 1
𝑥1 = + 3𝑠 − 𝑡, 𝑥2 = 𝑠, 𝑥3 = + 𝑡, 𝑥4 = 𝑡.
2 2 4 4

Remark.
The arbitrary values that are assigned to the free
variables are often called parameters. Although we shall
generally use the letters 𝑟, 𝑠, 𝑡, … for the parameters, any
letters that do not conflict with the variable names may
be used.
(b) Gauss-Jordan elimination.
We start from the end of (a). We put the augmented
matrix in reduced row-echelon form.
1 −3 2 1 2
1 −1/4 3/4
[0 0 | ]
0 0 0 0 0
0 0 0 0 0
1 −3 0 3/2 1/2
−2𝑅21 0 0 1 −1/4 3/4
→ [ | ]
0 0 0 0 0
0 0 0 0 0

- 58 -
 The last augmented matrix is in reduced row-echelon
form and the corresponding system is
3 1
𝑥1 − 3𝑥2 + 𝑥4 = ,
2 2
1 3
𝑥3 − 𝑥4 =
4 4
Putting 𝑥2 = 𝑠 and 𝑥4 = 𝑡 we obtain the set of solutions
3 1 3 1
𝑆 = {(3𝑠 − 𝑡 + , 𝑠, + 𝑡, 𝑡): 𝑡, 𝑠 ∈ ℝ}.◄
2 2 4 4

Example 3.
Solve
𝑥+𝑧=1
𝑦 − 𝑧 = −1
2𝑥 + 𝑦 + 𝑧 = 2
by Gauss and Gauss-Jordan elimination.
Solution.
(a)Gauss-Jordan elimination.
The corresponding augmented matrix is:
1 0 1 1 −2𝑅13 1 0 1 1
[𝐴|𝑏]=[ 0 1 −1| −1] → [ 0 1 −1| −1]
2 1 1 2 0 1 −1 0
−1𝑅23 1 0 1 1
→ [ 0 1 −1| −1]
0 0 0 1

- 59 -
 The corresponding system is
𝑥 + 𝑧 = 1, 𝑦 − 𝑧 = −1, 𝟎=𝟏
Which is impossible. Therefore, the given system of
equations has no solution. Hence it is inconsistent.
(b) Gauss-Jordan elimination.
Note that the augmented matrix is in reduced row-
echelon form. Therefore, we obtain the same results as
in (a). ◄
Remark.
Every system of linear equations has either no solutions,
exactly one solution, or infinitely many solutions.
Theorem 1.
If 𝐴𝑥 = 𝑏 is a consistent system of equations, then it has
exact one solution, or infinitely many solutions.
Proof.
Let 𝑥1 and 𝑥2 be two solutions of 𝐴𝑥 = 𝑏 has infinitely
solution, we prove that 𝑥1 + 𝑘(𝑥1 − 𝑥2 ), 𝑘 ∈ ℝ is also a
solution of 𝐴𝑥 = 𝑏. Now,
𝐴(𝑥1 + 𝑘 (𝑥1 − 𝑥2 )) = 𝐴𝑥1 + 𝑘 (𝐴𝑥1 ) − 𝑘 (𝐴𝑥1 )
= 𝑏 + 𝑘𝑏 − 𝑘𝑏 = 𝑏.

- 60 -
Therefore 𝑥1 + 𝑘 (𝑥1 − 𝑥2 ), 𝑘 ∈ ℝ is a solution of 𝐴𝑥 =
𝑏.◄
The above three examples are for system of equations
with number equal the number of unknowns.
Now, we give examples for system of equations with
number does not equal the number of unknowns.
Example 3. Use Gauss-Jordon to solve
3𝑥1 − 2𝑥2 = 4, 5𝑥1 + 𝑥2 = 1, 9𝑥1 + 7𝑥2 = −5
Solution.
3 −2 4 1/3𝑅1 1 −2/3 4/3
[𝐴|𝑏]=[5 1 | 1 ] → [5 1 | 1 ]
9 7 −5 9 7 −5
−5𝑅12 1 −2⁄3 4⁄3
(0 13⁄3 |−17⁄3)
−9𝑅13
0 13 −17
3⁄13𝑅2 1 −2/3
4/3
→ [0 1 | −17⁄13]
0 13 −17
−13𝑅23 1 −2/3
4/3 2⁄3𝑅21 1 0
6/13
→ [0 1 | −17⁄13] → [0 1| −17⁄13]
0 0 0 0 0 0
The corresponding system of equations is 𝑥1 = 6⁄13 and
𝑥2 = −17⁄13.◄

- 61 -
Exercise.
Use Gauss elimination to solve
3𝑥1 + 2𝑥2 = 4
9𝑥1 + 6𝑥2 = 12
−36𝑥1 − 24𝑥2 = −48
Example 3. Use Gauss-Jordon to solve
𝑥1 − 2𝑥2 + 𝑥3 = 3
𝑥1 − 2𝑥2 + 𝑥3 = 4
Solution.

[𝐴|𝑏]=[1 −2 1 3 −1𝑅12 1 −2 1 3
| ]→ [ | ].
1 −2 1 4 0 0 0 1
The corresponding system is
𝑥1 − 2𝑥2 + 𝑥3 = 3
0 = 1.
Therefore, the system has no solution. ◄
Remark.
If the number of unknowns is greater than the number of
equations, then the system either has infinitely many
solutions or it is inconsistent.

- 62 -
Example 4.
For which values of a will the following system have no
solutions? Exactly one solution? Infinitely many
solutions?
𝑥1 + 𝑥2 + 𝑥3 + 𝑥4 = 4
𝑥1 + 𝑎𝑥2 + 𝑥3 + 𝑥4 = 4
𝑥1 + 𝑥2 + 𝑎𝑥3 + (3 − 𝑎)𝑥4 = 6
2𝑥1 + 2𝑥2 + 2𝑥3 + 𝑎𝑥4 = 6.
Solution.
We use Gauss elimination to put the augmented matrix in
row-echelon form.
1 1 1 1 4
[𝐴|𝑏] = [1 𝑎 1 1 | 4] =
1 1 𝑎 3−𝑎 6
2 2 2 𝑎 6

1 1 1 1 4
−1𝑅12 , −1𝑅13 0 𝑎−1 0 0
[ |0]
−2𝑅14 0 0 𝑎−1 2−𝑎 2
0 0 0 𝑎−2 −2
… … . . (∗)
If 𝑎 ≠ 1, 2 , then

- 63 -
 (*)
1 1 1 1 4
1 1 1
𝑅 , 𝑅 , 𝑅
𝑎−1 2 𝑎−1 3 𝑎−2 4 0 1 0 0 0
→ 2−𝑎 | 2/(𝑎 − 1)
0 0 1
𝑎−1
[0 0 0 1 −2/(𝑎 − 2)]

2−𝑎 1 1 1 1 4
−( )𝑅
𝑎−1 43 0
→ [0 1 0 0 | 4/(𝑎 − 1) ]
0 0 1 0
0 0 0 1 −2/(𝑎 − 2)
1 1 1 0 4 + 2/(𝑎 − 2)
−1𝑅41 0 1 0 0 0
→ [ | 4/(𝑎 − 1) ]
0 0 1 0
0 0 0 1 −2/(𝑎 − 2)
2 4
1 1 0 0 4+ −
(𝑎 − 2) (𝑎 − 1)
−1𝑅31 0 1 0 0 |
→ 0
0 0 1 0 |
4/(𝑎 − 1)
0 0 0 1
[ −2/(𝑎 − 2) ]
2 4
1 0 0 0 4+ −
(𝑎 − 2) (𝑎 − 1)
−1𝑅21 0 1 0 0 |
→ 0
0 0 1 0 |
4/(𝑎 − 1)
0 0 0 1
[ −2/(𝑎 − 2) ]

- 64 -
Therefore, the unique solution is
2 4 4 −2
(4 + − , 0, , )
(𝑎 − 2) (𝑎 − 1) (𝑎 − 1) 𝑎−2
Now, if 𝑎 = 2 , then the matrix (*) becomes
1 1 1 1 4
0 1 0 0 0
[ | ]
0 0 1 0 2
0 0 0 0 −2
which implies 0 = −2, impossible. Then the system has
no solution. Finally, if 𝑎 = 1, then (*) becomes
1 1 1 1 4
0 0 0 0 0
[ | ]
0 0 0 1 2
0 0 0 −1 −2
which is equivalent to the following reduced row-
echelon form.
1 1 1 1 4
0 0 0 1 2
[ | ]
0 0 0 0 0
0 0 0 0 0
The corresponding system is

𝑥1 + 𝑥2 + 𝑥3 = 2,

𝑥4 = 2.

- 65 -
Let 𝑥3 = 𝑡 and 𝑥2 = 𝑠. Then the solution set is 𝑆 =

{(2 − 𝑠 − 𝑡, 𝑠, 𝑡, 2): 𝑠, 𝑡 ∈ ℝ} which is an infinitely

many solutions. ◄

Theorem 2.
If A is invertible 𝑛 × 𝑛 matrix, for each 𝑛 × 1 matrix 𝑏,
the system of equation 𝐴𝑥 = 𝑏 has exactly one solution,
namely 𝑥 = 𝐴−1 𝑏.
Proof.
Since 𝐴(𝐴−1 𝑏) = 𝑏, it follows that 𝑥 = 𝐴−1 𝑏 is a
solution of 𝐴𝑥 = 𝑏. To show that this is the only
solution, we will assume that 𝑥0 is an arbitrary solution
and then we show that 𝑥0 must be the solution 𝐴−1 𝑏.
If 𝑥0 is any solution, then 𝐴𝑥0 = 𝑏. Multiplying both
sides by 𝐴−1 , we obtain 𝑥0 = 𝐴−1 𝑏.◄
Example 5.
Consider the system of linear equations
𝑥1 − 2𝑥2 + 2𝑥3 = 3
2𝑥1 + 𝑥2 + 𝑥3 = 0
𝑥1 + 𝑥3 = −2

- 66 -
The matrix form of this system can be written as 𝐴𝑥 =
𝑏, where
1 −2 2 𝑥1 3
𝐴 = [2 1 1] , 𝑥 = [𝑥2 ] 𝑏 = [ 0 ].
1 0 1 𝑥3 −2
By elementary row operation, we obtain that
1 2 −4
−1
𝐴 = [−1 −1 3 ].
−1 −2 5
Therefore, the unique solution is
𝑥1 1 2 −4 3 11
𝑥 = [𝑥2 ] = 𝐴 𝑏 = [−1 −1 3 ] [ 0 ] = [ −9 ].◄
−1
𝑥3 −1 −2 5 −2 −13
Corollary 1.
If is A is 𝑛 × 𝑛 matrix, then the following are equivalent.
(a) A is invertible.
(b) The reduced row-echelon form of A is In.
(c) A is expressible as a product of elementary
matrices.
(d) 𝐴𝑥 = 𝑏 is consistent for every 𝑛 × 1 matrix b.
(e) 𝐴𝑥 = 𝑏 has exactly one solution for every 𝑛 × 1
matrix b.
(f) Det 𝐴 ≠ 0.

- 67 -
Example 6.
What conditions must 𝑏1 , 𝑏2 , and 𝑏3 satisfy for the
system of equations
𝑥1 − 2𝑥2 + 5𝑥3 = 𝑏1
4𝑥1 − 5𝑥2 + 8𝑥3 = 𝑏2
−3𝑥1 + 3𝑥2 − 3𝑥3 = 𝑏3
to be consistent?
Solution.
The augmented matrix is
1 −2 5 𝑏1
[𝐴|𝑏]=[ 4 −5 8 | 𝑏2 ]
−3 3 −3 𝑏3
which can be reduced to row-echelon form as follows.
𝑏1
−4𝑅12 1 −2 5
[𝐴|𝑏] [0 3 −12| 𝑏2 − 4𝑏1 ]
3𝑅13
0 −3 12 𝑏3 + 3𝑏1

1/3𝑅2 1 −2 5 𝑏1
→ [0 1 −4| (𝑏2 − 4𝑏1 )/3]
0 −3 12 𝑏3 + 3𝑏1

3𝑅23 1 −2 5 𝑏1
→ [0 1 −4| (𝑏2 − 4𝑏1 )/3].
0 0 0 𝑏3 + 𝑏2 −𝑏1

- 68 -
It is now evident from the third row in the matrix that
the system has a solution if and only if 𝑏1 , 𝑏2 , and 𝑏3
satisfy the condition
𝑏3 + 𝑏2 − 𝑏1 = 0 or 𝑏3 = 𝑏1 − 𝑏2 .
To express this condition another way, 𝐴𝑥 = 𝑏 is
consistent if and b is a matrix of the form
𝑏1
𝑏 = [ 𝑏2 ],
𝑏1 − 𝑏2
where 𝑏1 and 𝑏2 are arbitrary.

- 69 -
 Exercise Set (2.1)
In exercise (1)-(15), Find all solutions of the given
systems using Gauss and Gauss-Jordan.
(1) 𝑥 + 𝑦 = 1 (2) 𝑥 + 2𝑦 = 2
𝑥−𝑦 =0 𝑥 − 4𝑦 = −1

(3) 3𝑥 + 7𝑦 − 3𝑧 = 2 (4) 3𝑥 + 7𝑦 − 3𝑧 = 2
2𝑥 + 5𝑦 + 𝑧 = −4 2𝑥 + 5𝑦 + 𝑧 = −4
2𝑥 + 6𝑦 + 10𝑧 = 3 2𝑥 + 6𝑦 + 10𝑧 = −20

(5) 𝑥 + 𝑦 + 𝑧 = 4 (6) 𝑥 − 2𝑦 − 𝑧 = 1
2𝑥 + 5𝑦 − 2𝑧 = 3 𝑥+𝑦−𝑧 =2
𝑥 + 7𝑦 − 7𝑧 = 5 𝑥 + 2𝑦 − 2𝑧 = 2
(7) 𝑥1 − 𝑥2 + 2𝑥3 − 𝑥4 = −1 (8) 𝑥1 − 2𝑥2 + 𝑥3 − 4𝑥4 = 1
2𝑥1 + 𝑥2 − 2𝑥3 − 2𝑥4 = −2 𝑥1 + 3𝑥2 + 7𝑥3 + 2𝑥4 = 2
−𝑥1 − 2𝑥2 − 4𝑥3 + 𝑥4 = 1 𝑥1 − 12𝑥2 − 11𝑥3 − 16𝑥4 = 5
3𝑥1 − 3𝑥4 = −3 2𝑥1 + 3𝑥2 + 3𝑥3 + 5𝑥4 = 6
(9) 2𝑥1 + 𝑥2 + 𝑥3 + 4𝑥4 = 2 (10) 𝑥 + 𝑦 = 0
𝑥1 + 𝑥2 + 𝑥3 − 2𝑥4 = 0 2𝑥 − 𝑦 = 1
2𝑥2 + 𝑥3 − 4𝑥4 = 1 𝑥 + 2𝑦 = −1

- 70 -
(11)𝑥1 + 𝑥2 + 𝑥3 + 𝑥4 = 2 (12) 2𝑥1 + 𝑥2 − 3𝑥3 = 1
2𝑥1 + 𝑥2 + 3𝑥3 − 𝑥4 = 1 𝑥1 − 𝑥2 + 2𝑥3 = 2

(13) 3𝑥 − 2𝑦 = 4 (14)𝑥2 + 𝑥3 = 1
5𝑥 + 𝑦 = 1 𝑥1 − 𝑥3 = 2
9 𝑥 + 7𝑦 = −5 −2 𝑥1 − 𝑥2 = 3

(15) 𝑥1 + 3𝑥2 − 2𝑥3 + 2𝑥5 = 0

2𝑥1 + 6𝑥2 − 5𝑥3 − 2𝑥4 + 4𝑥5 − 3𝑥6 = −1
5𝑥3 + 10𝑥4 + 15𝑥6 = 5
2𝑥1 + 6𝑥2 + 8𝑥4 + 4𝑥5 + 18𝑥6 = 6

(16) Find conditions that b's must satisfy for the system to be
consistent
(𝑎) 2𝑥 + 3𝑦 + 5𝑧 = 𝑏1 (𝑏)𝑥1 + 𝑥2 + 2𝑥3 = 𝑏1
4𝑥 − 5𝑦 + 8𝑧 = 𝑏2 𝑥1 + 𝑥3 = 𝑏2
3𝑥 + 3𝑦 + 2𝑧 = 𝑏3 2𝑥1 + 𝑥2 + 3𝑥3 = 𝑏3

- 71 -
(𝑐) 𝑥1 − 𝑥2 + 3𝑥3 + 2𝑥4 = 𝑏1
−2𝑥1 + 𝑥2 + 5𝑥3 + 𝑥4 = 𝑏2
−3𝑥1 + 2𝑥2 + 2𝑥4 − 𝑥4 = 𝑏3
4𝑥1 − 3𝑥2 + 𝑥3 + 3𝑥4 = 𝑏4
(17) Solve the following systems:

(𝑎)𝑥 2 + 𝑥𝑦 − 𝑦 2 = 1 (𝑏) 𝑥𝑦 − 2√𝑦 + 3𝑧𝑦 = 8

2𝑥 2 − 𝑥𝑦 + 3𝑦 2 = 13 2𝑥𝑦 − 3√𝑦 + 2𝑧𝑦 = 7
𝑥 2 + 3𝑥𝑦 + 2𝑦 2 = 0
−𝑥𝑦 + √𝑦 + 2𝑧𝑦 = 4
(18) Find a, b, and c such that the following system has the
unique solution (1,-1,2).
𝑎𝑥 + 𝑏𝑦 − 3𝑧 = −3
−2𝑥 − 𝑏𝑦 + 𝑧 = −1
𝑎𝑥 + 3𝑦 − 𝑐𝑧 = −1

- 72 -
2. HOMOGENEOUS SYSTEMS OF LINEAR EQUATIONS
A system of linear equation 𝐴𝑥 = 𝑏, where 𝐴 is a
matrix of size 𝑚 × 𝑛, 𝑥 is a matrix of size 𝑛 × 1, and b a
matrix of size 𝑚 × 1, is said to be homogeneous if the
matrix 𝑏 is a zero matrix, 𝑖. 𝑒. , 𝑏 = 𝟎. If 𝑏 ≠ 𝟎. then
𝐴𝑥 = 𝑏 is a non- homogeneous system.
Every homogeneous system of linear equations is
consistent, since (0, 0, … ,0) is a solution of this system.
This solution is called the trivial solution; if there are
other solutions, they are called nontrivial solutions,
nontrivial solutions.
Because a homogeneous linear system always has the
trivial solution, there are only two possibilities for its
solutions.
• The system has only the trivial solution.
• The system has infinitely many solutions in addition
to the trivial solution.

- 73 -
 Theorem 1.
If 𝑥 and 𝑦 are solutions for the homogeneous system
𝐴𝑥 = 𝟎, then both 𝑥 + 𝑦 and 𝑘𝑥 are solutions for the
system.
Proof.
Since 𝑥, 𝑦 are solutions for the system, then 𝐴𝑥 = 𝟎 and
𝐴𝑦 = 𝟎. Therefore 𝐴(𝑥 + 𝑦) = 𝐴𝑥 + 𝐴𝑦 = 𝟎 + 𝟎 = 𝟎
and 𝐴(𝑘𝑥 ) = 𝑘(𝐴𝑥 ) = 𝑘. 𝟎 = 𝟎. Hence both 𝑥 + 𝑦 and
𝑘𝑥 are solution of 𝐴𝑥 = 𝟎.◄
Definition 2.
Let 𝑈1 , 𝑈2 , … … , 𝑈𝑘 be an 𝑛 × 1 (or 1 × 𝑛) matrices.
Then an 1 × 𝑛 matrix is a linear combination of
𝑈1 , 𝑈2 , … … , 𝑈𝑘 if there exist reals 𝛼1 , 𝛼2 , … … , 𝛼𝑘 such
that 𝑈 = 𝛼1 𝑈1 + 𝛼2 𝑈2 + ⋯ + 𝛼𝑘 𝑈𝑘 .
Note that by the above theorem, any linear combination
of solutions of 𝐴𝑥 = 𝟎 is also a solution. We use Gauss
and Gauss-Jordan in solving homogeneous system. Also,
we write this solution as a linear combination of other
solutions.

- 74 -
Example 1.
Solve the following homogeneous system of linear
equations by Gauss-Jordan elimination:
𝑥1 − 2𝑥2 + 𝑥3 + 𝑥4 = 0
−𝑥1 + 2𝑥2 + 𝑥4 = 0
2𝑥1 − 4𝑥2 + 𝑥3 = 0
Write the solution as a linear combination of the two
solutions (1,0, −2, 1) and (2, 1, 0, 0).
Solution.
The augmented matrix for the system is
1 −2 1 1 0
[−1 2 0 1| 0].
2 −4 1 0 0
Reducing this matrix to reduced row-echelon form, we
obtain
1 −2 0 1 0
[ 0 0 1 2 | 0]
0 0 0 0 0
The corresponding system of equations is
𝑥1 − 2𝑥2 − 𝑥4 = 0
𝑥3 + 2𝑥4 = 0
Solving for the leading variables yields
𝑥1 = 2𝑥2 + 𝑥4
- 75 -
 𝑥3 = −2𝑥4
Thus, the general solution is
𝑥1 = 2𝑠 + 𝑡, 𝑥2 = 𝑠, 𝑥3 = −2𝑡, 𝑥4 = 𝑡.
Hence the set of solutions 𝑆 = {(2𝑠 + 𝑡, 𝑠, −2𝑡, 𝑡) ∶
𝑆, 𝑡 ∈ ℝ}. Therefore
(2𝑠 + 𝑡, 𝑠, −2𝑡, 𝑡) = (2𝑠, 𝑠, 0,0) + (𝑡, 0, −2𝑡, 𝑡)
= 𝑠(2,1,0,0) + 𝑡(0,0, −2,1).
Which is a linear combination of (2,1,0,0) and
(0,0, −2,1).◄
Theorem 2.
A homogeneous system of linear equations with more
unknowns than equations has infinitely many solutions.
Remark. Note that the above theorem applies only to
homogeneous system. A non-homogeneous system with
more unknowns than equations need not be consistent;
however, if the system is consistent, it will have infinitely
many solutions.
Theorem 3.
If 𝑥∘ is a solution of 𝐴𝑥 = 𝑏, then any solution of 𝐴𝑥 =
𝑏 is in the form 𝑥∘ + 𝑥1 , where 𝑥1 is a solution of 𝐴𝑥 =
𝟎.
- 76 -
Proof.
Since 𝐴(𝑥∘ + 𝑥1 ) = 𝐴𝑥∘ + 𝐴𝑥1 = 𝑏 + 𝟎 = 𝑏, then any
solution of 𝐴𝑥 = 𝑏 is in the form 𝑥∘ + 𝑥1 . Now suppose
that 𝑦 is a solution of 𝐴𝑥 = 𝑏. We prove that 𝑦 = 𝑥∘ +
𝑥1 . Put 𝑥1 = 𝑦 − 𝑥∘ . Then
𝐴𝑥1 = 𝐴(𝑦 − 𝑥∘ ) = 𝐴𝑦 − 𝐴𝑥∘ = 𝑏 − 𝑏 = 𝟎.◄
Example 2.
If 𝑥∘ = (1, −1, 0, 0) is a solution of the system
𝑥1 + 2𝑥4 = 1, 𝑥2 − 𝑥4 = −1, 𝑥3 + 𝑥4 = 0, then find all
solutions of the system.
Solution. By Gauss-Jordan, (−2𝑡, 𝑡, −𝑡, 𝑡) is a solution
of the homogenous system. Then the solutions of the
given system are (−2𝑡, 𝑡, −𝑡, 𝑡) + (1, −1, 0, 0), 𝑡 ∈
ℝ.◄
Corollary 1.
If A is a square matrix, then 𝐴𝑥 = 𝟎 has the trivial
solution if and only if A is invertible.
Example 3.
Prove that the following system has only the trivial
solution.
𝑥1 − 2𝑥2 + 2𝑥3 = 0
- 77 -
 2𝑥1 + 𝑥2 + 𝑥3 = 0
𝑥1 + 𝑥3 = 0
Solution.
1 −2 2
Since det 𝐴 = |2 1 1| = 1 ≠ 0, then 𝐴 is
1 0 1
invertible. Therefore, the given system has only the trivial
solution
We end this section by the following theorem.

Theorem 4.

If A is an 𝑛 × 𝑛 matrix, then the following are equivalent.

(1) 𝐴 is in veritable.

(2) 𝐴𝑥 = 𝟎 has only the trivial solution.

(3) The reduced row-echelon from A is 𝐼𝑛 .

(4) 𝐴 is expressible as a product of elementary matrices.

(5) 𝐴𝑥 = 𝑏 is consistent for every 𝑛 × 1 matrix 𝑏.

(6) 𝐴𝑥 = 𝑏 has exactly one solution for every 𝑛 × 1

matrix 𝑏.

(7) det (𝐴) ≠ 0.

- 78 -
 Exercise Set (2.2)
(1) Find a such that the following two systems has
infinitely many solutions?
(a) 𝑥 − 2𝑦 + 𝑧 = 0,
𝑥 + 𝑎𝑦 − 3𝑧 = 0,
−𝑥 + 𝑏𝑦 − 5𝑧 = 0
(b) 𝑥 − 𝑦 − 𝑧 = 0, 𝑥𝑦 − 𝑧 = 0, 𝑥 + 𝑦 + 𝑎𝑧 = 0
(3) Write the general solution of the following system
as a linear combination of particular solutions.
𝑥1 + 2𝑥2 + 𝑥3 − 𝑥4 + 3𝑥5 = 0
𝑥1 + 2𝑥2 + 2𝑥3 + 𝑥4 + 2𝑥5 = 0
2𝑥1 + 4𝑥2 + 2𝑥3 − 𝑥4 + 7𝑥5 = 0
(3) Prove that the following system has only the trivial
solution.
𝑥 + 𝑦 = 0, 𝑥 − 𝑦 = 0, 2𝑥 + 3𝑦 = 0

- 79 -
CHAPTER (III)
DETERMINATES

- 80 -
 CHAPTER (III)
DETERMINATES
1. DETERMINATES
For any 𝑛 × 𝑛 matrix 𝐴 we can define the determinant of
𝐴 as function. Its domain is the set of 𝑛 × 𝑛 matrix and its
co-domain is the set of real numbers. The determinant
function is very important. It gives us some information
about the matrix, such as its inverse and how we compute
it. Also, we can solve a system of linear equation as we
will see. There are many methods to define the
determinant of the matrix. We choose the inductive
definition, 𝑖. 𝑒., we define the determinat for a matrix of
degree 1 and 2, and we define the determinat of an 𝑛 × 𝑛
matrix.
Let 𝐴 be an 𝑛 × 𝑛 matrix. The (𝑛 − 1) × (𝑛 − 1) sub-
matrix 𝑀𝑖𝑗 is obtained by deleting the 𝑖𝑡ℎ row and
𝑗𝑡ℎ column from 𝐴.
1 1 0
2 1
For example, if 𝐴 = [ 2 1 1], then 𝑀12 = [ ]
−1 3
−1 0 3
1 0
and 𝑀22 = [ ].
−1 3
- 81 -
Definition 1.
Let 𝐴 = [𝑎𝑖𝑗 ] be an 𝑛 × 𝑛 matrix the determinat of 𝐴 is
denoted by det 𝐴 or |𝐴|and is defined by inductive as:
(1) If 𝑛 = 1, then det 𝐴 = 𝑎11
(2) If 𝑛 = 2, then det 𝐴 = 𝑎11 𝑎22 − 𝑎12 𝑎21
(3) If 𝑛 > 2, then
det 𝐴 = 𝑎11 det 𝑀11 − 𝑎12 det 𝑀12 + ⋯ +
𝑛

(−1)𝑛+1 𝑎1𝑛 det 𝑀1𝑛 = ∑(−1)𝑗+1 𝑎𝑖𝑗 det 𝑀𝑖𝑗

𝑗=1

Example 1.
1 3 −2
If 𝐴 = [−1 2 4 ], then
3 1 0
2 4 −1 4 −1 2
det 𝐴 = 1 | | − 3| | + (−2) | |
1 0 3 0 3 1
= (0 − 4) − 3(0 − 12) − 2(−1 − 6)
= 46.◄

Example 2.
2 1 −3 1
−3−2 0 2
Let 𝐴 be the matrix [ ].
2 1 0 −1
1 0 1 2
- 82 -
Then
det 𝐴 = 2 det 𝑀11 − det 𝑀12 − det 𝑀13 − det 𝑀14 ,where
−2 0 2 −3 0 2
𝑀11 = [ 1 0 −1] , 𝑀12 = [ 2 0 −1],
1 0 2 1 1 2
−3 −2 −2 −3 −2 0
𝑀13 = [ 2 1 −1] , 𝑀14 = [ 2 1 0].
1 0 2 1 0 1

Therefore,
det 𝑀11 = 0, det 𝑀12 = 1, det 𝑀13 = 2, det 𝑀14 = 1.
Hence det 𝐴 = 2(0) − 1 − 3(2) − 1 = −8.◄
●Note that we defined the determinant by inductive using
the first row of the matrix. But we use any row or any
column.
Definition 2.
Let 𝐴 = [𝑎𝑖𝑗 ] be can 𝑛 × 𝑛 matrix. The minor of 𝑎𝑖𝑗 is
det 𝑀𝑖𝑗 and the cofactor 𝐶𝑖𝑗 of 𝑎𝑖𝑗 is (−1)𝑖+𝑗 det 𝑀𝑖𝑗 .
Example 3.
2 5 −21
1−2 0 1
Let 𝐴 = [ ].Then
33 40
4 1 −16

- 83 -
 2 −2 1
4+2
𝐶42 = (−1) det 𝑀42 = [1 0 1] = −10.◄
3 4 0
● Note that det 𝐴 is the sum of the product of the

component of first row by their cofactors 𝑖. 𝑒.,

det 𝐴 = 𝑎11 𝑐11 + 𝑎12 𝑐12 + ⋯ + 𝑎1𝑛 𝑐1𝑛 .

In fact, we can find the determinant using any row or any

column in the matrix.

That is, for each 𝑖 and 𝑗, 1 ≤ 𝑖 ≤ 𝑛 and 1 ≤ 𝑗 ≤ 𝑛, we

have det 𝐴 = 𝑎𝑖1 𝑐𝑖1 + 𝑎𝑖2 𝑐𝑖2 + ⋯ + 𝑎𝑖𝑛 𝑐𝑖𝑛 ,

where the summation is called the cofactor expansion of

det 𝐴 along the 𝑖 𝑡ℎ row.

Similarly, for each 𝑖 and 𝑗, 1 ≤ 𝑖 ≤ 𝑛 and 1 ≤ 𝑗 ≤ 𝑛, we

have

det 𝐴 = 𝑎1𝑗 𝑐1𝑗 + 𝑎2𝑗 𝑐2𝑗 + ⋯ + 𝑎𝑛𝑗 𝑐𝑛𝑗 ,

where the summation is called the cofactor expansion of

det 𝐴 along the 𝑗𝑡ℎ column.

- 84 -
Example 4.
2 5 −2 1
10 0 1
If 𝐴 = [ ], then we choose the second column
30 4 0
4 0 −1 2
to compute det 𝐴 as it contains many zeros. Therefore
det 𝐴 = 𝑎12 𝑐12 + 𝑎22 𝑐22 + 𝑎32 𝑐32 + 𝑎42 𝑐42
= 𝑎12 𝑐12 + 0 + 0 + 0
1 0 1
= −5 |3 4 0|
4 −1 2
= −5[1(8 + 0) − 0(6 − 0) + 1(−3 − 16)] = 55.◄

Exercise set (3.1)

In exercise (1) – (8).
Find the determinate of the given matrix:
2 3 3𝑎 𝑎
(1) [ ] (2) [ ]
1 −2 3𝑏 𝑏
1 1 0 1 0 4
(3) [−1 2 3] (4) [0 −2 3]
3 −4 4 2 −2 2
00 1 0 1 −2 1 0
00 2 0 −1 0 −3 3
(5) [ ] (6) [ ]
0 3 −1 2 0 3 4 −1
5−2 3 −3 0 0 −1 2

- 85 -
 −11 0 1 −1 4 5 −2
311 0 3 −12 7 −3
(7) [ ] (8) [ ]
0 1 2 −3 2 −8 8 4
1 2−3 2 1 −4 −1 2
𝑘−1 1
(9) Find 𝑘 such that | | = 0.
2 𝑘
1 −1 1
(10) Find 𝑘 such that |−1 𝑘 1| = 0.
−1 −1 1
𝑘−4 0 0
(11) Find 𝑘 such that | 0 𝑘 2 | = 0.
0 3 𝑘−1
1 0 −3
𝜆 −1
(12) Find 𝑘 such that | | = |2 𝜆 −6 |.
3 1−𝜆
1 3 𝜆−5
sin 𝜃 cos 𝜃 0
(13) Find | − cos 𝜃 sin 𝜃 0|.
sin 𝜃 − cos 𝜃 sin 𝜃 + cos 𝜃 1

- 86 -
3.2 Cramer' Rule
Theorem 1. (Cramer' Rule).
If 𝐴𝑥 = 𝑏 is a system of 𝑛 linear equations in 𝑛
unknowns such that det (𝐴) ≠ 0. Then the system has
a unique solution. This solution is
det (𝐴1 ) det (𝐴2 ) det (𝐴𝑛 )
𝑥1 = , 𝑥2 = , … , 𝑥𝑛 =
det (𝐴) det (𝐴) det (𝐴)
Where 𝐴𝑗 is the matrix obtained by replacing the
entries in the 𝑗𝑡ℎ column of 𝐴 by the entries in the
𝑏1
𝑏
matrix 𝑏 = [ 2 ]
⋮
𝑏𝑛
Proof.
Since det (𝐴) ≠ 0, then 𝐴 is invertible and𝑥 = 𝐴−1 𝑏
is the unique solution of 𝐴𝑥 = 𝑏.
Therefore
1
𝑥 = 𝐴−1 𝑏 = 𝑎𝑑j (𝐴). 𝑏
det(𝐴)
𝑐11 𝑐21 …𝑐𝑛1 𝑏1
1 𝑐12 𝑐22 …𝑐𝑛2 𝑏
= [ ⋮ ⋮ … ⋮ ] [ 2]
det (𝐴) ⋮
𝑐1𝑛 𝑐2𝑛 …𝑐𝑛𝑛 𝑏𝑛

- 87 -
 1
Therefore𝑥𝑖 = (𝑏1 𝑐1𝑖 + 𝑏2 𝑐2𝑖 + ⋯ + 𝑏𝑛 𝑐𝑛𝑖 ),
det(𝐴)

𝑖 = 1,2, … , 𝑛.
Also, computing det𝐴𝑖 using column 𝑖 we obtain
det 𝐴𝑖 = 𝑏1 𝑐1𝑖 + 𝑏2 𝑐2𝑖 + ⋯ + 𝑏𝑛 𝑐𝑛𝑖
det 𝐴𝑖
Therefore 𝑥𝑖 = , 𝑖 = 1,2, … , 𝑛.◄
det 𝐴

Example 1.
Use Cramer's Rule to solve
𝑥1 + 2𝑥3 = 6
−3𝑥1 + 4𝑥2 + 6𝑥3 = 30
−𝑥1 − 2𝑥2 + 3𝑥3 = 8
Solution.
1 0 2 6 0 2
𝐴 = [−3 4 6] , 𝐴1 = [30 4 6]
−1 −2 3 8 −2 3
1 6 2 1 0 6
𝐴2 = [−3 30 6] , 𝐴3 = [−3 4 30]
−1 8 3 −1 2 8
Therefore
det(𝐴1 ) −10 det (𝐴2 ) 72 18
𝑥1 = = , 𝑥2 = = =
det(𝐴) 11 det (𝐴) 44 11
𝑑𝑒𝑡 (𝐴3 ) 152 38
𝑥3 = − = .◄
det (𝐴) 44 11

- 88 -
 Exercise Set (3.2)
(1) Use Cramer's Rule to solve the following systems
(a) 2𝑥 − 3𝑦 = 7 (b) 8𝑥 − 6𝑦 = −4
8𝑥 + 𝑦 = −2 3𝑥 + 2𝑦 = 6
(c) 2𝑥 − 6𝑦 + 𝑍 = 2 (d) 𝑥 − 3𝑦 − 𝑍 = −7
𝑦+𝑍 =1 𝑥 − 𝑦 − 𝑍 = −2
𝑥−𝑦−𝑍 =0 𝑥 − 6𝑦 − 2𝑍 = −3
(e) 𝑥1 + 𝑥2 + 𝑥3 = 𝑎 (f) 6𝑥 − 𝑦 + 3𝑍 = −3
𝑥1 + (1 + 𝑎)𝑥2 + 𝑥3 = 2𝑎 9𝑥 + 5𝑦 + 2𝑍 = 7
𝑥1 + 𝑥2 + (1 + 𝑎)𝑥3 = 0 5𝑥 + 𝑦 − 8𝑍 = −2
(5) Is it possible to use Cramer's Ruler in solving the
following System?
2𝑥 + 3𝑦 + 𝑍 = 9
𝑥 + 2𝑦 + 3𝑍 = 6
𝑥 + 𝑦 + 2𝑍 = 3
Why? Is there a solution for the system?

- 89 -
CHAPTER IV
LINEAR SPACE

- 90 -
 CHAPTER IV
LINEAR SPACE
1. Euclidean n-space
Definition 1.
Let 𝑛 ≥ 1 . The set of all ordered n- tuples of real
numbers is called Euclidean space of dimension n and
will be denoted by ℝ𝑛 . That is:
ℝ𝑛 = {(𝑎1 , 𝑎2 , … , 𝑎𝑛 ) ∶ 𝑎𝑖 ∈ ℝ, 1 ≤ 𝑖 ≤ 𝑛}
Note.
We will use the row or the column to express the
elements of ℝ𝑛 and will be called n- vector. The addition
and scalar multiplication on ℝ𝑛 are defined as follows.
Definition 2.
Let 𝒖 = (𝑢1 , 𝑢2 , … , 𝑢𝑛 ), 𝒗 = (𝑣1 , 𝑣2 , … , 𝑣𝑛 ) ∈ ℝ𝑛 and
𝛼 ∈ ℝ , then 𝒖 + 𝒗 = (𝑢1 + 𝑣1 , 𝑢2 + 𝑣2 , … , 𝑢𝑛 + 𝑣𝑛 ) and
𝛼𝒖 = (𝛼𝑢1, 𝛼𝑢2 , … , 𝛼𝑢𝑛 ).
The zero element of ℝ𝑛 and the inverse of 𝒖 (𝑖. 𝑒, −𝒖)
are 𝟎 = (0,0, … ,0) and −𝒖 = (−𝑢1 , −𝑢2 , … , −𝑢𝑛 ),
respectively, 𝒖 − 𝒗 = 𝒖 + (−𝒗)
= (𝑢1 − 𝑣1 , 𝑢2 − 𝑣2 , … , 𝑢𝑛 − 𝑣𝑛 )
- 91 -
Example 1.
Let 𝒖 = (2, 6, −4) and 𝒗 = (−1, 3, 5) be elements of
ℝ3 . Find 𝒖 + 𝒗, 𝒖 − 𝒗, 3𝒖 and −2𝒗 .
Solution.
𝒖 + 𝒗 = (2, 6, −4) + (−1, 3, 5) = (1, 9, 1);
𝒖 − 𝒗 = 𝒖 + (−𝒗) = (2,6, −4) + (1, −3, −5)
= (3,3, −9);
3𝒖 = 3(2, 6, −4) = (6, 18, −12);
−2𝒗 = −2(−1, 3, 5) = (2, −6, −10).◄
For every positive integer n the set of ℝ𝑛 satisfies some
properties listed in the following theorem. Since these
properties are easy to establish, the proof of this theorem
is left as an exercise.
Theorem 1.
If 𝒖, 𝒗 and w are any vectors in ℝ𝑛 and if 𝛼, 𝛽 are any
real numbers, then
1. 𝒖 + 𝒗 is an n-vector;
2. 𝒖 + 𝒗 = 𝒗 + 𝒖 ;
3. (𝒖 + 𝒗) + 𝒘 = 𝒖 + (𝒗 + 𝒘);
4. There is a zero element 0 in ℝ𝑛 such that
𝒖 + 𝟎 = 𝟎 + 𝒖 = 𝒖 for every u in ℝ𝑛 ;
- 92 -
5. There is an n-vector –u such that 𝒖 + (−𝒖) = 𝟎 ;
6. 𝛼𝒖 is an n- vector;
7. (𝛼𝛽 )𝒖 = 𝛼 (𝛽𝒖);
8. (𝛼 + 𝛽 )𝒖 = 𝛼𝒖 + 𝛽𝒖 ;
9. 𝛼 (𝒖 + 𝒗) = 𝛼𝒖 + 𝛼𝒗 ;
10. 1. 𝒗 = 𝒗.
2. Linear spaces
In Section 1 above, we found that ℝ𝑛 possesses several
arithmetic properties that do not depend on the positive
integer. The operation of addition is both commutative
and associative, while the operation of scalar
multiplication satisfies several distributive properties. In
addition, ℝ𝑛 contains a zero element and every vector in
ℝ𝑛 has an additive inverse. We now define an abstract
mathematical system that embodies the arithmetic
properties of ℝ𝑛 .
Definition1.
A linear space V over the set of real numbers is a
nonempty set with two laws of combination, called
addition and scalar multiplication, satisfying the

- 93 -
following conditions for all 𝒖, 𝒗 and w in V and all real
numbers 𝛼 and 𝛽.
1. To every pair 𝒖, 𝒗 in V there is only one element in V,
called the sum of u and 𝒗, denoted 𝒖 + 𝒗 (closure
under addition).
2. 𝒖 + 𝒗 = 𝒗 + 𝒖 (Commutative property of addition).
3. (𝒖 + 𝒗) + 𝒘 = 𝒖 + (𝒗 + 𝒘) (Associate property of
addition).
4. There is a zero element 0 in V such that
𝒖 + 𝟎 = 𝟎 + 𝒖 = 𝒖 for all u in V (Existence of
additive identity).
5. To every element u in V there corresponds an additive
inverse element –u in V such that 𝒖 + (−𝒖) = 𝟎
(Existence of additive inverse).
6. To every real number 𝛼 and every element u in V
there is associated a unique element in V, called the
scalar product of u and 𝛼, which denoted by 𝛼𝒖
(closure under scalar multiplication).
7. (𝛼𝛽 )𝒖 = 𝛼(𝛽𝒖) (Associate property of scalar
multiplication).

- 94 -
8. (𝛼 + 𝛽 )𝒖 = 𝛼𝒖 + 𝛽𝒖 (Distributive property of scalar
multiplication).
9. 𝛼 (𝒖 + 𝒗) = 𝛼𝒖 + 𝛼𝒗 (Distributive property of scalar
multiplication).
10. 1. 𝒖 = 𝒖 (Scalar identity property).

By a scalar we mean a real number. If in the preceding

definition, we changed '' real numbers '' to '' complex
numbers "we would obtain a linear space over the
complex numbers. We shall restrict the discussion, in this
chapter, to linear spaces over the real numbers. A linear
space is also called a vector space. However, the
elements of a linear space may not be n–vectors.

Remark. It is important to realize that a vector space

consists of four entities:
1. A set V of vectors.
2. A set of scalars. In this class, it will always be the set
of real numbers ℝ. (Later on, this could be the set of
complex numbers ℂ.)
3. A vector addition denoted by " + ".
4. A scalar multiplication denoted by " ∙ ".

- 95 -
Example1.
By Theorem 1 of Section 1 in this chapter, ℝ𝑛 of all
ordered n-tuples of real numbers is a linear space over ℝ
for 𝑛 ≥ 1.◄
Remark.
We can express the elements of ℝ2 as vectors in the
plane. For example,𝒖 = (a1 , a1 ) ∈ ℝ2 is the vector line
with initial point(0, 0)and end point (a1 , a1 ) as shown in
the figure.

If 𝐚 = (a1 , a2 ) and 𝒃 = (𝑏1 , 𝑏2 ) ∈ ℝ2 , then

𝐚 + 𝒃 = (a1 + 𝑏1 , a2 + 𝑏2 ) is the vector with initial
(0, 0) and end (a1 + 𝑏1 , a2 + 𝑏2 ) which is as shown in
the following figure.

- 96 -
If 𝒗 = (𝑎, 𝑏) ∈ ℝ2 ,then 𝛼𝒗 = α(𝑎, 𝑏) = (𝛼𝑎, 𝛼𝑏) is the
vector with initial (0, 0) and end (𝛼𝑎, 𝛼𝑏) . If 𝛼 > 0, then
the direction of 𝛼𝒗 is the same direction of 𝒗. If 𝛼 < 0,
the direction of 𝛼𝒗 is opposite to the direction of 𝒗 as
shown in the following figure.

Example 2.
From Theorem 1 in Section 1 in Chapter I the set of all
𝑚 × 𝑛 matrices, with real entries, with addition and
scalar multiplication of matrices (𝑀𝑚×𝑛 , + , ∙) is a
linear space. ◄

- 97 -
Example 3.
Let V be the set of all 2- vectors whose components total
zero. That is 𝑉 = {(𝑥, 𝑦): 𝑥 + 𝑦 = 0}. The set V is
nonempty. Since (0, 0) is an element of V. The set V also
satisfied 2,3,7, 8,9 and 10 of Definition 1 in this section,
because all 2- vectors satisfies them and V is the set of 2-
vectors. We now verify that the remaining properties are
satisfied. Let 𝒖 = (𝑥1 , 𝑦1 ) and 𝒗 = (𝑥2 , 𝑦2 ) be any two
elements of V so that 𝑥1 + 𝑦1 = 0, 𝑥2 + 𝑦2 = 0. Then
𝒖 + 𝒗 = (𝑥1 + 𝑥2 , 𝑦1 + 𝑦2 ) and (𝑥1 + 𝑥2 ) +
(𝑦1 + 𝑦2 ) = (𝑥1 + 𝑦1 ) + (𝑥2 + 𝑦2 ) = 0 + 0 = 0. Thus
𝒖 + 𝒗 is an element of V, and property 1 is satisfied.
Since 𝟎 = (0, 0) is an element of V and 𝒖 + 𝟎 = 𝒖 for
every vector u in V, property 4 is satisfied. It is clear that
– 𝒖 = (−𝑥, −𝑦) is an additive inverse for 𝒖 = (𝑥, 𝑦)
since 𝒖 + (−𝒖) = (𝑥, 𝑦) + (−𝑥, −𝑦) = (0, 0) = 𝟎 .

Moreover, (−𝑥 ) + (−𝑦) = (−1)(𝑥 + 𝑦) = (−1)(0) = 0

so that −𝒖 is an element of V. Therefore property 5 is
satisfied. If 𝛼 is any number, then 𝛼𝒖 = 𝛼 (𝑥, 𝑦) =
(𝛼𝑥, 𝛼𝑦)and 𝛼𝑥 + 𝛼𝑦 = 𝛼 (𝑥 + 𝑦) = 𝛼. 0 = 0. Thus 𝛼𝒖

- 98 -
is an element of V, and property 6 is satisfied. We have
shown that V satisfied all the properties of Definition 1.
Therefore, V is a linear space. ◄
Example 4.
Let V be the set of all 2-vectors whose components total
one. That is 𝑉 = {(𝑥, 𝑦): 𝑥 + 𝑦 = 1} . Prove that 𝑉 is not
a linear space.
Solution:
Properties 1,4,5 and 6 of Definition 1 in this section are
not satisfied. In particular (0, 1) and (1, 0) are elements
of 𝑉, but (0, 1) + (1, 0) = (1, 1) is not. So, property 1 is
not satisfied. ◄
The important example of linear spaces is the set
of all polynomials. So, we remember the basic properties
of polynomials functions.
We define the polynomial function of degree n in the
variable 𝑥 as follows:
𝑃(𝑥 ) = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + ⋯ + 𝑎𝑛 𝑥 𝑛 ,
where 𝑎0 , 𝑎1 , 𝑎2 , … , 𝑎𝑛 ∈ ℝ are the real coefficients and
𝑎𝑛 ≠ 0. The zero polynomial is a polynomial with all

- 99 -
coefficients are zeros and its degree is undefined. Let P
be the set of all polynomials and
𝑝(𝑥 ) = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 + ⋯ and
𝑞 (𝑥 ) = 𝑏0 + 𝑏1 𝑥 + 𝑏2 𝑥 2 + ⋯
be two elements of P (may be with different degrees). We
say that 𝑝(𝑥 ) = 𝑞(𝑥) if and only if 𝑎𝑖 = 𝑏𝑖 for all i. We
define addition and scalar multiplication as follows.
𝑝(𝑥) + 𝑞(𝑥) = (𝑎0 + 𝑏0 ) + (𝑎1 + 𝑏1 )𝑥 + (𝑎2 + 𝑏2 )𝑥 2 + ⋯
𝛼𝑝(𝑥 ) = 𝛼𝑎0 + (𝛼𝑎1 )𝑥 + (𝛼𝑎2 )𝑥 2 + ⋯
It is obvious that if 𝑝(𝑥 ), 𝑞(𝑥) ∈ 𝑃 and 𝛼 ∈ ℝ , then
𝑝(𝑥 ) + 𝑞(𝑥) ∈ 𝑃 and 𝛼𝑝(𝑥) ∈ 𝑃. Also, It is easy to
verify that P satisfies all properties of Definition 1 above.
However, if 𝑉 denotes the set of all polynomial functions
with real coefficients of exactly degree 2. The operations
of addition and scalar multiplication are defined above.
Then 𝑉 is not a linear space, because the sum of two
second – degree polynomials need not be a second –
degree polynomial. For example, the sum of 𝑝(𝑥 ) = 𝑥 2
and 𝑞 (𝑥 ) = −𝑥 2 + 𝑥 is a first-degree polynomial.
Hence property 1 of Definition 1 above is not satisfied.
The polynomial that is zero for all x is not a second –
- 100 -
degree polynomial, so property 4 is not satisfied. There
are important subsets of P which is a linear space. For
example, if 𝑃𝑛 is the set of all polynomial functions of
degree less than or equal to n with the zero polynomial.
Then 𝑃𝑛 is a linear space for all 𝑛 ≥ 0.◄
Example 5.
Let 𝐹 [𝑎, 𝑏] be the set of all real valued functions defined
on closed interval [𝑎, 𝑏], i.e.,
𝐹 [𝑎, 𝑏] = {𝑓: 𝑓: [𝑎, 𝑏] → ℝ}
It is easy to verify that 𝐹 [𝑎, 𝑏] is a linear space, where
the addition and scalar multiplication is defined as
follows:
(𝑓 + 𝑔)(𝑥 ) = 𝑓 (𝑥 ) + 𝑔(𝑥 ) and (𝛼𝑓)(𝑥 ) = 𝛼 [𝑓(𝑥 )]for
all 𝑓, 𝑔 ∈ 𝐹 [𝑎, 𝑏] , 𝛼 ∈ ℝ 𝑎𝑛𝑑 𝑥 ∈ [𝑎, 𝑏].◄
Example 6.
Let 𝑉 be a set consists of only one element is 0. Then 𝑉 is
a linear space which is called a zero space and is denoted
by 0, where the addition and scalar multiplication is
defined as 𝟎 + 𝟎 = 𝟎, 𝛼𝟎 = 𝟎 for all 𝛼 ∈ ℝ.
All examples above on linear space are expected and
normal. Now, we introduce a stranger example.

- 101 -
Example 7.
Prove that the set 𝑉 = {𝑥 ∈ ℝ: 𝑥 > 0} with the addition
𝑥⨁𝑦 = 𝑥𝑦 and the scalar multiplication defined by
𝛼⨀𝑥 = 𝑥 𝛼 for all 𝑥, 𝑦 ∈ 𝑉 and α ∈ ℝ is a linear space.
Solution.
1- Since the product of two positive real is positive real
then 𝑥⨁𝑦 ∈ 𝑉.
2- 𝑥⨁𝑦 = 𝑥𝑦 = 𝑦𝑥 = 𝑦⨁𝑥.
3- 𝑥⨁(𝑦⨁𝑧) = 𝑥⨁(𝑦𝑧) = 𝑥(𝑦𝑧).
= (𝑥𝑦)𝑧 = (𝑥⨁𝑦)𝑧 = (𝑥⨁𝑦)⨁𝑧.
4- Since 1⨁𝑥 = 1𝑥 = 𝑥, then 1 is the additive identity.
1 1 1
5- Since 𝑥⨁ 𝑥 = 𝑥 𝑥 = 1 , then 𝑥 is the additive inverse
of 𝑥.
6- Since for all 𝑥 > 0 and α ∈ ℝ , we have 𝑥 𝛼 > 0 then
𝛼 ⊙ 𝑥 ∈ 𝑉.
7- 𝛼⨀(𝑥⨁𝑦) = 𝛼⨀(𝑥𝑦) = (𝑥𝑦)𝛼 = 𝑥 𝛼 𝑦 𝛼 =
𝑥 𝛼 ⨁𝑦 𝛼 = (𝛼⨀𝑥 )⨁(𝛼⨀𝑦).
8- (𝛼 + 𝛽)⨀𝑥 = 𝑥 𝛼+𝛽 = 𝑥 𝛼 𝑥 𝛽 = 𝑥 𝛼 ⨁𝑥 𝛽 =
(𝛼⨀𝑥 )⨁(𝛽⨀𝑥 ).
𝛼
9- (𝛼𝛽)⨀𝑥 = 𝑥 𝛼𝛽 = (𝑥 𝛽 ) = 𝛼⨀(𝛽⨀𝑥).
- 102 -
 10- 1⨀𝑥 = 𝑥1 = 𝑥.
Therefore, all properties of Definition 1 above are
satisfied and 𝑉 is a linear space. ◄
Example 8.
Prove that 𝑉 = ℝ2 with addition defined as
(𝑎, 𝑏) + (𝑐, 𝑑) = (𝑎 + 𝑐, 𝑏 + 𝑑) and scalar
multiplication defined as 𝛼 (𝑎, 𝑏) = (𝛼𝑎, 𝑏) is not a linear
space.
Solution.
Property 8 (for example) of Definition 1 above is not
satisfied because
𝛼 (𝑎, 𝑏) + 𝛽 (𝑎, 𝑏) = (𝛼𝑎, 𝑏) + (𝛽𝑎, 𝑏) = ((𝛼 + 𝛽)𝑎, 2𝑏)
while (𝛼 + 𝛽 )(𝑎, 𝑏) = ((𝛼 + 𝛽)𝑎, 𝑏) . Then 𝛼 (𝑎, 𝑏) +
𝛽 (𝑎, 𝑏) ≠ (𝛼 + 𝛽 )(𝑎, 𝑏).
Therefore 𝑉 is not a linear space. ◄
We conclude this section with a theorem that lists several
useful properties of linear spaces.

- 103 -
Theorem 1.
Let 𝑉 be a linear space, 𝒖, 𝒗, 𝒘 be elements of 𝑉 and 𝛼 is
a scalar. Then
1-There is precisely one zero element of 𝑉.
2- If 𝒖 + 𝒗 = 𝒖, then 𝒗 = 𝟎.
3-If 𝒖 + 𝒗 = 𝒖 + 𝒘, then 𝒗 = 𝒘.
4-There is exactly one element w of such that 𝒖 + 𝒘 = 𝟎.
5-0𝒗 = 𝟎.
6-𝛼𝟎 = 𝟎
7-If 𝛼𝒗 = 𝟎, then 𝛼 = 0 or 𝒗 = 𝟎.
8-(−𝛼 )𝒗 = −(𝛼𝒗) = 𝛼 (−𝒗)
9-(−1)𝒖 = −𝒖
Proof.
1-Assume that there are two elements 𝟎 and 𝟎′ such that
𝒖 + 𝟎 = 𝒖 and 𝒖 + 𝟎′ = 𝒖 for every u in V. Setting 𝒖 =
𝟎′ in the first identity and 𝒖 = 𝟎 in the second, we obtain
𝟎′ = 𝟎′ + 𝟎 = 𝟎 + 𝟎′ = 𝟎.
2-Suppose that 𝒖 + 𝒗 = 𝒖. By Property 5 of Definition 1
in this section, there is an inverse – 𝒖 for 𝒖. Adding – 𝒖
to each side of 𝒖 + 𝒗 = 𝒖 we have(−𝒖) + (𝒖 + 𝒗) =

- 104 -
(−𝒖) + 𝒖, or [(−𝒖) + 𝒖] + 𝒗 = (−𝒖) + 𝒖, or 𝟎 + 𝒗 =
𝟎, or 𝒗 = 𝟎.
3-𝒖 + 𝒗 = 𝒖 + 𝒘 ⇒
(−𝒖) + (𝒖 + 𝒗) = (−𝒖) + (𝒖 + 𝒘) ⇒
(−𝒖 + 𝒖) + 𝒗 = (−𝒖 + 𝒖) + 𝒘 ⇒
𝟎 + 𝒗 = 𝟎 + 𝒘 ⇒ 𝒗 = 𝒘.
4- Let 𝒘 and 𝒘′ be such that 𝒖 + 𝒘 = 𝟎 and 𝒖 + 𝒘′ = 𝟎
𝒘 = 𝒘 + 𝟎 = 𝒘 + (𝒖 + 𝒘′) = (𝒘 + 𝒖) + 𝒘′
= (𝒖 + 𝒘) + 𝒘′ = 𝟎 + 𝒘′ = 𝒘′.
5- Since 0𝒗 = (0 + 0)𝒗 = 0𝒗 + 0𝒗, then by adding
(−0𝒗) to both sides we have
0𝒗 + (−0𝒗) = (0𝒗 + 0𝒗) + (−0𝒗)
or
0𝒗 + (−0𝒗) = 0𝒗 + (0𝒗 + (−0𝒗))
or
𝟎 = 0𝒗 + 𝟎 = 0𝒗
6- Since 𝛼𝟎 = 𝛼 (𝟎 + 𝟎) = 𝛼𝟎 + 𝛼𝟎, then by adding
(−𝛼𝟎) to both sides we have
𝛼𝟎 + (−𝛼𝟎) = (𝛼𝟎 + 𝛼𝟎) + (−𝛼𝟎), or
𝟎 = 𝛼𝟎 + (𝛼𝟎 + (−𝛼𝟎)) = 𝛼𝟎 + 𝟎. Hence 𝛼𝟎 = 𝟎.

- 105 -
 1 1
7- Let 𝛼𝒗 = 𝟎 and 𝛼 ≠ 0 . Then 𝛼𝒗 = 𝟎.
𝛼 𝛼

Therefore,1. 𝒗 = 𝟎. So, 𝒗 = 𝟎.
8- Note that (−𝛼 )𝒗 + (𝛼𝒗) = (−𝛼 + 𝛼 )𝒗 = 0𝒗 = 𝟎.
By adding (−𝛼𝒗) or both sides we here (−𝛼 )𝒗 =
−(𝛼𝒗) . Similarly,𝛼 (−𝒗) = −𝛼𝒗.
9- Let 𝛼 = 1 in (8). ◄

- 106 -
 Exercise Set (4.2)
In Exercise (1) – (20) determine whether the given set
with the two operations are linear space.
1. 𝑉 = ℝ2 , (𝑥1 , 𝑥2 ) + (𝑦1 , 𝑦2 ) = (𝑥2 + 𝑦2 , 𝑥1 +
𝑦1 ), 𝛼 (𝑥1 , 𝑥2 ) = (𝛼𝑥1 , 𝛼𝑥2 ).
2. 𝑉 = {(𝑥1 , 0, 𝑥3 ) ∶ 𝑥1 , 𝑥3 ∈ ℝ},
(𝑥1 , 0, 𝑥3 ) + (𝑦1 , 0, 𝑦3 )
= (𝑥1 + 𝑦1 , 0, 𝑥3 + 𝑦3 ), 𝛼 (𝑥1 , 0, 𝑥3 )
= (𝛼𝑥1 , 0, 𝑥3 )
0 𝑏
3.𝑉 = {[ ] ∶ 𝑎, 𝑏, 𝑐 ∈ ℝ} with summation and scalar
𝑎 𝑐
multiplication of matrices.
𝑎 𝑏
4. 𝑉 = {[ ] ∶ 𝑎 ≠ 0} with summation and scalar
𝑐 𝑑
multiplication of matrices.
5.𝑉 = {𝐴: 𝐴 ∈ 𝑀2×3 , ∑ 𝑎𝑖𝑗 = 0} with summation and
scalar multiplication of matrices.
6.𝑉 = {𝑃 ∈ 𝑃𝑛 ∶ 𝑃(1) + 𝑃(6) − 𝑃(2) = 0} with
summation andscalar multiplication of polynomials.
7. 𝑉 = {𝑃 ∈ 𝑃𝑛 ∶ 𝑃(1) + 𝑃(6) − 𝑃(2) = 1} with
summation andscalarmultiplication of polynomials.

- 107 -
8. 𝑉 = ℝ𝟑 , (𝑥1 , 𝑦1 , 𝑧1 ) + (𝑥2 , 𝑦2 , 𝑧2 ) = (𝑥1 + 𝑥2 , 𝑦1 +
𝑦2 , 𝑧1 + 𝑧2 ), 𝛼 (𝑥, 𝑦, 𝑧) = (0,0,0)).
9.𝑉 = ℝ2 , (𝑥1 , 𝑦1 ) + (𝑥2 , 𝑦2 ) = (𝑥1 + 𝑥2 , 𝑦1 + 𝑦2 ),
(𝑥, 𝑦) = (2𝑥, 2𝑦).
10.𝑉 = {(𝑥, 0): 𝑥 ∈ ℝ} with the usual addition and scalar
multiplication onℝ2 .
11. 𝑉 = {(𝑥, 0): 𝑥 ≥ 0} with the usual addition and scalar
multiplication onℝ2 .
12. 𝑉 = ℝ2 , (𝑥, 𝑦) + (𝑥1 , 𝑦1 ) = (𝑥 + 𝑥1 + 1, 𝑦 + 𝑦1 +
1), (𝑥, 𝑦) = (𝑥, 𝑦).
𝑎 1
13.𝑉 = {[ ] : 𝑎, 𝑏 ∈ ℝ} with summation and scalar
1 𝑏
multiplication of matrices.
14.𝑉 = {𝑓 ∈ F[0,2]: 𝑓(1) = 0}with addition and scalar
multiplication in defined in Example 5 Section 2 above.
15.𝑉 = {(𝑎, 𝑎, −𝑎): 𝑎 ∈ ℝ} with usual addition and
scalar multiplication on ℝ𝟑 .
16. 𝑉 = {(𝑎, 1, −𝑎): 𝑎 ∈ ℝ} with usual addition and
scalar multiplication on ℝ𝟑 .
17.𝑉 = {𝑎0 + 𝑎1 (𝑥 + 1): 𝑎0 , 𝑎1 ∈ ℝ} with addition and
scalar multiplication of polynomials.

- 108 -
18. 𝑉 = {𝑎0 + 𝑥 + 1: 𝑎0 ∈ ℝ}with addition and scalar
multiplication of polynomials.
19. 𝑉 = {𝐴 ∈ 𝑀3×3 : 𝐴 = 𝐴𝑡 }with addition and scalar
multiplication of matrices.
20. 𝑉 = {(𝑎, 𝑏, 𝑐 ): ∈ ℝ3 }with addition and scalar
multiplication on ℝ𝟑 .
𝑎1 𝑎2 𝑎3
21. Find c such that 𝑉 = {[𝑏 𝑏2 𝑏3 ] : 𝑎1 + 𝑎2 +
1

𝑎3 = 𝑐} with addition and scalar multiplication of

matrices is a linear space.

22. Let 𝑉 be a linear space and 𝒗1 , 𝒗2 , … , 𝒗𝑛 ∈ 𝑉, 𝛼 ∈ ℝ.
Prove that𝛼 (𝒗1 , … , 𝒗𝑛 ) = 𝛼𝒗1 + ⋯ + 𝛼𝒗𝑛 .

- 109 -
 3. Subspaces
It is possible for one linear space to be contained within a
larger space. For example, planes through the origin are
linear spaces that are contained within the larger linear
space ℝ𝟑 . In this section we shall study this important
concept in more detail.
Definition 1.
A subset W of a linear space V is called a subspace of V if
W is itself a linear space under the addition and scalar
multiplication defined on V.
In general, one must verify the ten linear space axioms to
show that a set W with addition and scalar multiplication
forms a linear space. However, if W is part of a larger set
V that is already known to be a linear space, then certain
axioms needn’t be verified for W because they are
inherited from V. For example, there is no need to check
that 𝒖 + 𝒗 = 𝒗 + 𝒖 (Axiom 2 ) for W because this holds
for all vectors in V and consequently for all vectors in W.
other axioms inherited by W from V are 3, 7, 8, 9 and 10.
Thus, to show that a set W is a subspace of a linear space
V, we need only verify Axioms 1,4,5 and 6. The

- 110 -
following theorem shows that even Axioms 4 and 5 can
be dispensed with.
Theorem 1.
If W is a set of one or more vectors from a linear space V,
then W is a subspace of V if and only if the following
conditions hold:
(a) If u and v are vectors in W, then 𝒖 + 𝒗 is in W.
(b) If k is any scalar and u is any vector in W, then ku is
in W.
Proof.
If W is a subspace of V, then all the linear space axioms
are satisfied; in particular, Axioms 1 and 6 hold but these
are Precisely conditions (a) and (b).
Conversely, assume conditions (a) and (b) hold since
these conditions are linear space axioms1 and 6, we need
only show that W satisfies the remaining eight axioms.
Axioms 2,3,7,8,9, and 10 are automatically satisfied by
the vectors in W since they are satisfied by all vectors in
V. Therefore, to complete the proof, we need only verify
that Axioms 4 and 5 are satisfied by vectors in W. Let u
be any vector in W. By condition (b), ku is in W for every

- 111 -
scalar k. Setting 𝑘 = 0, it follows from Theorem 1 in
Section 2 of of this chapter that 0𝒖 = 𝟎 is in W, and
setting 𝑘 = −1, it follows that (−1)𝒖 = −𝒖 is in W.◄
Remark.
A set W of one or more vectors from a linear space V is
said to be closed under a addition if condition (a) in
Theorem 1 above holds and closed under scalar
multiplication if condition (b) holds. Thus, Theorem 1
above states that W is a subspace of V if and only if W is
closed under addition and under scalar multiplication.
Note.
We can use one condition equivalent as both conditions
(a) and (b) in Theorem 1 above as follows:
W is a subspace of V if and only if 𝛼𝒖 + 𝛽𝒗 ∈ 𝑾 for all
𝒖, 𝒗 ∈ 𝑊and all 𝛼, 𝛽 ∈ ℝ (why?).
Example 1.
Every nonzero linear space V has at least two subspaces:
V itself is a subspace, and the set {𝟎} consisting of just
the zero in V is a subspace called the zero subspace. ◄

- 112 -
Example 2.
Let 𝑋 = {(𝑥, 0): 𝑥 ∈ ℝ} ⊆ ℝ2 . Prove that X is a subspace
of ℝ2 .
Solution.
(1) (0,0) ∈ 𝑋, 𝑖. 𝑒, 𝑋 ≠ 𝜙.
(2) Let (𝑥1 , 0), (𝑥2 , 0) ∈ 𝑋. Then
(𝑥1 , 0) + (𝑥2 , 0) = (𝑥1 + 𝑥2 , 0) ∈ 𝑋
(3) 𝛼 (𝑥, 0) = (𝛼𝑥, 0) ∈ 𝑋 for all (𝑥, 0) ∈ 𝑋 and all
α ∈ ℝ. From (1) , (2) and (3), X is a subspace of ℝ2 .
Similarly
𝑌 = {(0, 𝑦): 𝑦 ∈ ℝ} is a subspace of ℝ2 .◄
Example 3.
Let 𝑊 = {(𝑥, 𝑦) ∈ ℝ2 : 𝑎𝑥 + 𝑏𝑦 = 0} ⊆ ℝ2 . Prove that
W is a subspace of ℝ2 .
Solution.
(1) Since 𝑎. 0 + 𝑏. 0 = 0, then (0,0) ∈ 𝑊and𝑊 ≠ ϕ.
(2) Let (𝑥, 𝑦), (𝑥1 , 𝑦1 ) ∈ 𝑊. Then 𝑎𝑥 + 𝑏𝑦 = 0 and
𝑎𝑥1 + 𝑏𝑦1 = 0. Hence 𝑎(𝑥 + 𝑥1 ) + 𝑏(𝑦 + 𝑦1 ) =
(𝑎𝑥 + 𝑏𝑦) + (𝑎𝑥1 + 𝑏𝑦1 ) = 0 + 0 = 0.
Therefore (𝑥, 𝑦) + (𝑥1 , 𝑦1 ) = (𝑥 + 𝑥1 , 𝑦 + 𝑦1 ) ∈ 𝑊.

- 113 -
(3) Let (𝑥, 𝑦) ∈ 𝑊 and 𝛼 ∈ ℝ . Then 𝑎𝑥 + 𝑏𝑦 = 0.
Therefore 𝛼 (𝑎𝑥 + 𝑏𝑦) = 0. Thus 𝛼 (𝑥, 𝑦) ∈ 𝑊. From (1),
(2) and (3), we have W is a subspace of ℝ2 . Note that the
subspace in Example 3 above of ℝ2 is the lines through
the origin. So, the subspaces of ℝ2 are {𝟎}, lines through
the origin and ℝ2 itself. Later, we will show that these are
the only subspaces of ℝ2 .◄
Example 4.
Let 𝑊 = {(𝑥, 𝑦, 𝑧) ∈ ℝ3 : 𝑎𝑥 + 𝑏𝑦 + 𝑐𝑧 = 0}. Prove W is
a subspace of ℝ3 .
Solution.
(1) Since 𝑎0 + 𝑏0 + 𝑐0 = 0, (0,0,0) ∈ 𝑊 and 𝑊 ≠ ∅.
(2) Let (𝑥, 𝑦, 𝑧), (𝑥1 , 𝑦1 , 𝑧1 ) ∈ 𝑊. Then 𝑎𝑥 + 𝑏𝑦 + 𝑐𝑧 =
0and 𝑎𝑥1 + 𝑏𝑦1 + 𝑐𝑧1 = 0. Therefore,
𝑎(𝑥 + 𝑥1 ) + 𝑏(𝑦 + 𝑦1 ) + 𝑐 (𝑧 + 𝑧1 ) = (𝑎𝑥 + 𝑏𝑦 + 𝑐𝑧) +
(𝑎𝑥1 + 𝑏𝑦1 + 𝑐𝑧1 ) = 0.
Thus (𝑥, 𝑦, 𝑧) + (𝑥1 , 𝑦1 , 𝑧1 ) ∈ 𝑊.
(3) Let (𝑥, 𝑦, 𝑧) ∈ 𝑊 and α ∈ ℝ. Then 𝑎𝑥 + 𝑏𝑦 + 𝑐𝑧 =
0. Therefore 𝛼 (𝑎𝑥 + 𝑏𝑦 + 𝑐𝑧) = 0. Thus 𝛼 (𝑥, 𝑦, 𝑧) ∈ 𝑊.
Hence W is a subspace of ℝ3 . Note that the subspace W is
the planes through the origin. Later, we will show that the
- 114 -
only subspaces of ℝ3 are {𝟎}, lines through the origin,
planes through the origin, and ℝ3 itself. ◄
Example 5.
𝑎 𝑏
Let 𝑊 = {[ ] ∶ 𝑎, 𝑏 ∈ ℝ} ⊆ 𝑀2×2 . Prove that W is
0 −𝑎
a subspace of 𝑀2×2 .
Solution.
0
0
[ ] ∈ 𝑊 and W ≠ ϕ.
0
0
𝑎 𝑏 𝑐
𝑑
(2)Let [ ], [ ] ∈ 𝑊. Then
0 −𝑎 0
−𝑐
𝑎 𝑏 𝑐 𝑑 𝑎+𝑐 𝑏+𝑑
[ ]+[ ]=[ ] ∈ 𝑊.
0 −𝑎 0 −𝑐 0 −(𝑎 + 𝑐)
𝑎 𝑏
(3) Let [ ] ∈ 𝑊 and α ∈ ℝ.
0 −𝑎
𝑎 𝑏 𝛼𝑎 𝛼𝑏
Then 𝛼 [ ]=[ ] ∈ 𝑊.
0 −𝑎 0 −𝛼𝑎
Therefore, W is a subspace of 𝑀2×2 .
Example 6.
Let P be the set of all polynomials and
𝑊 = {𝑝(𝑥 ) ∈ 𝑃: 𝑝(3) = 0} ⊆ 𝑃. Prove that W is a
subspace of P.
Solution.
(1) Since 𝟎(3) = 0, then 𝟎 ∈ 𝑊.
- 115 -
(2) Let 𝑝(𝑥 ), 𝑞 (𝑥 ) ∈ 𝑊. then 𝑝(3) = 𝑞 (3) = 0.
Therefore (𝑝 + 𝑞 )(3) = 𝑝(3) + 𝑞 (3) = 0 + 0 = 0. Thus
𝑝(𝑥 ) + 𝑞(𝑥) ∈ 𝑊.
(3) Let 𝑝(𝑥 ) ∈ 𝑊 and α ∈ ℝ. Then 𝑝(3) = 0 and
α𝑝(3) = 0. Hence 𝛼𝑝(𝑥 ) ∈ 𝑊.◄
Example 7.
The subset 𝑊 = {(𝑥, 𝑥 + 1): 𝑥 ∈ ℝ} is not a subspace of
ℝ2 since (0,0) ∉ 𝑊.
Example 8.
The subset 𝑊 = {𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 : 𝑎0 is integer} is not
1
a subspace of 𝑃2 , since, for example, 2 + 3𝑥 + 𝑥 2 ∈ 𝑊
3
1 1 1 2 1
and ∈ ℝ but (2 + 3𝑥 + 𝑥 2 ) = + 𝑥 + 𝑥 2 ∉ 𝑊.◄
3 3 3 3 9

Example 9.
The subset 𝑊 = {(𝑎, 𝑏) ∈ ℝ2 : 𝑎 ≠ 0, 𝑏 ≠ 0}is not a
subspace of ℝ2 , because, for example, (2,3), (−2,5) ∈ 𝑊
but (2,3) + (−2,5) = (0,8) ∉ 𝑊.
Definition 2.
If 𝐴𝒙 = 𝑏 is a system of linear equations, then each
vector 𝒙 that satisfies this equation is a solution vector of
the system. The following theorem shows that the

- 116 -
solution vectors of a homogeneous linear system form a
linear space, which we shall call the solution space of
the system.
Theorem 2.
If 𝐴𝒙 = 𝟎 is a homogeneous linear system of m
equations in n unknowns, then the set of solution vectors
is a subspace of ℝ𝑛 .
Proof.
Let W be the solution vectors set. There is at least one
vector in W, namely 0. To show that W is closed under
addition and scalar multiplication, we must show that if
𝒙 and 𝒙′ are any solution vectors and k is any scalar, then
𝒙 + 𝒙′ and 𝑘𝒙 are also solution vectors. But if 𝒙 and 𝒙′
are solution vectors, then 𝐴𝒙 = 𝟎 and 𝐴𝒙′ = 𝟎. It
follows that 𝐴(𝒙 + 𝒙′) = 𝐴𝒙 + 𝐴𝒙′ = 𝟎 + 𝟎 = 𝟎 and
𝐴(𝑘𝒙) = 𝑘(𝐴𝒙) = 𝑘𝟎 = 𝟎. This proves that 𝒙 +
𝒙′ and 𝑘𝒙 are solution vectors. Therefore, W is a
subspaceℝ𝑛 .◄

- 117 -
 Exercise set (4.3)
In Exercise (1) – (19) determine which of the following
subsets are subspaces of the giver linear spaces.
1.𝑊 = {𝑓: [a, b] ⟶ ℝ, f is diffrentiable on [a, b]} ⊆
𝐹 [𝑎, 𝑏]
2.W= {𝑓: [𝑎, 𝑏] → ℝ, 𝑓 is diffrentiable twice on [a, b]}
⊆ 𝐹 [𝑎, 𝑏] such that 𝑓 ′′ + 𝑓 = 0}
3.𝑊 = {𝑎, 𝑏, 𝑎 − 2𝑏: 𝑎, 𝑏 ∈ ℝ} ⊆ ℝ3 .
𝑎 0
4. 𝑊 = {[ ] , 𝑎, 𝑏, 𝑐 ∈ ℝ} ⊆ 𝑀2×2 .
𝑏 𝑐
5.𝑊 = {𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 ∈ 𝑃2 : 𝑎0 + 𝑎1 + 𝑎2 = 0} ⊆ 𝑃2 .
6.𝑊 = {𝑓 ∈ F[0,2]: 𝑓(1) = 0} ⊆ 𝐹 [0,2].
7.𝑊 = {(𝑥, 𝑦) ∈ ℝ2 : 2𝑥 + 3𝑦 = 0} ⊆ ℝ2 .
8.𝑊 = {𝐴 ∈ 𝑀2×2 : det 𝐴 = 0} ⊆ 𝑀2×2 .
(9)𝑊 = {𝐴 ∈ 𝑀2×2 : 𝐴2 = 𝐴} ⊆ 𝑀2×2 .
10.𝑊 = {(𝑥, 𝑦) ∈ ℝ2 : 𝑥 + 𝑦 = 1} ⊆ ℝ2 .
11.𝑊 = {𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 ∈ 𝑃2 : 𝑎1 = 1 𝑎0 = 𝑎2 } ⊆ 𝑃2 .
12.𝑊 = {(𝑎, 𝑏, 𝑐 ) ∈ ℝ3 : 𝑎 + 𝑏 + 𝑐 = 0} ⊆ ℝ3 .
13.𝑊 = {(𝑎, 𝑏, 𝑐 ) ∈ ℝ3 : 𝑎 − 𝑏 − 𝑐 = 0} ⊆ ℝ3 .
(14)𝑊 = {(𝑎, 𝑏, 𝑐) ∈ ℝ3 : 𝑎 + 𝑏 = 1, 𝑐 = 0} ⊆ ℝ3 .
15.𝑊 = {(𝑎, 𝑏, 𝑐 ) ⊂ ℝ3 : 𝑎𝑏 = 0} ⊆ ℝ3 .

- 118 -
 3 𝑎 3
16.𝑊 = {[ ] : 𝑎, 𝑏 ∈ ℝ} ⊆ 𝑀2×3 .
𝑎 𝑏 𝑎
17.𝑊 = {𝐴 ∈ 𝑀2×2 : 𝐴 = −𝐴𝑡 } ⊆ 𝑀2×2 .
1
18.𝑊 = {𝑓 ∈ F[0,1] : ∫0 𝑓(x)dx = 0} ⊆ 𝐹 [0,1].

4. Linear combination and spanning sets

Definition 1.
Let 𝑉 be a linear space and 𝒗1 , 𝒗2 , … , 𝒗𝑛 ∈ 𝑉.An
element w of 𝑉 is called a linear combination of
𝒗1 , 𝒗2 , … , 𝒗𝑛 if there are scalars 𝛼1 , 𝛼2 , … , 𝛼𝑛 ∈ ℝsuch
that 𝒘 = 𝛼1 𝒗1 + 𝛼2 𝒗2 + ⋯ + 𝛼𝑛 𝒗𝑛 .
Example 1.
Prove that the vector 𝒗 = (7, −2, 2) ∈ ℝ3 is a linear
combination of the vectors 𝒗1 = (1, −1,0), 𝒗2 =
(0,1,1), 𝒗3 = (2,0,1).
Solution.
We want to find 𝛼1 , 𝛼2 , 𝛼3 ∈ ℝ such that
(7, −2,2) = 𝛼1 (1, −1,0) + 𝛼2 (0,1,1) + 𝛼3 (2,0,1).Then
𝛼1 + 2𝛼3 = 7,−𝛼1 + 𝛼2 = −2, 𝛼2 + 𝛼3 = 2. By solving
this system we obtain
𝛼1 = 1, 𝛼2 = −1, 𝛼3 = 3
- 119 -
Hence (7, −2,2) = (1, −1,0) − (0,1,1) + 3(2,0,1).◄
Example 2.
Determine whether 𝒗 = (3, −1,4) ∈ ℝ3 is a linear
combination of𝒗1 = (1, −1,0), 𝒗2 = (0,1,1), 𝒗3 =
(3, −5,2), or not?
Solution.
Let 𝛼1 , 𝛼2 , 𝛼3 ∈ ℝ such that 𝒗 = 𝛼1 𝒗1 + 𝛼2 𝒗2 + 𝛼3 𝒗3
Therefore,
(3, −1,4) = 𝛼1 (1, −1,0) + 𝛼2 (0,1,1) + 𝛼3 (3, −5, −2)
So, we have the following system of equation
𝛼1 + 3𝛼3 = 3
−𝛼1 + 𝛼2 − 5𝛼3 = −1
𝛼2 − 2𝛼3 = 4
By using Gauss method, we find the row–echelon form of
the argument matrix is :
1 0 3 3
[0 1 −2| 2]
0 0 0 1
Hence the system is inconsistent. Therefore𝒗 is not a
linear combination of 𝒗1 , 𝒗2 , 𝒗3 .◄
Example 3.
Consider 𝑓(𝑥 ) = 𝑥 2 + 2, 𝑔(𝑥 ) = 𝑥 + 2,
- 120 -
ℎ(𝑥 ) = 3𝑥 2 + 2𝑥 − 2 and 𝑘 (𝑥 ) = 𝑥 3 in the linear space
of all functions defined on the interval
(−∞, ∞), . 𝑖. 𝑒. , 𝐹 (−∞, ∞). The function h is a linear
combination of f and g since ℎ(𝑥 ) = 3𝑓(𝑥 ) −
𝑔(𝑥)however 𝑘(𝑥 ) = 𝑥 3 is not a linear combination of f
and g.
This may be proved as follows.
Suppose that k is a linear combination of f and g. Then
there exists 𝛼1 , 𝛼2 ∈ ℝ such that 𝑘(𝑥 ) = 𝛼1 𝑓(𝑥 ) +
𝛼2 𝑔(𝑥) .Therefore𝑥 3 = 𝛼1 (𝑥 2 + 𝑥 ) + 𝛼2 (𝑥 + 2).
compare the coefficients inboth sides we have 0 = 1,
which is impossible hence 𝑘(𝑥 )is not a linear
combination of f and g.◄
Example 4.
Using the triangular equalitycos(3 + x) = cos 3 cos x −
sin 3 sin x,we find that cos(3 + x) is a linear combination
of sin x and cos x . However, sin2 𝑥is not a linear
combination of sin x and cos x . Indeed, Let sin2 𝑥 is a
linear combination of sin x and cos x . Then sin2 𝑥 =
𝛼1 sinx + 𝛼2 cos x,where 𝛼1 , 𝛼2 ∈ ℝ for every x ∈ ℝ.In
particular, sin2 0 = 𝛼1 sin 0 + 𝛼2 cos 0, 𝑖. 𝑒, 0 = 𝛼1 0 +
- 121 -
𝛼2 1, hence 𝛼2 = 0. Therefore sin2 x = 𝛼1 sinx.Ifsin x ≠
0, then sin x = α1 forevery𝑥 ∈ ℝ.This is impossible.
Therefore sin2 𝑥 is not a linear combination of sinx
andcos x. ◄
In Examples 1-3 above we need to solve system of
equations to show that a vector is a linear combination of
a given vectors. In special case that the linear space is a
subspace of ℝ𝑛 thereis a mechanical procedure for
discovering whether an n-vector b is a linear combination
of n- vectors𝒗1 , 𝒗2 , . . 𝒗𝑛 . In fact, we will show that b is a
linear combination of 𝒗1 , 𝒗2 , . . 𝒗𝑛 if and only if the
equation 𝐴𝒙 = 𝑏 is consistent, where A is the
matrix having 𝒗1 , 𝒗2 , . . 𝒗𝑛 as its columns. In order to
prove this, we first prove a theorem that relates the
conceptof linear combinations to matrix multiplications.
Theorem 1.
Let 𝐴 = [𝑎𝑖𝑗 ] be an 𝑚 × 𝑛 matrix and 𝑥 =
[𝑥1 , 𝑥2 , … , 𝑥𝑛 ]𝑡 . If 𝐴1 , 𝐴2 , … , 𝐴𝑛 are the columns of A,
then 𝐴𝑥 = 𝑥1 𝐴1 + 𝑥2 𝐴2 + ⋯ + 𝑥𝑛 𝐴𝑛
Proof.

- 122 -
 𝑎11 𝑎12 …𝑎1𝑛 𝑥1
𝑎21 𝑎22 …𝑎2𝑛 𝑥2
𝐴𝑥 = [ ⋮ ⋮ ⋮ ⋮ ][ ⋮ ]
𝑎𝑚1 𝑎𝑚2 …𝑎𝑚𝑛 𝑥𝑛
𝑎11 𝑥1 + 𝑎12 𝑥2 + ⋯ + 𝑎1𝑛 𝑥𝑛
𝑎 𝑥 + 𝑎22 𝑥2 + ⋯ + 𝑎2𝑛 𝑥𝑛
= [ 21 1 ⋮ ]
𝑎𝑚1 𝑥1 + 𝑎𝑚2 𝑥2 + ⋯ + 𝑎𝑚𝑛 𝑥𝑛
= 𝑥1 𝐴1 + 𝑥2 𝐴2 + ⋯ + 𝑥𝑛 𝐴𝑛 . ◄
Corollary1.
Let A be an 𝑚 × 𝑛 matrix the system 𝐴𝒙 = 𝑏 is
consistent if and only if b is a linear combination of
columns of A.
Example 5.
Find all vectors 𝒃 = [𝑏1 𝑏2 𝑏3 𝑏4 ]𝑡 such that b is a linear
combination of 𝒗1 = [1 2 3 4]𝑡
and𝒗2 = [4 3 2 1]t .
Solution.
Let = 𝛼1 𝒗1 + 𝛼2 𝒗2 , 𝛼1 , 𝛼2 ∈ ℝ . Then𝐴𝒙 = 𝒃, where
1 4
𝐴 = [2 3] , 𝑥 = [𝛼1 ]. From Corollary 1above, it is
3 2 𝛼2
4 1

- 123 -
sufficient to find 𝑏1 , 𝑏2 , 𝑏3 , 𝑏4 such that𝐴𝒙 = 𝒃 is
consistent. Therefore
1 4 𝑏1 1 4 𝑏1
[𝐴 | 𝑏 ] = [ 2 3] 𝑏2 ] ~ [0 −5] 𝑏2 − 2𝑏1
]
3 2 𝑏3 0 0 𝑏3 + 𝑏1 − 2𝑏2
4 1 𝑏4 0 0 𝑏4 + 2𝑏1 − 3𝑏2
This system is consistent if 𝑏3 + 𝑏1 − 2𝑏2 = 0 and
𝑏4 + 2𝑏1 − 3𝑏2 = 0 . Hence
𝒃 = [𝑏1 𝑏2 2𝑏2 − 𝑏1 3𝑏2 − 2𝑏1 ]𝑡 . ◄
If𝒗1 , 𝒗2 , … , 𝒗𝑟 are in vector space V, then generally some
vectors in V may be a linear combinations of
𝒗1 , 𝒗2 , … , 𝒗𝑟 and others may not. The following theorem
shows that if we construct a set W consisting of all those
vectors that are expressible as linear combinations
of𝒗1 , 𝒗2 , … , 𝒗𝑟 , then W forms a subspace of V.
Theorem 2.
If 𝒗1 , 𝒗2 , … , 𝒗𝑟 are vectors in a vector space V, then
(a)The set W of all linear combinations of 𝒗1 , 𝒗2 , … , 𝒗𝑟 is
a subspace of V.
(b)W is the smallest subspace of V that
contains𝒗1 , 𝒗2 , … , 𝒗𝑟 . In the since that every other
subspace of 𝒗1 , 𝒗2 , … , 𝒗𝑟 must contain W.
- 124 -
Proof.
(a)To show that W is a subspace of V, we prove it is
closed under addition and scalar multiplication.
There is at least one vector in W, namely, 0, since
𝟎 = 0𝒗1 + 0𝒗2 + ⋯ + 0𝒗𝑟 . If u and v are vectors in W,
then𝒖 = 𝑐1 𝒗1 + 𝑐2 𝒗2 + ⋯ + 𝑐𝑟 𝒗𝑟 and 𝒗 = 𝑘1 𝒗1 +
𝑘2 𝒗2 + ⋯ + 𝑘𝑟 𝒗𝑟 ,where 𝑐1 , 𝑐2 , … , 𝑐𝑟 , 𝑘1 , 𝑘2 , … , 𝑘𝑟 are
scalars. Therefore
𝒖 + 𝒗 = (𝑐1 + 𝑘1 )𝒗1 + (𝑐2 + 𝑘2 )𝒗2 + ⋯ + (𝑐𝑟 + 𝑘𝑟 )𝒗𝑟
and, for any scalar k,𝑘𝑢 = (𝑘𝑐1 )𝒗1 + (𝑘𝑐2 )𝒗2 + ⋯ +
(𝑘𝑐𝑟 )𝒗𝑟 . Thus, 𝒖 + 𝒗 and k u are linear combinations of
𝒗1 , 𝒗2 , … , 𝒗𝑟 and consequently elements in W. Therefore,
W is closed under addition and scalar multiplication.
(b) Each vector 𝒗𝑖 is a linear combination of
𝒗1 , 𝒗2 , … , 𝒗𝑟 since we can write
𝒗𝑖 = 0𝒗1 + 0𝒗2 + ⋯ + 1𝒗𝑖 + ⋯ + 0𝒗𝑟
Therefore, the subspace W contains each of the vectors
𝒗1 , 𝒗2 , … , 𝒗𝑟 . LetW' be any other subspace that contains
𝒗1 , 𝒗2 , … , 𝒗𝑟 . SinceW' is closed under addition and scalar
multiplication, it must contain all linear combinations of
𝒗1 , 𝒗2 , … , 𝒗𝑟 . Thus, W' contains each vector of W. ◄
- 125 -
We have the following definition.
Definition 2.
If 𝑆 = {𝒗1 , 𝒗2 , … , 𝒗𝑟 } is a set of vectors in a vector space
V, then the subspace W of V consisting of all linear
combinations of vectors of S is called the space spanned
by the vectors 𝒗1 , 𝒗2 , … , 𝒗𝑟 and we say that the vectors
𝒗1 , 𝒗2 , … , 𝒗𝑟 span W.
To indicate that W is the space spanned by the vectors in
the set 𝑆 = {𝒗1 , 𝒗2 , … , 𝒗𝑟 }, we write 𝑊 = span (𝑆) or
𝑊 = span{𝒗1 , 𝒗2 , … , 𝒗𝑟 } .
Theorem 3.
Let 𝑆 = {𝒗1 , 𝒗2 , … , 𝒗𝑟 } ⊆ ℝ𝑛 and A be an 𝑛 × 𝑘 matrix
whose columns are 𝑣1 , , … , 𝒗𝑟 . Then 𝑆 spansℝ𝑛 if and
only if the system 𝐴𝑥 = 𝑏 is consistent for every𝑏 ∈ ℝ𝑛 .
Example 6.
Determine whether 𝒗1 = (1,1,2), 𝒗2 = (1,0,1), and𝒗3 =
(2,1,3) span the vector space ℝ3 , or not.?
Solution.
We must determine whether an arbitrary vector 𝒃 =
(𝑏1 , 𝑏2 , 𝑏3 )in ℝ3 can be expressed as a linear
combination 𝒃 = 𝑘1 𝒗1 + 𝑘2 𝒗2 + 𝑘3 𝒗3 of the vectors
- 126 -
𝒗1 , 𝒗2 and 𝒗3 . Expressing this equation in terms of
components gives
(𝑏1 , 𝑏2 , 𝑏3 ) = 𝑘1 (1,1,2) + 𝑘2 (1,0,1) + 𝑘3 (2,1,3)
or
𝑘1 + 𝑘2 + 2𝑘3 = 𝑏1
𝑘1 + 𝑘3 = 𝑏2
2𝑘1 + 𝑘2 + 3𝑘3 = 𝑏3
This system is consistent for all 𝑏1 , 𝑏2 and 𝑏3 if and only if
1 1 2
the coefficient matrix 𝐴 = [1 0 1] is invertible.
2 1 3
Butd𝑒𝑡 (𝐴) = 0 (verify), so that A is not invertible.
Consequently, 𝑣1 , 𝑣2 and𝑣3 do not span ℝ3 .
We leave the proof of the following useful theorem as an
exercise.
Theorem 4.
If𝑆 = {𝒗1 , … , 𝒗𝑟 } and 𝑆′ = {𝒘1 , … , 𝒘𝑟 }are two sets of
vectors in a vector space V, then span{𝒗1 , … , 𝒗𝑟 } =
span{𝒘1 , … , 𝒘𝑟 }if ad only if each vector in 𝑆 is a linear
combination of those in 𝑆′. And conversely each vector in
𝑆′ is a linear combination of those in 𝑆.
Example 7.
- 127 -
Prove that 𝑝(𝑥 ) = 𝑥 2 + 1, 𝑞(𝑥 ) = 𝑥 − 2, and r(𝑥 ) = 𝑥 +
3span 𝑃2
Solution.
Let 𝑚(𝑥 ) = 𝑎𝑥 2 + 𝑏𝑥 + 𝑐 ∈ 𝑃2 such that 𝑚(𝑥 ) =
𝛼1 𝑝(𝑥 ) + 𝛼2 𝑞(𝑥 ) + 𝛼3 𝑟(𝑥) for 𝛼1 , 𝛼2 , 𝛼3 ∈ ℝ.
Therefore,
𝛼1 = 𝑎
𝛼2 + 𝛼3 = 𝑏
𝛼1 − 2𝛼2 + 3𝛼3 = 𝑐
Hence the coefficient matrix is
1 0 0
𝐴 = [0 1 1]
1 −2 3
Is invertible (verify?).
Thus, the system above is consistent. So, 𝑚(𝑥) is a linear
combination of 𝑝, 𝑞, 𝑟 and then span 𝑃𝑛 .◄
●The subspace W in Theorem 2 above is called the
subspace spanned by 𝑆 and is denoted by 〈𝑆〉.
Remark:
(1) If 𝑆 spans V, then 𝑉 = 〈𝑆〉.
(2) If = ϕ or𝑆 = {𝟎} , then 〈𝑆〉 = {𝟎}.
(3) If 𝑆 is a subspace of 𝑉, then 〈𝑆〉 = 𝑆
- 128 -
Example 8.
Find the vectors that spans the subspace W of ℝ3 , where
𝑊 = {(𝑥, 𝑦, 𝑧) ∈ ℝ3 : 𝑥 + 𝑦 + 𝑧 = 0}.
Solution: Since+𝑦 + 𝑧 = 0, then𝑧 = −𝑥 − 𝑦.Therefore
if (𝑥, 𝑦, 𝑧) ∈ 𝑊,then
(𝑥, 𝑦, 𝑧) = (𝑥, 𝑦, −𝑥 − 𝑦)
= (𝑥, 0, −𝑥 ) + (0, 𝑦, −𝑦)
= 𝑥 (1,0, −1) + 𝑦(0,1, −1)
Hence 𝑆 = {(1,0, −1), (0,1, −1)} spans W. ◄
Exercise Set (4.4)
In Exercise (1) – (9), express the first vector as a linear
combination of the other vectors.
(1)(1, −7,5) ; (1, −1,1) , (1,0,1) , (1,1,0);
(2) (−9, −7, −15) ; (2,1,4) , (1, −1,3) , (3,2,5);
(3) (0,0,0) ; (2, 1, 4) , (1, −1, 3) , (3, 2, 5);
(4) (2,3, −7,3) ; (2,1,0,3) , (3, −1,5,2) , (−1,0,2,1)
(5) 𝑥 2 + 1 ; −𝑥 + 1 , 3𝑥 2 + 𝑥 + 2;
(6) 2𝑥 2 − 3𝑥 + 1 ; 𝑥 + 1 , 𝑥 2 + 𝑥 , 𝑥 2 + 2;
(7) 9𝑥 2 + 8𝑥 + 7 ; 4𝑥 2 + 𝑥 + 2 , 3𝑥 2 − 𝑥 + 1 , 5𝑥 2 +
2𝑥 + 3;

- 129 -
 7 19 2 4 3 1
(8)𝐴 = [ ] ; 𝐴1 = [ ] , 𝐴2 = [ ] , 𝐴3 =
−3 8 0 2 3 2
0 10
[ ];
−6 2
6 −8 4 2
(9) 𝐴 = [ ] ; 𝐴1 = [ ] , 𝐴2 =
−1 −8 −2 −2
1 −1 0 2
[ ] , 𝐴3 = [ ];
2 3 1 4
(10) Is 𝑤 = (−1,4,15) ∈ 〈𝑆〉, where
𝑆 = {𝒗1 = (1,2,8) , 𝒗2 = (3,0,1)} ⊆ ℝ3 .
(11) Is 𝑝(𝑥 ) = 6𝑥 2 + 4𝑥 ∈ 〈𝑆〉, where
𝑆 = {𝑃1 (𝑥 ) = 𝑥 2 + 𝑥 + 1, 𝑃2 (𝑥 ) = 3𝑥 2 + 2𝑥 − 2,
𝑃3 (𝑥 ) = 6𝑥 3 + 5𝑥 2 + 5𝑥 + 1} ⊆ 𝑃2
0 5
(12) Is 𝐴 = [ ] ∈ 〈𝑆〉, where
−4−1
1 2 3 1
𝑆 = {𝐴1 = [ ], 𝐴2 = [ ] ⊆ 𝑀2×2
−1 0 1 1
(13) Which of the following subsets of ℝ3 spansℝ3 .
(a) 𝑆1 = {(2,0,2), (3,1,1), (−3,5,5).
(b) 𝑆2 = {(1,5,1), (2,6,1), (3,3,0), (4,6,2)}.
(14) Which of the following subsets of 𝑃2 spans 𝑃2
(a) 𝑆1 = {−𝑥 2 + 1, 𝑥 2 + 1, 6𝑥 2 + 5𝑥 + 2}.
(b) 𝑆2 = {3𝑥, 𝑥 + 1, 2𝑥 2 + 1}.

- 130 -
(c) 𝑆2 = {−8𝑥 2 + 15𝑥 + 3, −5𝑥 2 + 3𝑥, 3𝑥 2 + 5𝑥 +
2, −𝑥 2 + 4𝑥 + 1}.
(15) Which of the following subsets of 𝑀2×2 spans 𝑀2×2
1 0 1 0 0 1 1 1
(a) 𝑆1 = {[ ], [ ], [ ], [ ]}
0 0 0 1 1 0 0 1
2 1 3 1 2 2 2 3
(b) 𝑆2 = {[ ], [ ], [ ], [ ]}
−1 0 1 1 1 1 −4 −1
5. linear Independence
In the preceding section we learned that asset of vectors
𝑆 = {𝒗1 , 𝒗2 , … , 𝒗𝑟 } spans a linear space V if every vector
in V is expressible as a linear combination of the vectors
in S. In general, there may be more than one way to
express a vector in V as a linear combination of vectors in
a spanning set. In this section we will study conditions
under which
each vector in V is expressible as a linear combination of
the spanning vectors in exactly one way. Spanning sets
with this property play a fundamental role in the study of
vector spaces.
Definition 1.
If𝑆 = {𝒗1 , 𝒗2 , … , 𝒗𝑟 }is nonempty set of vectors, then the
vector equation

- 131 -
 𝑘1 𝒗1 + 𝑘2 𝒗2 + ⋯ + 𝑘𝑟 𝒗𝑟 = 𝟎
has at least one solution, namely
𝑘1 = 0, 𝑘2 = 0, … , 𝑘𝑟 = 0
If this is the only solution, then S is called a linearly
independent set. If there are other solutions, then S is
called linearly dependent set.
Example 1.
If𝒗1 = (2, −1, 0, 3), 𝒗2 = (1, 2, 5, −1) and 𝒗3 =
(7, −1, 5, 8), then the set vectors𝑆 = {𝑣1 , 𝒗2 , 𝒗3 } is
linealy dependent, since 3𝒗1 + 𝒗2 − 𝒗3 = 𝟎 .◄
Example 2.
The polynomials 𝑃1 = 1 − 𝑥, 𝑃2 = 5 + 3𝑥 − 2𝑥 2 , and
𝑃3 = 1 + 3𝑥 − 𝑥 2 form a linearly dependent set in 𝑃2
since 3𝑃1 − 𝑃2 + 2𝑃3 = 𝟎.◄
Example 3.
Consider the vectors 𝒊 = (1,0,0), 𝒋 = (0,1,0), and 𝒌 =
(0,0,1)in ℝ3 . In terms of components the vector
equation𝑘1 𝒊 + 𝑘2 𝒋 + 𝑘3 𝒌 = 𝟎 becomes
𝑘1 (1,0,0) + 𝑘2 (0,1,0) + 𝑘3 (0,0,1) = (0,0,0)
Or equivalently, (𝑘1 , 𝑘2 , 𝑘3 ) = (0,0,0)

- 132 -
This implies that 𝑘1 = 0, 𝑘2 = 0 and 𝑘3 = 0.So the set
𝑆 = {𝒊, 𝒋, 𝒌} is linearly independent.
●A similar argument can be used to show that the vectors
𝒆1 = (1,0,0, … ,0), 𝒆2 = (0,1,0, … ,0), …,
𝒆𝑛 = (0,0, … ,1)form a linearly independent set in ℝ𝑛 .◄
Example 4.
Determine whether the vectors
𝑣1 = (1, −2, 3), 𝑣2 = (5, 6, 1), 𝑣3 = (3, 2, 1)
from a linearly dependent set or linearly independent set.
Solution.
In terms of components the vector equation
𝑘1 𝒗1 + 𝑘2 𝒗2 + 𝑘3 𝒗3 = 0
becomes
𝑘1 (1, −2,3) + 𝑘2 (5,6, −1) + 𝑘3 (3,2,1) = (0,0,0)
Or equivalently
(𝑘1 + 5𝑘2 + 3𝑘3 , −2𝑘1 + 6𝑘2 + 2𝑘3 , 3𝑘1 − 𝑘2 + 𝑘3 )
= (0,0,0)
Equating corresponding components gives.
𝑘1 + 5𝑘2 + 3𝑘3 = 0
−2𝑘1 + 6𝑘2 + 2𝑘3 = 0
3𝑘1 − 𝑘2 + 𝑘3 = 0
- 133 -
Thus, 𝒗1 , 𝒗2 and 𝒗3 form a linearly dependent set if this
system has a nontrivial solution or a linearly independent
set if it has only the trivial solution. Solving this system
gives
1 1
𝑘1 = − 𝑡, 𝑘2 = − 𝑡, 𝑘3 = 𝑡, 𝑡 ∈ 𝑘
2 2
Thus, the system has nontrivial solutions and𝒗1 , 𝒗2 and
𝒗3 form a linearly dependent set.◄

●Alternatively, we could show the existence of nontrivial

solutions without solving the system by showing that the
coefficient matrix has determinant Zero and
consequently is not invertible (verify).

Theorem 1. A set 𝑆 with two or more vectors is:

(a)Linearly dependent if and only if at least one of the
vectors in 𝑆 is expressible as a linear combination of the
other vectors in 𝑆.
(b) Linearly independent if and only if no vector in 𝑆is
expressible as a linear combination of the other vectors in
𝑆.

- 134 -
Proof:
(a) Let 𝑆 = {𝒗1 , 𝒗2 , … , 𝒗𝑟 }be aset with two or more
vectors. If we assume that 𝑆is linearly dependent, then
there are scalars 𝑘1 , 𝑘2 , … , 𝑘𝑟 , not all zero, such that
𝑘1 𝒗1 + 𝑘2 𝒗2 + ⋯ + 𝑘𝑟 𝒗𝑟 = 𝟎 (∗)
To be specific, suppose that 𝑘𝑗 ≠ 0, 1 ≤ 𝑗 ≤ 𝑟then (*)
can be rewritten as
𝑘1 𝑘2 𝑘𝑗−1
𝒗𝑗 = (− ) 𝒗1 + (− ) 𝒗2 + ⋯ + (− ) 𝒗𝑗−1
𝑘𝑗 𝑘𝑗 𝑘𝑗
𝑘𝑗+1 𝑘
+ (− ) 𝒗𝑗+1 + ⋯ + (− 𝑟 ) 𝒗𝑟 .
𝑘𝑗 𝑘𝑗

Which expresses 𝒗𝑗 as a linear combination of the other

vectors in 𝑆.
Conversely, assume that at least one of the vectors in 𝑆 is
expressible as a linear combination ofthe others vectors.
To be specific, suppose that
𝒗𝑗 = 𝑐1 𝒗1 + ⋯ + 𝑐𝑗−1 𝒗𝑗−1 + 𝑐𝑗+1 𝒗𝑗+1 + ⋯ + 𝑐𝑟 𝒗𝑟
So
𝑐1 𝒗1 + ⋯ + (−1)𝒗𝑗 + ⋯ + 𝑐𝑟 𝒗𝑟 = 0.
It follows that S is linearly dependent since the
equation𝑘1 𝒗1 + 𝑘2 𝒗2 + ⋯ + 𝑘𝑟 𝒗𝑟 = 0 is satisfied

- 135 -
by𝑘1 = 𝑐1 , 𝑘2 = 𝑐2 , … , 𝑘𝑗 = −1, 𝑘𝑟 = 𝑐𝑟 which are all
not zero.
(b) Similar to (a).◄
Example 5.
In Example 1 above we saw that the vectors 𝒗1 =
(2, −1, 0, 3), 𝒗2 = (1, 2, 5, −1), and 𝒗3 = (7, −1, 5, 8)
form a linearly dependent set. It follows from Theorem 1
above that at least one of these vectors is expressible as a
linear combination of the other two. In this example each
vector is expressible as a linear combination of the other
two since it follows from the equation 3𝒗1 + 𝒗2 − 𝒗3 =
1 1
𝟎 that 𝒗1 = − 𝒗2 + 𝒗3 , 𝒗2 = −3𝒗1 + 𝒗3 , and 𝒗3 =
3 3

3𝒗1 + 𝒗2 .◄
Example 6.
In Example 3 in this section we saw that the vectors 𝒊 =
(1,0,0), 𝒋 = (0,1,0), and 𝒌 = (0,0,1) form a linearly
independent set. Thus, it follows from Theorem 1 above
that none of these vectors is expressible as a linear
combination of the other two. To see directly that,

- 136 -
suppose that k is expressible as𝒌 = 𝛼1 𝒊 + 𝛼2 𝒋.Then, in
terms of components, (0,0,1) = 𝛼1 (1,0,0) + 𝛼2 (0,1,0).
Or(0,0,1) = (𝛼1 , 𝛼2 , 0). But this equation is not satisfied
by any values of 𝛼1 and𝛼1 .So k cannot be expressed as a
linear combination of i and j. Similarly, i is not
expressible as a linear combination of j and k, and j is not
expressible as a linear combination of i and k.◄
Theorem 2.
(a) A finite set of vectors that contains the zero vector is
linearly dependent.
(b) A set with exactly two vectors is linearly independent
if and only if neither of the vectors is a scalar multiple of
the other.
Theorem 3.
Let 𝑆 = {𝒗1 , 𝒗2 , … , 𝒗𝑟 } be a set of vectors in ℝ𝑛 . If > 𝑛 ,
then𝑆 is linearly dependent.
Remark.
The preceding theorem tells us that a set in ℝ2 with more
than two vectors is linearly dependent, and a set in ℝ3
with more than three vectors is linearly dependent.

- 137 -
Example 7.
Determine whether the following set of vectors in 𝑃2 is
linearly dependent
1 + 𝑥, 3𝑥 + 𝑥 2 , 2 + 𝑥 − 𝑥 2 .
Solution.
Let 𝛼1 (1 + 𝑥 ) + 𝛼2 (3𝑥 + 𝑥 2 ) + 𝛼3 (2 + 𝑥 − 𝑥 2 ) = 0.
Then
𝛼1 + 2𝛼3 = 0
𝛼1 + 3𝛼2 + 𝛼3 = 0
𝛼2 − 𝛼3 = 0
The determinant of the coefficient's matrix A for this
system of equation is
1 0 2
det 𝐴 = |1 3 1 | = −2 ≠ 0.
0 1 −1
Therefore A is Invertible, and the only solution is the

trivial solution 𝛼1 = 𝛼2 = 𝛼3 = 0.Thus the vectors are

linearly dependent. ◄

- 138 -
 Exercise Set (4.5)
In Exercise (1) – (5), determine whether the given set of
vectors is linearly dependent or independent in the given
vectors spaces:
(1) 𝑆 = {(1, −1), (2, 0)}, 𝑉 = ℝ2 ;
(2) 𝑆 = {(1, 1, 2), (1, 4, 5), (1, 2, 7), (−1, 8, 3)},
𝑉 = ℝ4 ;
(3) 𝑆 = {4𝑥 2 − 𝑥 + 2, 2𝑥 2 + 6𝑥 + 3, −4𝑥 2 + 10𝑥 +
2}, 𝑉 = 𝑃2 ;
1 1 1 0 1 0
(4) 𝑆 = {[ ], [ ], [ ]} , 𝑉 = 𝑀2×2 ;
0 1 1 1 0 1
(5) Find x such that 𝑆 is linearly independent in ℝ3 ,
where𝑆 = {(2, 𝑥, 1), (1,0,1), (0,1,3)}
(7)Find 𝑎, 𝑏 such that the vectors
(1, 𝑎, 2, 𝑏), (1, −1,2,0), (1,2,1,3) are linearly dependent
inℝ4 .
(8)Prove that any nonempty subset of linearly
independent set is linearly independent.
(9)Let 𝑆1 , 𝑆2 be subsets of a vector space V such that 𝑆1 ⊆
𝑆2 and 𝑆1 is linearly dependent.Prove that 𝑆2 is linearly
dependent.

- 139 -
6. Basis and Dimension

Let 𝑆1 = {(2, 1), (1, −1), (3, −2)} and 𝑆2 =

{(1, −1), (1, 1)} and be two subset of ℝ2 . It is easy to

show that 𝑆1 is linearly dependent and 𝑆2 is linearly

independent. Therefore, 𝑆2 is a basis for ℝ2 but 𝑆1 is not.

In this section we discuss the notion of basis in vector

spaces. We will show that each basis for a vector space

contains the same number of elements. This number is

called the dimension of vector space. The number of

elements in a basis for a vector space V depends only on

V and not on the way we chose the basis.

Definition 1.
If V is any vector space and 𝑆 = {𝒗1 , 𝒗2 , … , 𝒗𝑛 } is a set
of vectors in V, then 𝑆 is called a basis for V if the
following two conditions hold:
(a) 𝑆 is linearly independent

(b) S spans V.
- 140 -
Theorem 1.
If 𝑆 = {𝒗1 , 𝒗2 , … , 𝒗𝑛 } is a basis for a vector spaceV, then
every vector v in V can be expressed in the form 𝒗 =
𝑐1 𝒗1 + 𝑐2 𝒗2 + ⋯ + 𝑐𝑛 𝒗𝑛 in exactly one way.
Proof.
Since 𝑆 spans V, it follows from the definition of the
spanning set that every vector in V is expressible as a
linear combination of the vectors in 𝑆. To see that there is
only one way to express a vector as a linear combination
of the vectors in 𝑆,supposethat some vector v can be
written as
𝒗 = 𝑐1 𝒗1 + 𝑐2 𝒗2 + ⋯ + 𝑐𝑛 𝒗𝑛
and also, as
𝒗 = 𝑘1 𝒗1 + 𝑘2 𝒗2 + ⋯ + 𝑘𝑛 𝒗𝑛
Subtracting the second equation from the first gives
0 = (𝑐1 − 𝑘1 )𝒗1 + (𝑐2 − 𝑘2 )𝒗2 + ⋯ + (𝑐𝑛 − 𝑘𝑛 )𝒗𝑛
The linear independence of 𝑆 implies that
𝑐1 − 𝑘1 = 0 , 𝑐2 − 𝑘2 = 0 … , 𝑐𝑛 − 𝑘𝑛 = 0.
That is, 𝑐1 = 𝑘1 , 𝑐2 = 𝑘2 , … , 𝑐𝑛 = 𝑘𝑛 .
Thus, the two expressions for V are the same.

- 141 -
If 𝑆 = {𝒗1 , 𝒗2 , … , 𝒗𝑛 } is a basis for a vector space V, and
𝒗 = 𝑐1 𝒗1 + 𝑐2 𝒗2 + ⋯ + 𝑐𝑛 𝒗𝑛
is the expression for a vector v in terms of the basis 𝑆,
then the scalars 𝑐1 , 𝑐2 , … , 𝑐𝑛 are called the coordinates of
v relative to the basis 𝑆. The vector (𝑐1 , 𝑐2 , … , 𝑐𝑛 ) in ℝ𝑛
constructed from these coordinates is called the
coordinates vector of V relative to ; it is denoted by
(𝒗)𝑆 = (𝑐1 , 𝑐2 , … , 𝑐𝑛 ). ◄
Remark: It should be noted that coordinate vectors
depend not only on the basis 𝑆 but also on the order in
which the basis vectors are written; a change in the order
of the basis vectors results in a corresponding change of
order for the entries in the coordinate vectors.
Example 1.
If 𝒊 = (1,0,0), 𝒋 = (0,1,0), and = (0,0,1) ,then it is easy
to show that 𝑆 = {𝒊, 𝒋, 𝒌} is a linearly independent set in
ℝ3 . This set also spansℝ3 since any vector 𝒗 = (𝑎, 𝑏, 𝑐)
in ℝ3 can be written as
𝒗 = (𝑎, 𝑏, 𝑐 ) = 𝑎(1,0,0) + 𝑏(0,1,0) + 𝑐(0,0,1)
= 𝑎𝒊 + 𝑏𝒋 + 𝑐𝒌 (1)
Thus, 𝑆 is a basis for ℝ3 ; it is called the standard basis
forℝ3 . Looking at the coefficients of i, j and k in (1), it

- 142 -
follows that the standard basis are 𝑎, 𝑏, and c. So (𝒗)𝑠 =
(𝑎, 𝑏, 𝑐). Comparing this result we see that 𝒗 = (𝒗)𝑠 . ◄
●The result in the preceding example is a special case of
those in the next example.
Example 2.
As shown in the preceding section = {𝒆1 , 𝒆2 , … , 𝒆𝑛 } ,
where 𝒆1 = (1,0, … ,0), 𝒆2 = (0,1,0, … ,0) , … , 𝒆𝑛 =
(0,0, … ,1), is a linearly independent set in ℝ𝑛 . This set
also spans ℝ𝑛 . Since any vector 𝒗 = (𝒗1 , 𝒗2 , … , 𝒗𝑛 ) in
ℝ𝑛 can be written as
𝒗 = 𝒗1 𝒆1 + 𝒗2 𝒆2 + ⋯ + 𝒗𝑛 𝒆𝑛 (2)
Thus 𝑆 is a basis for ℝ𝑛 ; it is called the standard basis
for ℝ𝑛 . It follows from (2) that the coordinates of v
relative to the standard basis are
(𝒗)𝑆 = (𝒗1 , 𝒗2 , … , 𝒗𝑛 ) = 𝒗.◄
Remark.
We will see in a subsequent example that a vector and its
coordinate vector will not be the same; the equality that
we observed in the two preceding examples is a special
situation that occurs only with the standard basis for ℝ𝑛 .

- 143 -
Example 3.
Let 𝒗1 = (1, 2, 1), 𝒗2 = (2, 4, 0),and𝒗3 = (3, 3, 4). Show
that the set 𝑆 = {𝒗1 , 𝒗2 , 𝒗3 } is a basis for ℝ3 .
Solution.
To show that the set 𝑆spansℝ3 , we must show that any
arbitrary vector 𝒃 = (𝑏1 , 𝑏2 , 𝑏3 ) can be expressed as a
linear combination 𝑏 = 𝑐1 𝒗1 + 𝑐2 𝒗2 + 𝑐3 𝒗3 of vectors
in 𝑆. Expressing this equation in terms of components
gives
(𝑏1 , 𝑏2 , 𝑏3 ) = 𝑐1 (1, 2, 1) + 𝑐2 (2, 4, 0) + 𝑐3 (3, 3, 4)
or on equating corresponding components
𝑐1 + 2𝑐2 + 3𝑐3 = 𝑏1
2𝑐1 + 9𝑐2 + 3𝑐3 = 𝑏2 (3)
𝑐1 + 4𝑐3 = 𝑏3
Thus, to show that 𝑆 spasmℝ3 , we must demonstrate the
system (3) has a solution for all choice of = (𝑏1 , 𝑏2 , 𝑏3 ) .
To prove that 𝑆 is linearly independent, we must show
that the only solution of
𝑐1 𝒗1 + 𝑐2 𝒗2 + 𝑐3 𝒗3 = 𝟎 (4)
Is 𝑐1 = 𝑐2 = 𝑐3 = 0.

- 144 -
As above, if (4) is expressed in terms of components, the
verification of independence reduces to showing that the
homogeneous system
𝑐1 + 2𝑐2 + 3𝑐3 = 0
2𝑐1 + 9𝑐2 + 3𝑐3 = 0 (5)
𝑐1 + 4𝑐3 = 0
has only the trivial solution. Observe that the systems (3)
and (5) have the same coefficient matrix. Thus from
Theorem 4, in Section 2 of Chapter (III) we can
simultaneously prove that 𝑆is linearly independent and
spans ℝ3 by demonstrating that in systems (3) and (5) the
matrix of coefficients
1 2 3
𝐴 = [2 9 3]
1 0 4
has anon zero determinant. But det (𝐴) = −1. So, 𝑆 is a
basis for ℝ3 . ◄
Example 4.
Let 𝑆 = {𝒗1 , 𝒗2 , 𝒗3 }be the basis for ℝ3 in the preceding
example.
(a)Find the coordinate vector of𝒗 = (5, −1, 9) with
respect to S.
- 145 -
(b)Find the vector v inℝ3 whose coordinate vector with
respect to the basis 𝑆 is (𝒗)𝑆 = (−1, 3, 2) .
Solution.
(a)We must find scalars 𝑐1 , 𝑐2 , 𝑐3 such that
𝒗 = 𝑐1 𝒗1 + 𝑐2 𝒗2 + 𝑐3 𝒗3 or, in terms of components,
(5, −1, 9) = 𝑐1 (1, 2, 1) + 𝑐2 (2, 9, 0) + 𝑐3 (3, 3, 4)
Equating corresponding components gives
𝑐1 + 2𝑐2 + 3𝑐3 = 5
2𝑐1 + 9𝑐2 + 3𝑐3 = −1
𝑐1 + 4𝑐3 = 9
Solving this system, we obtain 𝑐1 = 1, 𝑐2 = −1, 𝑐3 = 2
(verify). Therefore, (𝒗)𝑆 = (1, −1, 2).
(b)Using the definition of the coordinate vector (𝒗)𝑆 ,we
obtain
𝒗 = (−1)𝒗1 + 3𝒗2 + 2𝒗3
= (−1)(1, 2, 1) + 3 (2, 9, 0) + 2(3, 3, 4)
= (11, 31, 7). ◄
Example 5.
(a) Show that 𝑆 = {1, 𝑥, 𝑥 2 , … , 𝑥 𝑛 } is a basis for the
vector space 𝑃𝑛 of polynomials of the form 𝑎0 + 𝑎1 𝑥 +
⋯ + 𝑎𝑛 𝑥 𝑛 .
- 146 -
(b) Find the coordinate vector of the polynomial
𝑝 = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 relative to basis 𝑆 = {1, 𝑥, 𝑥 2 } for
𝑃2 .
Solution.
(a)The student can easily show that 𝑆 spans 𝑃𝑛 and 𝑆is
linearly independent set. Thus, 𝑆 is a basis for 𝑃𝑛 ; it is
called the standard basis for 𝑃𝑛 .
(b)The coordinates of 𝑝 = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 are the
scalar coefficients of the basis vectors 1, 𝑥, and 𝑥 2 so
(𝑝)𝑆 = (𝑎0 , 𝑎1 , 𝑎2 ).◄
Example 6.
1 0 0 1 0 0
Let 𝑴1 = [ ] , 𝑴2 = [ ] , 𝑴3 = [ ],
0 1 0 0 1 0
0 0
𝑴4 = [ ]. The set 𝑆 = {𝑴1 , 𝑴2 , 𝑴3 , 𝑴4 } is a basis
0 1
for the vector space𝑀2×2 of 2 × 2 Matrices.
Solution.
To see that 𝑆 spans 𝑀2×2 , note that an arbitrary vector
𝑎 𝑏
(matrix) [ ] can be written as
𝑐 𝑑
𝑎 𝑏 1 0 0 1 0 0 0 0
[ ] = 𝑎[ ]+𝑏[ ]+𝑐[ ]+𝑑[ ]
𝑐 𝑑 0 1 0 0 1 0 0 1
= 𝑎𝑴1 + 𝑏𝑴2 + 𝑐𝑴3 + 𝑑𝑴4

- 147 -
To see that 𝑆 is linearly independent, assume that
𝑎𝑴1 + 𝑏𝑴2 + 𝑐𝑴3 + 𝑑𝑴4 = 𝟎
That is,
1 0 0 1 0 0 0 0 0 0
𝑎[ ]+𝑏[ ]+𝑐[ ]+𝑑[ ]=[ ].
0 0 0 0 1 0 0 1 0 0
It follows that
𝑎 𝑏 0 0
[ ]=[ ]
𝑐 𝑑 0 0
Thus, 𝑎 = 𝑏 = 𝑐 = 𝑑 = 0.So𝑆 is linearly independent.
The basis 𝑆 in this example is called the standard basis
for𝑀2×2 . ◄
●More generally, the standard basis for 𝑀𝑚×𝑛 consists
of the 𝑚 × 𝑛 different Matrices with a single l and zeros
for theremaining entries.
Definition 2.
A nonzero vector space V is called finite – dimensional
if it contains a finite set of vectors {𝒗1 , 𝒗2 , … , 𝒗𝑛 } that
forms abasis. If no such set exists, V is called infinite –
dimensional. In addition, we shall regard the zero vector
space to be finite – dimensional.
Example 7.
The vector spacesℝ𝑛 ,𝑃𝑛 , and 𝑀𝑚×𝑛 are finite –
dimensional. The vector spaces𝐹 (−∞, ∞),
𝐶 (−∞, ∞), 𝐶 𝑚 (−∞, ∞) and 𝐶 ∞ (−∞, ∞) are infinite –
dimensional. The next theorem will provide the key to the
concept of dimensions.
- 148 -
Theorem 2.
If V is a finite – dimensional vector space and
{𝒗1 , 𝒗2 , … , 𝒗𝑛 }is any basis. Then
(a) Every set with more than n vectors is linearly
dependent.
(b) No set with fewer than n vectors spans V.
Proof.
(a) Let 𝑆′ = {𝒘1 , 𝒘2 , … , 𝒘𝑚 } be any set of m vectors in
V, where 𝑚 > 𝑛.
We want to Show that 𝑆′ is linearly dependent. Since
𝑆 = {𝒗1 , 𝒗2 , … , 𝒗𝑛 } is a basis, each𝒘𝑖 can be
expressedas a linear combination of the vectors in𝑆,say
𝒘1 = 𝑎11 𝒗1 + 𝑎21 𝒗2 + ⋯ + 𝑎𝑛1 𝒗𝑛
𝒘2 = 𝑎12 𝒗1 + 𝑎22 𝒗2 + ⋯ + 𝑎𝑛2 𝒗2
⋮⋮⋮⋮
𝒘𝑚 = 𝑎1𝑚 𝒗1 + 𝑎2𝑚 𝒗2 + ⋯ + 𝑎𝑛𝑚 𝒗𝑛
(6)
To show that 𝑆′ is linearly dependent, we must find
scalars𝑘1 , 𝑘2 , … , 𝑘𝑚 , not all zero, such that
𝑘1 𝒘1 + 𝑘2 𝒘2 + ⋯ + 𝑘𝑚 𝒘𝑚 = 0 (7)
Using the equations in (6), we can rewrite (7) as
- 149 -
 (𝑘1 𝑎11 + 𝑘2 𝑎12 + ⋯ + 𝑘𝑚 𝑎1𝑚 )𝒗1 +
(𝑘1 𝑎21 + 𝑘2 𝑎22 + ⋯ + 𝑘𝑚 𝑎2𝑚 )𝒗2 + ⋯
+(𝑘1 𝑎𝑛1 + 𝑘2 𝑎𝑛2 + ⋯ + 𝑘𝑚 𝑎𝑛𝑚 )𝒗𝑛 = 0
Thus, from the linear independence of S, the problem of
proving that𝑆 ′ is linearly dependent set reduces to
showing there are scalars 𝑘1 , 𝑘2 , … , 𝑘𝑚 , not all zero, that
satisfy
𝑎11 𝑘1 + 𝑎12 𝑘2 + ⋯ + 𝑎1𝑚 𝑘𝑚 = 0
𝑎21 𝑘1 + 𝑎22 𝑘2 + ⋯ + 𝑎2𝑚 𝑘𝑚 = 0
⋮⋮⋮⋮
𝑎𝑛1 𝑘1 + 𝑎𝑛2 𝑘2 + ⋯ + 𝑎𝑛𝑚 𝑘𝑚 = 0
(8)
But (8) has more unknowns than equations, then the
system has nontrivial solution. So the proof is complete.
(b) Let 𝑆 ′ = {𝒘1 , 𝒘2 , … , 𝒘𝑚 } be any set of m vectors in
V, where𝑚 < 𝑛.
We want to show that 𝑆′does not span V. The proof will
be by contradiction. We will show that assuming 𝑆′spans
V leads to a contradiction of the linear independence of
{𝒗1 , 𝒗2 , … , 𝒗𝑛 } .

- 150 -
If 𝑆′ spans V, then every vector in V is a linear
combination of the vectors in 𝑆 ′ .In particular each basis
vector 𝒗𝑖 is a linear combination of the vectors in 𝑆′ , say
𝒗1 = 𝑎11 𝒘1 + 𝑎21 𝒘2 + ⋯ + 𝑎𝑚1 𝒘𝑚
𝒗2 = 𝑎12 𝒘1 + 𝑎22 𝒘2 + ⋯ + 𝑎𝑚2 𝒘𝑚
⋮⋮⋮⋮
𝒗𝑛 = 𝑎1𝑛 𝒘1 + 𝑎2𝑛 𝒘2 + ⋯ + 𝑎𝑚𝑛 𝒘𝑚
(9)
To obtain our contradiction we will show that there are
scalars 𝑘1 , 𝑘2 , … , 𝑘𝑛 not all zero, such that
𝑘1 𝒗1 + 𝑘2 𝒗2 + ⋯ + 𝑘𝑛 𝒗𝑛 = 0 (10)
But observe that (9) and (10) have the same form as (6)
and (7) except that m and n are interchanged and the w’s
and v’s are interchanged. Thus, the computation that led
to(8) now yields
𝑎11 𝑘1 + 𝑎12 𝑘2 + ⋯ + 𝑎1𝑛 𝑘𝑛 = 0
𝑎21 𝑘1 + 𝑎22 𝑘2 + ⋯ + 𝑎2𝑛 𝑘𝑛 = 0
⋮⋮⋮⋮
𝑎𝑚1 𝑘1 + 𝑎𝑚2 𝑘2 + ⋯ + 𝑎𝑚𝑛 𝑘𝑛 = 0
This linear system has more unknowns than equations,
and therefore has nontrivial solution.◄
- 151 -
Theorem 3.
All basis for a finite – dimensional vector space has the
same number of vectors.
Definition 3.
The dimension of a finite – dimensional vector space V,
denoted by dim (V), is defined to be the number of
vectors in a basis for V. In addition, we define the zero-
vector space to have dimension zero.
Example 8.
dim (ℝ𝑛 ) = 𝑛 [The standard basis has n vectors].
dim (𝑃𝑛 ) = 𝑛 + 1[The standard basis has 𝑛 + 1 vectors].
dim (𝑀𝑚×𝑛 ) = 𝑚𝑛[The standard basis has 𝑚𝑛
vectors].◄

Example 9.
Determine a basis for and the dimension of the solution
space of the homogenous system
2𝑥1 + 2𝑥2 − 𝑥3 + 𝑥5 = 0
−𝑥1 − 𝑥2 + 2𝑥3 − 3𝑥4 + 𝑥5 = 0
𝑥1 + 𝑥2 − 2𝑥3 − 𝑥5 = 0
𝑥3 + 𝑥4 + 𝑥5 = 0

- 152 -
Solution.
The general solution of the given system is
𝑥1 = −𝑠 − 𝑡 , 𝑥2 = 𝑠 , 𝑥3 = −𝑡, 𝑥4 = 0 , 𝑥5 = 𝑡
(Verify). Therefore, the solution vectors can be written
as
𝑥1 −𝑠 − 𝑡 −𝑠 −𝑡 −1 −1
𝑥2 𝑠 𝑠 0 1 0
𝑥3 = −𝑡 = 0 + −𝑡 = 𝑠 0 + 𝑡 −1
𝑥4 0 0 0 0 0
[𝑥5 ] [ 𝑡 ] [ 0 ] [𝑡 ] [0] [1]
Which shows that the vectors
𝒗1 = (−1, 1, 0, 0, 0) and 𝒗2 = (−1, 0, −1, 0, 1)
span the solution space. Since they are also linearly
independent (verify),{𝒗1 , 𝒗2 } is a basis and the solution
space is two- dimensional. ◄
Theorem 4.

If W is a subspace of finite – dimensional vectors space

V, thendim(𝑊 ) ≤ dim (𝑉); moreover, if dim (𝑊) =

dim (𝑉) ,then 𝑊 = 𝑉

- 153 -
 Exercise Set (4.6)
(1) Which of the following sets of vectors are basis for
ℝ2 ?
(a) {(1, 3) , (3, 5)}(𝑏){(7, 1) , (10, 1)}
3 4 −4 3
(c) {(7, 1) , (21, 3)}(𝑑) {( , ) , ( , )}
5 5 5 5

(2) Which of the following sets of vectors are basis for

ℝ3 ?
(a) {(0, 17, 8) , (1, 7, 5) , (2, −3, 2)}
(b) {(√2, 0, √2) , (√3, 0, √6) , (0, −√6, √3)}
(c) {(1, 2, 1) , (1, 3, 1) , (2, 3, 1)}
(d) {(2, 1, 3) , (−1, 2, 1) , (3, 0, 0)}
(3) Which of the following sets of vectors are basis for
ℝ4 ?
(a) {(1,1,0,0) , (1,0,1,0) , (0,1,0,1) , (0,0,1,1)}
(b) {(1,1,3, −1) , (1,1, −1, −1) , (2,2,2, −2) , (3,3,1, −3)}
(c) {(3,0, −3,6) , (0,2,3,1) , (0, −2, −2,0) , (−2,1,2,1)}
(4) Which of the following sets of vectors are basis for
𝑃2 ?
(a) {1, 𝑥 + 1, 𝑥 2 − 𝑥 }
(b) {𝑥 3 + 1 , 𝑥 3 − 1 , 𝑥 2 + 1 , 𝑥 2 − 1}

- 154 -
(c) {𝑥 2 + 3𝑥 + 5 , 5𝑥 2 + 3𝑥 + 1 , 𝑥 2 + 𝑥 + 1}
(d) {𝑥 + 5 , −𝑥 2 − 𝑥 + 1 , 2𝑥 2 + 𝑥 + 3}
(5) Which of the following sets of vectors are basis for
𝑀2×2 ?
1 0 0 −1 2 2 0 3
(a) = {[ ] ,[ ],[ ] ,[ ]}
0 1 −1 0 2 0 3 3
0 1 0 5 0 0 1 7
(b) = {[ ] ,[ ] ,[ ] ,[ ]}
4 4 5 0 2 −3 0 0
1 1 1 1 1 0 0 1
(c) = {[ ] ,[ ] ,[ ] ,[ ]}
1 0 0 1 1 1 1 1
(6) Determine basis for and dimensions of the given
subspaces (≤ denotes subspace).
(a)𝑊 = {(𝑎, 𝑏, 𝑐 ): 𝑎 + 𝑏 + 𝑐 = 0} ≤ ℝ3 .
(b)𝑊 = {(𝑎, 𝑏, 𝑐 ): 𝑎 + 𝑏 = 𝑏 + 𝑐 = 0} ≤ ℝ3 .
(c)𝑊 = {(𝑎 − 𝑏 , 𝑏 + 𝑐 , 𝑏, 𝑏 − 𝑐): 𝑎, 𝑏, 𝑐 ∈ ℝ} ≤ ℝ4 .
(d) 𝑊 = {𝐴 ∈ 𝑀2×2 ∶ 𝐴𝑡 + 𝐴 = 0} ≤ 𝑀2×2 .
(e) 𝑊 = {𝐴 ∈ 𝑀2×2 ∶ 2𝑎11 + 𝑎22 = 𝑎12 + 𝑎21 = 0} ≤ 𝑀2×2
(f) 𝑊 = {𝑎(𝑥 + 1) + 𝑏(𝑥 2 + 𝑥 ): 𝑎, 𝑏 ∈ ℝ} ≤ 𝑃2
(7) Determine the dimension of and a basis for the
solution space of the systems:
(a)𝑥1 + 𝑥2 − 𝑥3 = 0
−2𝑥1 − 𝑥2 + 2𝑥3 = 0

- 155 -
 −𝑥1 + 𝑥3 = 0.
(b)3𝑥1 + 𝑥2 + 𝑥3 + 𝑥4 = 0
5𝑥1 − 𝑥2 + 𝑥3 − 𝑥4 = 0.
(8) Find the coordinate vector of w relative to the basis
𝑆 = {𝒖1 , 𝒖2 } for ℝ2 .
(a)𝒖1 = (1, 0), 𝒖2 = (0, 1); 𝒘 = (3, −7);
(b) 𝒖1 = (1, 1), 𝒖2 = (0, 2) ; 𝒘 = (𝑎, 𝑏).
(10) Find the coordinate vector of p relative to the basis
𝑆 = {𝒑1 , 𝒑2 , 𝒑3 }, where
𝒑 = 4 − 3𝑥 + 𝑥 2 ; 𝒑1 = 1, 𝒑2 = 𝑥 , 𝒑3 = 𝑥 2
(11) Find the coordinate vector of A relative to the basis
= {𝐴1 , 𝐴2 , 𝐴3 , 𝐴4 } , where
2 0 −1 1 1 1
𝑨=[ ] , 𝑨1 = [ ] , 𝑨2 = [ ],
−1 3 0 0 0 0
0 0 0 0
𝑨3 = [ ] , 𝑨4 = [ ]
1 0 0 1

- 156 -
 CHAPTER (V)
INNER PRODUCT SPACES

- 157 -
 CHAPTER (V)
INNER PRODUCT SPACES
1. Inner product
Definition 1.
An inner product on a linear space V having the real
numbers as scalars is a function 〈 , 〉: 𝑉 × 𝑉 → ℝthat
associates with each pair of elements u and v real
number〈𝒖, 𝒗〉 in such a way that the following axioms are
satisfied for all elements u , v and w of V and𝛼 ∈ ℝ:
(1) 〈𝒖, 𝒗〉 = 〈𝒗, 𝒖〉; [Symmetry axiom]
(2) 〈𝒖 + 𝒗, 𝒘〉 = 〈𝒖, 𝒘〉 + 〈𝒗, 𝒘〉; [Additively axiom]
(3) 〈𝛼𝒖, 𝒗〉 = 𝛼 〈𝒖, 𝒗〉; [Homogeneity axiom]
(4) 〈𝒗, 𝒗〉 ≥ 0 [Positivity axiom]
(5) 〈𝒗, 𝒗〉 = 0 if and only if 𝒗 = 𝟎.
Areal linear space with an inner product is called an inner
product space.
Example 1.
Let 𝑉 = ℝ𝑛 and 𝒖 = (𝑎1 , … , 𝑎𝑛 ) and 𝒗 = (𝑏1 , … , 𝑏𝑛 )be
elements in V. Then the function defined by :
〈𝒖, 𝒗〉 = 𝒖 ∙ 𝒗 = 𝑎1 𝑏1 + ⋯ + 𝑎𝑛 𝑏𝑛

- 158 -
is an inner product (called Euclidean inner product). In
this case, V is called Euclidean Product Space.
Solution.
Let 𝒖 = (𝑎1 , … , 𝑎𝑛 ) , 𝒗 = (𝑏1 , … , 𝑏𝑛 ),and 𝒘 =
(𝑐1 , … , 𝑐𝑛 )be elements in ℝ𝑛 and 𝛼𝜖ℝ. Then
(1) 〈𝒖, 𝒗〉 = 𝑎1 𝑏1 + ⋯ + 𝑎𝑛 𝑏𝑛
= 𝑏1 𝑎1 + ⋯ + 𝑏𝑛 𝑎𝑛
= 〈𝒗, 𝒖〉.
(2) 〈𝒖 + 𝒗, 𝒘〉 = (𝑎1 + 𝑏1 )𝑐1 + ⋯ + (𝑎𝑛 + 𝑏𝑛 )𝑐𝑛

= 𝑎1 𝑐1 + 𝑏1 𝑐1 + ⋯ + 𝑎𝑛 𝑐𝑛 + 𝑏𝑛 𝑐𝑛

= (𝑎1 𝑐1 + ⋯ + 𝑎𝑛 𝑐𝑛 ) + (𝑏1 𝑐1 + ⋯ + 𝑏𝑛 𝑐𝑛 )

= 〈𝒖, 𝒘〉 + 〈𝒗, 𝒘〉.

(3) 〈𝜶𝒖, 𝒗〉 = (𝛼𝑎1 )𝑏1 + ⋯ + (𝛼𝑎𝑛 )𝑏𝑛

= 𝛼 (𝑎1 𝑏1 + ⋯ + 𝑎𝑛 𝑏𝑛 ) = 𝛼 〈𝒖, 𝒗〉.

(4) 〈𝒖, 𝒖〉 = 𝑎12 + ⋯ + 𝑎𝑛2 ≥ 0

(5) 〈𝒖, 𝒖〉 = 0 ⟺ 𝑎12 + ⋯ 𝑎𝑛2 = 0 ⟺

𝑎1 = ⋯ = 𝑎𝑛 = 0 ⟺ 𝒖 = 𝟎.

Therefore ℝ2 is an inner product space.

- 159 -
Remark.

The Euclidian inner product is the most important inner

product on ℝ𝑛 . However there are various applications in
which it is desirable to modify the Euclidean inner
product by weighting its terms differently. More
precisely, if 𝛼1 , … , 𝛼𝑛 are positive real numbers, which we
shall call weights, and if𝒖 = (𝑢1 , … , 𝑢𝑛 ) and 𝒗 =
(𝑣1 , … , 𝑣𝑛 ) are vectors in ℝ𝑛 . Then it can be shown (left
as exercise) that the formula

〈𝒖, 𝒗〉 = 𝛼1 𝑢1 𝑣1 + ⋯ + 𝛼𝑛 𝑢𝑛 𝑣𝑛 (∗)

defines an inner product on ℝ𝑛 ,it is called weighted

Euclideaninner product with weights 𝛼1 , … , 𝛼𝑛 .

It will always be assumed that ℝ𝑛 has the Euclidean inner

product unless some other inner product is specially
specified.

Example 2.

Let 𝒖 = (𝑎1 , 𝑎2 ) , 𝒗 = (𝑏1 , 𝑏2 ) ∈ ℝ2 . Verify the

weighted Euclidean inner product 〈𝒖, 𝒗〉 = 7𝑎1 𝑏1 +
5𝑎2 𝑏2 satisfies the five inner product axioms.

- 160 -
Solution.

Let 𝒖 = (𝑎1 , 𝑎2 ), 𝒗 = (𝑏1 , 𝑏2 ), 𝒘 = (𝑐1 , 𝑐2 ) ∈ ℝ2

and ∈ ℝ . Then

(1) 〈𝒖, 𝒗〉 = 7𝑎1 𝑏1 + 5𝑎2 𝑏2 = 7𝑏1 𝑎1 + 5𝑏2 𝑎2 = 〈𝒗, 𝒖〉.

(2) 〈𝒖 + 𝒗, 𝒘〉 = 7(𝑎1 + 𝑏1 )𝑐1 + 5(𝑎2 + 𝑏2 )𝑐2

= (7𝑎1 𝑐1 + 5𝑎2 𝑐2 ) + (7𝑏1 𝑐1 + 5𝑏2 𝑐2 )

= 〈𝒖, 𝒘〉 + 〈𝒗, 𝒘〉.

(3) 〈𝜶𝒖, 𝒗〉 = 7(𝛼𝑎1 )𝑏1 + 5(𝛼𝑎2 )𝑏2

= 𝛼 (7𝑎1 𝑏1 + 5𝑎2 𝑏2 )

= 𝛼 〈𝒖, 𝒗〉.

(4) 〈𝒖, 𝒖〉 = 7𝑎12 + 5𝑎22 ≥ 0.

(5) 〈𝒖, 𝒖〉 = 0 ⇔ 7𝑎12 + 5𝑎22 = 0 ⟺ 𝑎1 = 𝑎2 = 0

⟺ 𝒖 = 𝟎.◄

Example 3.

Let𝑝(𝑥 ) = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 , 𝑞 (𝑥 ) = 𝑏0 + 𝑏1 𝑥 + 𝑏2 𝑥 2 ∈
𝑃2 ,where 〈𝑝(𝑥 ), 𝑞 (𝑥 )〉 = 𝑎0 𝑏0 + 𝑎1 𝑏1 + 𝑎2 𝑏2 . Prove
that 𝑃2 is an inner product space.

- 161 -
Solution.

Let 𝛼 ∈ ℝ and𝑝(𝑥 ) = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 , 𝑞 (𝑥 ) = 𝑏0 +
𝑏1 𝑥 + 𝑏2 𝑥 2 and 𝑟(𝑥 ) = 𝑐0 + 𝑐1 𝑥 + 𝑐2 𝑥 2 ∈ 𝑃2

(1)〈𝑝, 𝑞 〉 = 𝑎0 𝑏0 + 𝑎1 𝑏1 + 𝑎2 𝑏2

= 𝑏0 𝑎0 + 𝑏1 𝑎1 + 𝑏2 𝑎2

= 〈𝑞, 𝑝〉.

(2) 〈𝑝 + 𝑞, 𝑟〉 = (𝑎0 + 𝑏0 )𝑐0 + (𝑎1 + 𝑏1 )𝑐1 +

(𝑎2 + 𝑏2 )𝑐2

= (𝑎0 𝑐0 + 𝑎1 𝑐1 + 𝑎2 𝑐2 ) + (𝑏0 𝑐0 + 𝑏2 𝑐1 + 𝑏2 𝑐2 )

= 〈𝑝, 𝑟〉 + 〈𝑞, 𝑟〉 .

(3) 〈𝛼𝑝, 𝑞 〉 = (𝛼𝑎0 )𝑏0 + (𝛼𝑎1 )𝑏1 + (𝛼𝑎2 )𝑏2

= 𝛼 (𝑎0 𝑏0 + 𝑎1 𝑏1 + 𝑎2 𝑏2 ) = 𝛼 〈𝑝, 𝑞 〉 .

(4) 〈𝑝, 𝑝〉 = 𝑎02 + 𝑎12 + 𝑎22 ≥ 0

(5) 〈𝑝, 𝑝〉 = 0 ⇔ 𝑎02 + 𝑎12 + 𝑎22 = 0

⟺ 𝑎0 = 𝑎1 = 𝑎2 = 0

⟺ 𝑝 = 𝟎.

Therefore 𝑃2 is an inner product space.◄

- 162 -
Example 4.

Let 𝑓, 𝑔 ∈ 𝐶 [𝑎, 𝑏], where 𝐶 [𝑎, 𝑏] is the set consists of all

continuous function on the closed interval [𝑎, 𝑏].
𝑏
Prove that the function 〈𝑓, 𝑔〉 = ∫𝑎 𝑓(𝑥 )𝑔(𝑥 )𝑑𝑥 defines
an inner product space on 𝐶 [𝑎, 𝑏] .
Solution:

Let 𝑓, 𝑔, ℎ ∈ 𝐶 [𝑎, 𝑏] and 𝛼 ∈ ℝ. Then

𝑏 𝑏
(1)〈𝑓, 𝑔〉 = ∫𝑎 𝑓(𝑥)𝑔(𝑥)𝑑𝑥 = ∫𝑎 𝑔(𝑥)𝑓(𝑥)𝑑𝑥 = 〈𝑔, 𝑓〉.

𝑏
(2)〈𝑓 + 𝑔, ℎ〉 = ∫𝑎 [𝑓(𝑥 ) + 𝑔(𝑥 )] ℎ(𝑥 )𝑑𝑥
𝑏 𝑏

= ∫ 𝑓(𝑥 )ℎ(𝑥 )𝑑𝑥 + ∫ 𝑔(𝑥 )ℎ(𝑥 )𝑑𝑥

𝑎 𝑎

= 〈𝑓, ℎ〉 + 〈𝑔, ℎ〉.

𝑏
(3) 〈𝛼𝑓, 𝑔〉 = ∫𝑎 (𝛼𝑔(𝑥 ))𝑔(𝑥 )𝑑𝑥

= 𝛼 ∫ 𝑓(𝑥 )𝑔(𝑥 )𝑑𝑥

𝑎

= 𝛼 〈𝑓, 𝑔〉.

- 163 -
(4) Since f is continuous on [𝑎, 𝑏], then [𝑓(𝑥)]2 ≥ 0 for
all 𝑥 ∈ [𝑎, 𝑏] Therefore 〈𝑓, 𝑓〉 ≥ 0.
𝑏
(5)〈𝑓, 𝑓〉 = 0 ⇔ ∫𝑎 [𝑓(𝑥)]2 𝑑𝑥 = 0 ⟺ 𝑓 (𝑥 ) = 𝟎.

Thus 𝐶 [𝑎, 𝑏] is an inner product space.◄

Theorem 4.

Let V be an inner product space and 𝒖, 𝒗, 𝒘 ∈ 𝑉,

𝛼 ∈ ℝ. Then

(1) 〈𝟎, 𝒗〉 = 〈𝒗, 𝟎〉 = 0;

(2) 〈𝒖, 𝒗 + 𝒘〉 = 〈𝒖, 𝒗〉 + 〈𝒖, 𝒘〉;

(3) 〈𝒖, 𝛼𝒗〉 = 𝛼 〈𝒖, 𝒗〉;

(4) 〈𝒖 − 𝒗, 𝒘〉 = 〈𝒖, 𝒘〉 − 〈𝒗, 𝒘〉;

(5) 〈𝒖, 𝒗 − 𝒘〉 = 〈𝒖, 𝒗〉 − 〈𝒖, 𝒘〉.

Proof.

(1) Note that〈𝟎, 𝒗〉 = 〈𝟎 + 𝟎, 𝒗〉 = 〈𝟎, 𝒗〉 + 〈𝟎, 𝒗〉.

Then〈𝟎, 𝒗〉 + (−〈𝟎, 𝒗〉) = 〈𝟎, 𝒗〉.Therefore 〈𝟎, 𝒗〉 = 𝟎.

(2) 〈𝒖, 𝒗 + 𝒘〉 = 〈𝒗 + 𝒘, 𝒖〉 = 〈𝒗, 𝒖〉 + 〈𝒘, 𝒖〉

= 〈𝒖, 𝒗〉 + 〈𝒖, 𝒘〉.

- 164 -
(3) 〈𝒖, 𝛼𝒗〉 = 〈𝛼𝒗, 𝒖〉 = 𝛼 〈𝒗, 𝒖〉 = 𝛼〈𝒖, 𝒗〉.

(4) 〈𝒖 − 𝒗, 𝒘〉 = 〈𝒖 + (−𝒗), 𝒘〉

= 〈𝒖, 𝒘〉 + 〈(−𝒗), 𝒘〉

= 〈𝒖, 𝒘〉 − 〈𝒗, 𝒘〉.

(5) 〈𝒖, 𝒗 − 𝒘〉 = 〈𝒗 − 𝒘, 𝒖〉

= 〈𝒗, 𝒖〉 − 〈𝒘, 𝒖〉

= 〈𝒖, 𝒗〉 − 〈𝒖, 𝒘〉 .◄

- 165 -
 Exercise Set (5.1)

In Exercise 1-8 determine which of the given functions

are inner products on ℝ3 , where𝒖 = (𝑎1 , 𝑎2 , 𝑎3 ) and𝒗 =
(𝑏1 , 𝑏2 , 𝑏3 ).
(1)〈𝒖, 𝒗〉 = 𝑎1 𝑏2 + 𝑎2 𝑏1 .
(2)〈𝒖, 𝒗〉 = 2𝑎1 𝑏1 + 5𝑎2 𝑏2 + 𝑎3 𝑏3 − 𝑎2 𝑏3 − 𝑎3 𝑏2 .
(3)〈𝒖, 𝒗〉 = 𝑎1 𝑏1 + 𝑎2 𝑏2 + 𝑎3 𝑏3 + 𝑎1 𝑎2 𝑏3 .
(4)〈𝒖, 𝒗〉 = 𝑎1 𝑏1 + 𝑎2 𝑏2 .
(5)〈𝒖, 𝒗〉 = 𝑎1 + 𝑏1 .
(6)〈𝒖, 𝒗〉 = 𝑎12 𝑏12 + 𝑎22 𝑏22 + 𝑎32 𝑏32 .
(7)〈𝒖, 𝒗〉 = 𝑎12 𝑏1 + 𝑎2 𝑏2 + 𝑎3 𝑏3 .
(8)〈𝒖, 𝒗〉 = 𝑎1 𝑏1 − 𝑎2 𝑏2 + 𝑎3 𝑏3 .
(9) Let 𝑝, 𝑞 ∈ 𝑃2 . Prove that

〈𝑝, 𝑞 〉 = 𝑝(−1)𝑞 (−1) + 𝑝(0)𝑞 (0) + 𝑝(1)𝑞(1) is an

inner product function on 𝑃2 .

(10) Let 𝑝(𝑥 ) = 𝑎0 + 𝑎1 𝑥 + 𝑎2 𝑥 2 , 𝑞 (𝑥 ) = 𝑏0 + 𝑏1 𝑥 +

𝑏2 𝑥 2 ∈ 𝑃2 . Is〈𝑝, 𝑞 〉 = 𝑎0 𝑏0 − 𝑎1 𝑏1 + 𝑎2 𝑏2an inner
productfunction on 𝑃2 .

(11) Let V be an inner product space and 𝒖, 𝒗, 𝒘 ∈ 𝑉.

Simplify the following:

- 166 -
〈𝒖 + 𝒗, 𝒗 + 𝒘〉, 〈2𝒗 − 𝒘, 3𝒖 + 2𝒘〉, 〈𝒖 − 𝒗 + 4𝒘, 3𝒖 − 𝒗〉.

(12) Let 𝑓, 𝑔 ∈ 𝐷[𝑎, 𝑏] (The linear space of all

differential function on [𝑎, 𝑏]). Is 〈𝒇, 𝒈〉 = 𝑓 ′ (𝑥 ) + 𝑔′ (𝑥)
is an inner product function on 𝐷[𝑎, 𝑏].

(13) Let 𝑓, 𝑔 ∈ 𝐶 [0,1]. Is 〈𝒇, 𝒈〉 is an inner product on

𝐶 [0,1].

(14) Let 𝑝, 𝑞 ∈ 𝑃3 . Is 〈𝒑, 𝒒〉 is an inner product function

on 𝑃3 .

(15) Let 𝐴, 𝐵 ∈ 𝑀𝑛×𝑛 . Is 〈𝐴, 𝐵〉 = 𝑡𝑟 (𝐴𝐵) is an inner

productfunction on 𝑀𝑛×𝑛 .

(16) Let 𝐴, 𝐵 ∈ 𝑀𝑛×𝑛 . Is 〈𝐴, 𝐵〉 = 𝑡𝑟 (𝐴𝑡 𝐵)is an inner

product function on 𝑀𝑛×𝑛 .

(17) Let A be 𝑛 × 𝑛non singular matrix and 𝒖, 𝒗 ∈

ℝ𝑛 Prove that 〈𝒖, 𝒗〉 = (𝐴𝒖)𝑡 (𝐴𝒗) is an inner product
function on ℝ𝑛 .

(18) Let 𝐴 = [𝑎𝑎1 𝑎𝑎2 ] , 𝐵 = [𝑏𝑏1 𝑏𝑏2 ] ∈ 𝑀2×2 and

3 4 3 4

〈𝐴, 𝐵〉 = 𝑎1 𝑏1 + 𝑎2 𝑏2 + 𝑎3 𝑏3 + 𝑎4 𝑏4 .

Prove that 𝑀2×2 is an inner product space.

- 167 -
2.Orthogonally

In this section we shall define the notion of an angle

between two vectors in an inner product space, and we
shall use this concept to obtain basic relations between
vectors in an inner product and fundamental geometric
concepts in ℝ2 .

Definition 1.

Let u be an element of an inner product space V. The

norm‖𝒖‖ of V is defined by ‖𝒖‖ = √〈𝒖, 𝒖〉.

If w is another element in V, then we define the distance

between u and w denoted by (𝒖, 𝒘) , as follows:

𝑑(𝒖, 𝒘) = ‖𝒖 − 𝒘‖ = √〈𝒖 − 𝒘, 𝒖 − 𝒘〉

Example 1.

Let 𝒖 = (𝑎1 , … , 𝑎𝑛 ) , 𝒘 = (𝑏1 , … , 𝑏𝑛 ) ∈ ℝ𝑛 and the inner

product defined on ℝ𝑛 be the Euclidean inner
product.Then‖𝒖‖ = √𝑎12 + ⋯ + 𝑎𝑛2 and

𝑑(𝒖, 𝒘) = √(𝑎1 − 𝑏1 )2 + ⋯ + (𝑎𝑛 − 𝑏𝑛 )2

- 168 -
In particular, if 𝑛 = 3, 𝒖 = (1, 2, 3) and 𝒘 =
(−1, −2 ,4), then‖𝒖‖ = √(1)2 + (2)2 + (3)2 =

√14and𝑑(𝒖, 𝒘) = √(1 + 1)2 + (2 + 2)2 + (3 − 4)2 =

√21. ◄

Example 2.

Let 𝒖 = (𝑎1 , 𝑎2 ) , 𝒗 = (𝑏1 , 𝑏2 ) ∈ ℝ2 and the inner

product function defined on ℝ2 is 〈𝒖, 𝒗〉 = 7𝑎1 𝑏1 +
5𝑎2 𝑏2 . Then‖𝒖‖ = √7𝑎12 + 5𝑎22 . Also,

𝑑 (𝒖, 𝒘) = √7(𝑎1 − 𝑏1 )2 + 5(𝑎2 − 𝑏2 )2 . ◄

Example 3.

Let 𝑝(𝑥 ) = 2 − 𝑥 + 𝑥 2 , 𝑞 (𝑥 ) = 1 + 2𝑥 − 𝑥 2 ∈ 𝑃2 and

the inner product be as defined in Example 3 in Section 1
of this chapter. Then

‖𝑝(𝑥 )‖ = √〈𝑝(𝑥), 𝑝(𝑥 )〉

= √(2)2 + (−1)2 + (1)2 = √6.

𝑑(𝑝(𝑥 ), 𝑞 (𝑥 )) = √〈𝑝(𝑥) − 𝑞 (𝑥 ), 𝑝(𝑥 ) − 𝑞(𝑥)〉

= √(1)2 + (−3)2 + (2)2 = √14. ◄

- 169 -
Example 4.

Let 𝑓(𝑥 ) = cos 𝑥 , 𝑔(𝑥 ) = sin 𝑥 ∈ 𝐶 [0, 𝜋] and the inner

product on 𝐶 [0, 𝜋]be as defined in Example 4 of Section
1 in this chapter. Then
1 1
𝜋 2 𝜋 2
𝜋
‖𝑓(𝑥)‖ = (∫[𝑓(𝑥)]2 𝑑𝑥 ) = (∫ cos2 𝑥𝑑𝑥 ) = √ .
2
0 0

1
𝜋 2
Also,𝑑(𝑓(𝑥 ), 𝑔(𝑥 )) = (∫0 (cos 𝑥 − sin 𝑥) 𝑑𝑥 ) = √𝜋.◄
2

Theorem 1. (Cauchy – Schwartz Inequality).

If u and v are any vectors of an inner product space,

then|〈𝒖, 𝒗〉| ≤ ‖𝒖‖‖𝒗‖.

Proof.

If 𝒖 = 𝟎, then the result holds. Now, Let 𝒖 ≠ 𝟎 and

〈𝒗, 𝒖〉
𝒘=𝒗− 𝒖.
‖𝒖‖2

Therefore,

〈𝒗, 𝒖〉 〈𝒗, 𝒖〉
𝟎 ≤ ‖𝒘‖2 = 〈𝒘, 𝒘〉 = 〈𝒗 − 𝒖, 𝒗 − 𝒖〉
‖𝒖‖2 ‖𝒖‖2

- 170 -
 〈𝒗, 𝒖〉 〈𝒗, 𝒖〉 〈𝒗, 𝒖〉2
= 〈𝒗, 𝒗〉 − 〈𝒗, 𝒖〉 − 〈 𝒖, 𝒗〉 + 〈𝒖, 𝒖〉
‖𝒖‖2 ‖𝒖‖2 ‖𝒖‖4

〈𝒗, 𝒖〉2 〈𝒗, 𝒖〉2 〈𝒗, 𝒖〉2

= ‖𝒗 ‖2 − − +
‖𝒖‖2 ‖𝒖‖2 ‖𝒖‖2
〈𝒗,𝒖〉2
= ‖𝒗‖2 − ‖𝒖‖2
.

Then ‖𝒖‖2 ‖𝒗‖2 − 〈𝒖, 𝒗〉2 ≥ 0 or 〈𝒖, 𝒗〉2 ≤ ‖𝒖‖2 ‖𝒗‖2 .

Thus|〈𝒖, 𝒗〉| ≤ ‖𝒖‖‖𝒗‖. ◄

Remarks.
(1) Applying the Caushy-Schwarz inequality to ℝ𝑛 we
obtain the inequality:
(𝑎1 𝑏1 + ⋯ + 𝑎𝑛 𝑏𝑛 )2 ≤ (𝑎12 + ⋯ + 𝑎𝑛2 )(𝑏12 + ⋯ + 𝑏𝑛2 )
(2) Applying the Caushy-Schwarz inequality to 𝐶 [𝑎, 𝑏]
we obtain the inequality:
𝑏 2
(∫ 𝑓 (𝑥 )𝑔(𝑥 )𝑑𝑥 )
𝑎
𝑏 𝑏
2 2
≤ (∫ (𝑓(𝑥 )) 𝑑𝑥 ) (∫ (𝑔(𝑥 )) 𝑑𝑥 )
𝑎 𝑎

- 171 -
Lemma 1.

|〈𝒖, 𝒗〉| = ‖𝒖‖‖𝒗‖ if and only if u and v be linearly

dependent.

Proof.
〈𝒗,𝒖〉
Let |〈𝒖, 𝒗〉| = ‖𝒖‖‖𝒗‖ and 𝒘 = 𝒗 − ‖𝒖‖2 𝒖. Then
〈𝒗,𝒖〉
〈𝒘, 𝒘〉 = 𝟎. Therefore 𝒘 = 𝟎. Hence𝒗 = 𝒖.
‖𝒖‖2

So u and v are linearly dependent.

Conversely, let u and v be linearly dependent. Hence =
𝛼𝒖 , where 𝛼 ∈ ℝ. Then
〈𝒖, 𝒗〉2 = 〈𝒖, 𝛼𝒖〉2 = 𝛼 2 〈𝒖, 𝒖〉2 = 𝛼 2 〈𝒖, 𝒖〉〈𝒖, 𝒖〉
= 〈𝒖, 𝒖〉〈𝛼𝒖, 𝛼𝒖〉
= 〈𝒖, 𝒖〉〈𝒗, 𝒗〉 = ‖𝒖‖2 ‖𝒗‖2 .
Therefore |〈𝒖, 𝒗〉| = ‖𝒖‖‖𝒗‖.◄
The following Theorem lists some of the basic properties
ofany norm.
Theorem 1.
If V is an inner product space, then for every u and vin
Vand every scalar 𝛼 ∈ ℝ.
(1)‖𝒖‖ ≥ 0 ;

- 172 -
(2)‖𝒖‖ = 0 if and only if 𝒖 = 𝟎;
(3)‖𝛼𝒖‖ = |𝛼 |‖𝒖‖;
(4)‖𝒖 + 𝒗‖ ≤ ‖𝒖‖ + ‖𝒗‖ (Triangular inequality).
Proof.
(1) Since 〈𝒖, 𝒖〉 ≥ 0 ,‖𝒖‖ = √〈𝒖, 𝒖〉 ≥ 0.
(2) Since 〈𝒖, 𝒖〉 = 0 if and only if 𝒖 = 𝟎 . Then‖𝒖‖ =
𝟎if and only if 𝒖 = 𝟎.
(3) ‖𝛼𝒖‖2 = 〈𝛼𝒖, 𝛼𝒖〉
= 𝛼 2 〈𝒖, 𝒖〉 = 𝛼 2 ‖𝒖‖2 .
Therefore ‖𝛼𝒖‖ = 𝛼 ‖𝒖‖.
(4) ‖𝒖 + 𝒗‖2 = 〈𝒖 + 𝒗, 𝒖 + 𝒗〉
= 〈𝒖, 𝒖〉 + 2〈𝒖, 𝒗〉 + 〈𝒗, 𝒗〉
≤ 〈𝒖, 𝒖〉 + 2|〈𝒖, 𝒗〉| + 〈𝒗, 𝒗〉
≤ ‖𝒖‖2 + 2‖𝒖‖‖𝒗‖ + ‖𝒗‖2
= (‖𝒖‖ + ‖𝒗‖)2 .
Therefore‖𝒖 + 𝒗‖ ≤ ‖𝒖‖ + ‖𝒗‖. ◄
Lemma 2.
Let V be an inner product space,𝒖, 𝒗, 𝒘 ∈ 𝑉 and
𝛼 ∈ ℝ.Then
(1) 𝑑(𝒖, 𝒗) ≥ 0 ;

- 173 -
(2) 𝑑 (𝒖, 𝒗) = 0 if and only if 𝒖 = 𝒗 ;
(3) 𝑑 (𝒖, 𝒗) = 𝑑(𝒗, 𝒖)
(4) 𝑑 (𝒖, 𝒗) ≤ 𝑑(𝒖, 𝒘) + 𝑑(𝒘, 𝒗)(Triangular
inequality).

Proof.
(1) 𝑑 (𝒖, 𝒗) = ‖𝒖 − 𝒗‖ ≥ 0
(2) 𝑑 (𝒖, 𝒗) = 0 ⇔ ‖𝒖 − 𝒗‖ = 0 ⟺ 𝒖 − 𝒗 = 𝟎 ⟺
𝒖 = 𝒗.
(3) 𝑑 (𝒖, 𝒗) = ‖𝒖 − 𝒗‖ = ‖𝒗 − 𝒖‖ = 𝑑(𝒗, 𝒖).
(4) 𝑑 (𝒖, 𝒗) = ‖𝒖 − 𝒗‖ = ‖𝒖 − 𝒘 + 𝒘 − 𝒗‖
≤ ‖𝒖 − 𝒘‖ + ‖𝒘 − 𝒗‖
= 𝑑 (𝒖, 𝒘) + 𝑑(𝒘, 𝒗).◄
♠Now we show how to use Cauchy – Schwarz inequality
in defining the angle between two vectors in an inner
product space. Let V be an inner product space and 𝒖, 𝒗 ∈
𝑉. From Cauchy Schwarz inequality we have
|〈𝒖, 𝒗〉| ≤ ‖𝒖‖‖𝒗‖ ⟺
|〈𝒖, 𝒗〉|
≤1⟺
‖𝒖‖‖𝒗‖

- 174 -
 〈𝒖, 𝒗〉
−1 ≤ ≤1
‖𝒖‖‖𝒗‖
Therefore, there exists unique angle 𝜃 such that
〈𝒖, 𝒗〉
cos 𝜃 =
‖𝒖‖‖𝒗‖
where 0 ≤ 𝜃 ≤ 𝜋.
Definition 2.
Let V be an inner product space and 𝒖and𝒗 be anonzero
vectors in V. Then
(1) The cosine of 𝜃 between u and v is

〈𝒖, 𝒗〉
cos 𝜃 =
‖𝒖‖‖𝒗‖
(2) The vectors u and v are orthogonal if 〈𝒖, 𝒗〉 = 0.
Example 5.
Let ℝ2 be the Euclidean inner product space. Then the
angle between 𝒖 = (1, 2)and 𝒗 = (−1, 3)satisfies the
(1)(−1)+(2)(3) 1
equationcos 𝜃 = = .
√1+4√1+9 √2

Furthermore, the vectors 𝒖 = (1, 2) and 𝒗 = (−2, 1)

are orthogonal as 〈𝒖, 𝒗〉 = (−2)(1) + (1)(2) = 0.

- 175 -
But 𝒖 = (1, 2) and 𝒗 = (−1, 1) are not orthogonal if the
inner product on ℝ2 is the inner product defined in
Example 2,in Section 2 of Chapter (IV), where
〈𝒖, 𝒗〉 = 7(1)(−2) + 5(2)(1) ≠ 0.◄
Example 6.
Let 𝐴 = [20 −4
0 0 1
] , 𝐵 = [−7 0
] ∈ 𝑀2×2 and the inner product
is as defined in Exercise (5.1)-(18). Then A and B are
orthogonal as
〈𝐴, 𝐵〉 = (2)(0) + (0)(1) + 0(−7) + (−4)(0) = 0.◄
Example 7.
The vectors sin 𝑥 , cos 𝑥 ∈ 𝐶 [0, 𝜋]is orthogonal, where
the inner product on 𝐶 [0, 𝜋] is defined in Example 4 of
Section 1 in this Chapter as
𝜋 1 𝜋
〈sin 𝑥 , cos 𝑥 〉 = 2
∫0 sin 𝑥 cos 𝑥 𝑑𝑥 = [ sin 𝑥] .◄
2 0

Example 8.
Let 𝑝(𝑥 ) = 2 − 𝑥 + 4𝑥 2 , 𝑞 (𝑥 ) = 3 + 5𝑥 − 𝑥 2 ∈ 𝑃2 ,
where theinner product is defined in Example 3 of
Section 1 in this Chapter. The angle 𝜃 between𝑝(𝑥) and
𝑞(𝑥) satisfies the following equation:
6−5−4 −3
cos 𝜃 = = .◄
√4+1+16√9+25+1 √735

- 176 -
Theorem 2.
Let u and v be any vectors of an inner product space V.
Then u and v are orthogonal if and only if
(‖𝒖 + 𝒗‖2 = ‖𝒖‖2 + ‖𝒗‖2 ) (*)
Proof.
‖𝒖 + 𝒗‖2 = 〈𝒖 + 𝒗, 𝒖 + 𝒗〉
= 〈𝒖, 𝒖〉 + 2〈𝒖, 𝒗〉 + 〈𝒗, 𝒗〉
= ‖𝒖‖2 + 2〈𝒖, 𝒗〉 + ‖𝒗‖2
Thus the identity (*) holds if and only if 〈𝒖, 𝒗〉 = 0, i.e.,
ifand only if u and v are orthogonal.◄
Exercise Set (5.2)
In Exercise (1) – (5) determine the angle between the
given vectors and whose vectors orthogonal with respect
to the Euclidean inner produce on ℝ3 .
(1)𝒖 = (−1, 3, 2), 𝒗 = (4, 2, −1);
(2) 𝒖 = (−2, −2, −2), 𝒗 = (−1, 1, −1);
(3) 𝒖 = (−1, 1, 0), 𝒗 = (4, 0, 9);
(4) 𝒖 = (−1, 5, 2), 𝒗 = (2, 4, −9);
(5)𝒖 = (8, −4, −2), 𝒗 = (−4, 2, 1);
(6) Do the Exercise (1) – (5) if the inner product on ℝ3
is:〈𝒖, 𝒗〉 = 7𝑎1 𝑏1 + 3𝑎2 𝑏2 + 4𝑎3 𝑏3 .

- 177 -
In Exercise (7) – (11) determine the angle between the
given elements and whose are orthogonal with respect to
the Euclidean inner product ℝ4 .
(7)𝒖 = (−4, 0, −10, 1), 𝒗 = (2, 1, −2, 9);
(8)𝒖 = (−1, 1, 0, 2), 𝒗 = (1, −1, 3, 0);
(9)𝒖 = (0, −2, 2, 1), 𝒗 = (−1, −1, 1, 1);
(10)𝒖 = (1, 2, 0, 1), 𝒗 = (−2, 3, 1, 2);
(11)𝒖 = (−2, 3, 1, 4), 𝒗 = (−3, 1, 4, 6).
(12) Do the Exercise (7) – (11) if the inner product on
ℝ4 isdefined as
〈𝒖, 𝒗〉 = 𝑎1 𝑏1 + 𝑎2 𝑏2 + 2𝑎3 𝑏3 + 3𝑎4 𝑏4 .
In Exercise (13) – (17) Determine the angle 𝜃between
the given vectors and whose are orthogonal where
theinner product on 𝑃2 is defined in Example 3of Section
1 in this Chapter
(13) 𝑝(𝑥 ) = 𝑥 2 + 2𝑥 + 3 , 𝑞 (𝑥 ) = 𝑥 2 + 1 ;
(14) 𝑝(𝑥 ) = 2𝑥 2 + 𝑥 − 3 , 𝑞 (𝑥 ) = 4𝑥 2 − 𝑥 ;
(15) 𝑝(𝑥 ) = 2𝑥 2 + 5𝑥 − 1 , 𝑞 (𝑥 ) = 9𝑥 2 + 2𝑥 + 1 ;
(16) 𝑝(𝑥 ) = 2𝑥 2 − 𝑥 + 1 , 𝑞 (𝑥 ) = 𝑥 2 + 2𝑥 ;
(17) 𝑝(𝑥 ) = 2𝑥 2 − 𝑥 + 1 , 𝑞 (𝑥 ) = 𝑥 2 + 2𝑥 ;

- 178 -
In Exercise (18) – (22) Find the angle 𝜃 between the given
vectors and determine whose are orthogonal, where the inner
product on 𝑀2×2 is defined in Exercise (5.1)-(18).
(18)= [21 −3
6
] , 𝐵 = [01 23] ;

(19) 𝐴 = [−1
6 1
2
] , 𝐵 = [33 01] ;

(20) 𝐴 = [13 24] , 𝐵 = [51 −3

2
]

(21) = [52 −4
6
] , 𝐵 = [−2 −1
1 −5
];

(22) 𝐴 = [10 00] , 𝐵 = [03 −2

5
].
(23) Prove thatsin 3𝑥, cos 4𝑥 ∈ 𝐶 [0, 𝜋] are orthogonal,
where the inner product defined in Example 4of Section
1 in this Chapter.
(24) Let V be an inner product space and , 𝒗, 𝒘 ∈ 𝑉 . If u
is orthogonal to v and w, prove that u and 𝛼𝒗 + 𝛽𝒘are
orthogonal for every 𝛼, 𝛽 ∈ ℝ .
(25) Let V be an inner product space and 𝒖, 𝒗 ∈ 𝑉.Prove
(1)‖𝒖 + 𝒗‖2 − ‖𝒖 − 𝒗‖2 = 4〈𝒖, 𝒗〉
(2) ‖𝒖 + 𝒗‖2 + ‖𝒖 − 𝒗‖2 = 2‖𝒖‖2 + 2‖𝒗‖2
(3) ‖𝒖 + 𝒗‖2 = ‖𝒖‖2 + ‖𝒗‖2 + 2‖𝒖‖‖𝒗‖ cos 𝜃
Where 𝜃 is the angle between u and v.
(4)𝒖 − 𝒗 and 𝒖 + 𝒗 are orthogonal when ‖𝒖‖ = ‖𝒗‖.
(5) ‖𝒖‖ − ‖𝒗‖ ≤ ‖𝒖 − 𝒗‖.

- 179 -
3. Orthonormal Basis
In many problems involving vector spaces, the problem
solver is free to choose any basis for the vector space that
seems appropriate. In inner product spaces the solution
of a problem is often greatly simplified by choosing a
basis in which the vectors are orthogonal to one another.
In this section we shall show how such bases can be
obtained.
Definition 1.
A set of vectors in an inner product space is called an
orthogonal set if all pairs of distinct vectors in the set are
orthogonal. An orthogonal set in which each vector has
norm 1 is called orthonormal.
Example 1.
1 1
Let 𝒖1 = (1, 1, 0, 1) , 𝒖2 = (1, −2, 0, 1)and
√3 √6
1
𝒖3 = (−1,0,2,1)be three vectors in the Euclidean
√6

inner product space ℝ4 . It follows that the set of vectors

𝑆 = {𝒖1 , 𝒖2 , 𝒖3 } is orthonormal since〈𝒖1 , 𝒖2 〉 =
〈𝒖2 , 𝒖3 〉 = 〈𝒖1 , 𝒖3 〉 = 0and also,‖𝒖1 ‖ = ‖𝒖2 ‖ =
‖𝒖3 ‖ = 1.◄

- 180 -
Example 2.
1 3 3 2 1
Let𝑆 = { , √ 𝑥, √ (𝑥 2 − )} ⊆ 𝐶 [−1,1], where the
√ 2 2 2 5 3

inner product in 𝐶 [−1,1] is defined in Example 4 of

Section 1 in this chapter. Then S is orthonormal set since
1 1
1 3 √3 √3
〈 , √ 𝑥〉 = ∫ 𝑥 𝑑𝑥 = [ 𝑥 2 ] = 0 ;
√2 2 2 4 −1
−1
1
1 3 2 1 3 1
〈 , √ (𝑥 2 − )〉 = ∫ (𝑥 2 − ) 𝑑𝑥 = 0 ;
√2 2 5 3 2√5 3
−1
1
√3 3 2 1 3√6 3 1
〈 𝑥, √ (𝑥 2 − )〉 = ∫ (𝑥 − 𝑥) 𝑑𝑥 = 0 ;
2 2 5 3 4√5 3
−1

Also, we can show that

1 3 3 2 1
‖ ‖ = ‖√ 𝑥‖ = ‖ √ (𝑥 2 − )‖ = 1 .◄
√2 2 2 5 3

Remark.
If V is a nonzero vector in an inner product space, then
1 1 1 1
‖𝒗‖
𝒗 has norm 1, as ‖‖𝒗‖ 𝒗‖ = |‖𝒗‖| ‖𝒗‖ = ‖𝒗‖ ‖𝒗‖ = 1.

The process of multiplying nonzero vector v by the

reciprocal of its length to obtain a vector of norm 1 is

- 181 -
called normalizing v. An orthogonal set of nonzero
vectors can always be converted to an orthonormal set by
normalizing each of its vectors.
Example 3.
Let 𝒖1 = (0,1,0), 𝒖2 = (1,0,1), 𝒖3 = (1,0, −1)and
assume that ℝ3 has the Euclidean inner product. It
follows that the set of vectors 𝑆 = {𝒖1 , 𝒖2 , 𝒖3 } is
orthogonal since 〈𝒖1 , 𝒖2 〉 = 〈𝒖1 , 𝒖3 〉 = 〈𝒖2 , 𝒖3 〉 =
0.Also the Euclidean norms of the vectors are
‖𝒖1 ‖ = 1, ‖𝒖2 ‖ = √2, ‖𝒖3 ‖ = √2
Consequently, normalizing 𝒖1 , 𝒖2 , and 𝒖3 yields
𝒖 𝒖 1 1
𝒗1 = ‖𝒖1 ‖ = (0,1,0) , 𝒗2 = ‖𝒖2 ‖ = ( , 0, ),
1 2 √2 √2
𝒖 1 −1
𝒗3 = ‖𝒖3 ‖ = ( , 0, ). We leave it for you to verify that
3 √2 √2

the set 𝑆 = {𝒗1 , 𝒗2 , 𝒗3 }is orthonormal by showing that

〈𝒗1 , 𝒗2 〉 = 〈𝒗1 , 𝒗3 〉 = 〈𝒗2 , 𝒗3 〉 = 0 and ‖𝒗1 ‖ = ‖𝒗2 ‖ =
‖𝒗3 ‖ = 1.◄
Theorem 1.
If 𝑆 = {𝒗1 , … , 𝒗𝑛 } is an orthogonal set of nonzero
vectors in an inner product space, then S is linearly
independent.
- 182 -
Proof:
Assume that
k1 𝒗1 + ⋯ + k 𝑛 𝒗𝑛 = 𝟎. (*)
To demonstrate that 𝑆 = {𝒗1 , … , 𝒗𝑛 }is linearly
independent,we must prove that k1 = ⋯ = k 𝑛 = 0.
For each 𝒗𝑖 in S, it follows from (*) that
〈k1 𝒗1 + ⋯ + k 𝑛 𝒗𝑛 , 𝒗𝑖 〉 = 〈𝟎, 𝒗𝑖 〉 = 0 .
Or equivalently,k1 〈𝒗1 , 𝒗𝑖 〉 + ⋯ + k 𝑛 〈𝒗𝑛 , 𝒗𝑖 〉 = 0 .
From the orthogonality of S it follows that〈𝒗𝑗 , 𝒗𝑖 〉 =
0 .when ≠ 𝑖, so that this equation reduces to
k 𝑖 〈𝒗𝑖 , 𝒗𝑖 〉 = 0
Since the vectors in S are assumed to be nonzero,
〈𝒗𝑖 , 𝒗𝑖 〉 ≠ 0by the positivity axiom for inner product.
Therefore k 𝑖 = 0. Since the subscript I is arbitrary, we
have k1 = ⋯ = k 𝑛 = 0. Thus S is linearly
independent.◄
● In an inner product space, a basis consisting of
orthonormal vectors is called an orthonormal basis, and
a basis consisting of orthogonal vectors is called
orthogonal basis.

- 183 -
Theorem 2.
If 𝑆 = {𝒗1 , … , 𝒗𝑛 }is an orthonormal basis for an inner
product space V, and u is any vector in V, then
𝒖 = 〈𝒖, 𝒗1 〉𝒗1 + 〈𝒖, 𝒗2 〉𝒗2 + ⋯ + 〈𝒖, 𝒗𝑛 〉𝒗𝑛 .
Proof.
Since 𝑆 = {𝒗1 , … , 𝒗𝑛 } is a basis, a vector u can be
expressed in the form𝒖 = k1 𝒗1 + ⋯ + k 𝑛 𝒗𝑛 .
We shall complete the proof by showing that k 𝑖 =
〈𝒖, 𝒗𝑖 〉for = 1, … , 𝑛 . For each vector 𝒗𝑖 in S we have
〈𝒖, 𝒗𝑖 〉 = 〈k1 𝒗1 + ⋯ + k 𝑛 𝒗𝑛 , 𝒗𝑖 〉
= k1 〈𝒗1 , 𝒗𝑖 〉 + ⋯ + k 𝑛 〈𝒗𝑛 , 𝒗𝑖 〉.
Since 𝑆 = {𝒗1 , … , 𝒗𝑛 } is an orthonormal set, we have
〈𝒗𝑖 , 𝒗𝑖 〉 = ‖𝒗𝑖 ‖2 = 1 and 〈𝒗𝑗 , 𝒗𝑖 〉 = 0 if 𝑗 ≠ 𝑖.
Therefore, the above expression for 〈𝒖, 𝒗𝑖 〉 simplifies to
〈𝒖, 𝒗𝑖 〉 = k 𝑖 . The scalars 〈𝒖, 𝒗1 〉, 〈𝒖, 𝒗2 〉, … , 〈𝒖, 𝒗𝑛 〉 in the
above theorem are the coordinates of u relative to the
orthonormal basis 𝑆 = {𝒗1 , … , 𝒗𝑛 } and
(𝒖)𝑆 = (〈𝒖, 𝒗1 〉, 〈𝒖, 𝒗2 〉, … , 〈𝒖, 𝒗𝑛 〉)
is the coordinate vector of u relative to this basis.◄

- 184 -
Example 4.
4 3 3 4
Let 𝒗1 = (0,1,0), 𝒗2 = (− , 0, ) , 𝒗3 = ( , 0, ). It is
5 5 5 5

easy to check that𝑆 = {𝒗1 , 𝒗2 , 𝒗3 }is an orthonormal basis

for ℝ3 with the Euclidean inner product. Express 𝒖 =
(1, 1, 1) as a linear combination of the vectors in S, and
find the coordinate (𝒖)𝑆 .
Solution.
1 7
〈𝒖, 𝒗1 〉 = 1, 〈𝒖, 𝒗2 〉 = − and 〈𝒖, 𝒗3 〉 = .
5 5

Therefore by Theorem 2 in Section 3 of this chapter we

1 7
have𝒖 = 𝒗1 − 𝒗2 + 𝒗3 . That is,
5 5
1 4 3 7 3 4
(1,1,1) = (0,1,0) − (− , 0, ) + ( , 0, ).
5 5 5 5 5 5

The coordinate vector of u relative to S is

1 7
(𝒖)𝑠 = (〈𝒖, 𝒗1 〉, 〈𝒖, 𝒗2 〉, 〈𝒖, 𝒗3 〉) = (1, − , ).◄
5 5

●We have seen that orthonormal bases enjoy variety

useful properties. Our next theorem shows that every
nonzero finite dimensional vector space has an
orthonormal basis. The proof of this result provides an
algorithm (called Gram – Schmidt Process) for
converting an arbitrary basis into orthonormal basis.

- 185 -
Theorem 3 (Gram – Schmidt algorithm).
Every nonzero finite – dimensional inner product space
has an orthonormal basis.
Proof.
Let V be any nonzero finite – dimensional inner product
space, and let {𝒖1 , … , 𝒖𝑛 } be any basis for V. It suffice to
show that V has an orthogonal basis, since the vectors in
the orthogonal basis can by normalized to produce an
orthonormal basis for V. The following sequence of steps
will produce an orthogonal basis {𝒗1 , … , 𝒗𝑛 } for V. Let
Step 1.𝒗1 = 𝒖1
〈𝒖2 ,𝒗1 〉
Step 2.𝒗2 = 𝑢2 − ‖𝒗1 ‖2
𝒗1
〈𝒖3 ,𝒗1 〉 〈𝒖3 ,𝒗2 〉
Step 3.𝒗3 = 𝒖3 − ‖𝒗1 ‖2
𝒗1 − ‖𝒗2 ‖2
𝒗2

⋮
〈𝒖𝑛 , 𝒗𝑛−1 〉 〈𝒖𝑛 , 𝒗1 〉
𝒗 𝑛 = 𝒖𝑛 − 𝒗 − ⋯ − 𝒗
‖𝒗𝑛−1 ‖2 𝑛−1 ‖𝒗1 ‖2 1
Since {𝒖1 , … , 𝒖𝑛 } is linearly independence, then
{𝒗1 , … , 𝒗𝑛 } is nonzero.
Now, we proof by induction that{𝒗1 , … , 𝒗𝑛 } is
orthogonal.

- 186 -
Let = 2 , then
〈𝒖2 , 𝒗1 〉
〈𝒗2 , 𝒗1 〉 = 〈𝒖2 − 𝒗 ,𝒗 〉
‖𝒗1 ‖2 1 1
〈𝒖2 , 𝒗1 〉
= 〈𝒖2 , 𝒗1 〉 − 〈𝒗 , 𝒗 〉
‖𝒗1 ‖2 1 1
= 〈𝒖2 , 𝒗1 〉 − 〈𝒖2 , 𝒗1 〉 = 0.
Therefore 𝒗1 , 𝒗2 are orthogonal.
Now assume{𝒗1 , 𝒗2 , … , 𝒗𝑛−1 }is orthogonal, where 𝑛 > 2
. We prove that {𝒗1 , 𝒗2 , … , 𝒗𝑛 } is orthogonal.
For 𝑖 = 1,2, … , 𝑛 − 1,we have
〈𝒖𝑛 , 𝒗𝑛−1 〉
〈 𝒗 𝑛 , 𝒗 𝑖 〉 = 〈 𝒖𝑛 , 𝒗 𝑖 〉 − 〈𝒗𝑛−1 , 𝒗𝑖 〉 −
‖𝒗𝑛−1 ‖2
〈𝒖𝑛 , 𝒗1 〉
…− 〈𝒗1 , 𝒗𝑖 〉
‖𝒗1 ‖2
〈 𝒖𝑛 , 𝒗 𝑖 〉
= 〈 𝒖𝑛 , 𝒗 𝑖 〉 − 〈𝒗𝑖 , 𝒗𝑖 〉 = 0
‖𝒗𝑖 ‖2
Therefore {𝒗1 , … , 𝒗𝑛 } is orthogonal. From Theorem 1 in
Section 3 of Chapter (IV) {𝒗1 , … , 𝒗𝑛 } is linearly
independent. Hence {𝒗1 , … , 𝒗𝑛 } is an orthogonal basis
𝒗
for the inner product space V. Finally if 𝒘𝑖 = ‖𝒗𝑖‖ for 𝑖 =
𝑖

- 187 -
1, … , 𝑛, we obtain an orthonormal basis {𝒘1 , … , 𝒘𝑛 } for
the inner product space V.◄
Example 5.
Consider the vector space ℝ3 with the Euclidean inner
product. Apply the Gram – Schmidt process to transform
the basis vectors 𝒖1 = (1,1,1) , 𝒖2 = (0,1,1) , 𝒖3 =
(0,0,1)into an orthogonal basis {𝒗1 , 𝒗2 , 𝒗3 } ; then
normalize the orthogonalbasis vectors to obtain an
orthonormal basis {𝒒1 , 𝒒2 , 𝒒3 } .
Solution:
Step 1. 𝒗1 = 𝒖1 = (1,1,1)
〈𝒖2 ,𝒗1 〉
Step 2.𝒗2 = 𝒖2 − ‖𝒗1 ‖2
𝒗1
2 2 1 1
= (0,1,1) − (1,1,1) = (− , , ).
3 3 3 3
〈𝒖3 ,𝒗1 〉 〈𝒖3 ,𝒗2 〉
Step 3. 𝒗3 = 𝒖3 − ‖𝑣1 ‖2
𝒗1 − ‖𝒗2 ‖2
𝒗2

1 1 ⁄3 2 1 1
= (0,0,1) − (1,1,1) − (− , , )
3 2 ⁄3 3 3 3
1 1
= (0, − , ).
2 2
2 1 1 1 1
Thus 𝒗1 = (1,1,1) , 𝒗2 = (− , , ) , 𝒗3 = (0, − , )
3 3 3 2 2

form an orthogonal basis for ℝ3 . The norms of these

- 188 -
 √6 1
vectors are ‖𝒗1 ‖ = √3, ‖𝒗2 ‖ = , ‖𝒗3 ‖ = .So an
3 √2

orthonormal basis for ℝ3 is

𝒗1 1 1 1 𝒗2
𝒒1 = =( , , ) , 𝒒2 =
‖𝒗1 ‖ √3 √3 √3 ‖𝒗2 ‖
2 1 1
= (− , , )
√6 √6 √6
𝒗 −1 1
𝒒3 = ‖𝒗3 ‖ = (0, , ).◄
3 √2 √2

Example 6.
Consider the subspace W of ℝ5 spanned by {𝒗1 , 𝒗2 , 𝒗3 } ,
where𝒗1 = (−1, −1,1,0,0), 𝒗2 = (0, −1,0,0,1) , 𝒗3 =
(1, −1,0,1,0). Find an orthonormal basis for W, where the
inner product is the Euclidean product.
Solution.
Step 1:𝒖1 = 𝒗1 = (−1, −1,1,0,0).
〈𝒗2 ,𝒖1 〉
Step 2:𝒖2 = 𝒗2 − ‖𝒖1 ‖2
𝒖1 = (0, −1,0,0,1) −
1 1 −2 1
(−1, −1,1,0,0) = ( , , − , 0,1)
3 3 3 3
〈𝒗3 ,𝒗𝒖1 〉 〈𝒗3 ,𝒖2 〉 4 −3 1 3
Step 3.𝒖3 = 𝒗3 − ‖𝑢1 ‖2
𝒖1 − ‖𝒖2 ‖2
𝒖2 = ( , , , 1, )
5 5 5 5

Thus {𝒖1 , 𝒖2 , 𝒖3 } is orthogonal basis. Therefore the

orthonormal basis is

- 189 -
 1 1 1 1 −2 1 √15
(− ,− , , 0,0) , ( , ,− , 0, )
√3 √3 √3 √15 √15 √15 5
2 √15 √15 √15 √15
( ,− , , , − ). ◄
√15 10 30 6 10

Example 7.
Let 𝑝, 𝑞 ∈ 𝑃4 and the inner product defined on 𝑃4 is as
follows
〈𝑝, 𝑞 〉 = 𝑝(−2)𝑞 (−2) + 𝑝(−1)𝑞 (−1) + 𝑝(0)𝑞 (0) +
𝑝(1)𝑞 (1) + 𝑝(2)𝑞(2) and W be a subspace of 𝑃4
spanned by {𝑝1 (𝑥 ) = 1, 𝑝2 (𝑥 ) = 𝑥, 𝑝3 (𝑥) = 𝑥 2 }.
Find an orthogonal basis for W.
Solution.
Step 1. 𝑞1 = 𝑝1 = 1
Step 2.
〈𝑝2 , 𝑞1 〉 = 𝑝2 (−2)𝑞1 (−2) + 𝑝2 (−1)𝑞1 (−1)
+ 𝑝2 (0)𝑞1 (0) + 𝑝2 (1)𝑞1 (1) + 𝑝2 (2)𝑞1 (2)
=0
〈𝑝2 ,𝑞1 〉
Hence 𝑞2 = 𝑝2 − ‖𝑞1 ‖2
𝑞1 = 𝑥 − 0 = 𝑥
〈𝑝3 ,𝑞2 〉 〈𝑝3 ,𝑞1 〉
Step 3:𝑞3 = 𝑝3 − ‖𝑞2 ‖2
𝑞2 − ‖𝑞1 ‖2
𝑞1 ,

- 190 -
where 〈𝑝3 , 𝑞1 〉 = 10, ‖𝑞1 ‖2 = 〈𝑞1 , 𝑞1 〉 = 5, and
10
〈𝑝3 , 𝑞2 〉 = 0. Therefore, 𝑞3 = 𝑥 3 − 0 − (1) = 𝑥 2 − 2
5

Hence {1, 𝑥, 𝑥 2 − 2} is orthogonal basis for W.◄

Exercise Set (5.3)
(1)Let the vector space 𝑃2 have the inner product
1

〈𝑝, 𝑞 〉 = ∫ 𝑝(𝑥 )𝑞 (𝑥 )𝑑𝑥

−1

Apply the Gram – Schmidt process to transform the

standard basis 𝑆 = {1, 𝑥, 𝑥 2 } into orthonormal basis.
(2) Let ℝ3 be the inner product space, where the inner
function is 〈𝒖, 𝒗〉 = 𝑎1 𝑏1 + 2𝑎2 𝑏2 + 3𝑎3 𝑏3 .
(3) Apply the Gram – Schmidt process to transform the
basis {𝒗1 , 𝒗𝟐 , 𝒗3 } to orthonormal basis.
(𝑎)𝒗1 = (1,1,1) , 𝒗2 = (−1,1,0) , 𝒗3 = (1,2,1)
(𝑏)𝒗1 = (1,1,1) , 𝒗2 = (1,1,0) , 𝒗3 = (1,0,0)
(𝑐)𝒗1 = (1,0,0) , 𝒗2 = (3,7, −2) , 𝒗3 = (0,4,1).

- 191 -
Definition.
Let A be an 𝑛 × 𝑚 matrix. The subspace of ℝ𝑛 spanned
by the rows (columns) of A is called the row (column)
space.
(4) Apply the Gram – Schmidt process to find an
orthonormal basis for the column space of the matrix A,
where the inner product is the Euclidean
3 0 5 1 0 0
(𝑎) 𝐴 = [0 1 −2] , (𝑏) 𝐴 = [1 1 0]
4 1 1 1 1 1
(5) Find an orthonormal basis for the row space for the
matrix A in Exercise (4).

4. EigenValues and EigenVectors

In matrix algebra you will come across equations of the

form:
Scalar (number)
𝐴. 𝒙 = 𝜆𝒙
(Will be called eigenvalue)

Square matrix Vector

(Will be called eigenvector)

- 192 -
The equation can be rearranged as follows:
Rearrange: 𝐴𝒙 − 𝜆𝒙 = 0
Factorize: (𝐴 − 𝜆)𝒙 = 0
Make 𝜆 into a matrix:(𝐴 − 𝜆𝐼)𝒙 = 0
This form is useful for finding eigenvalues and
eigenvectors.
We can see from the last version of the equation that 𝒙 =
0 is a solution (trivial).
To find the eigenvalues (𝜆), matrix theory tells us that we
must set the determinant of 𝐴 − 𝜆𝐼 equal to 0.
i.e.|𝐴 − 𝜆𝐼| = 0
To find the eigenvectors, put the eigenvalues back into
the original equation and solve.
The characteristic polynomial for the matrix A is:
𝑝(𝜆) = |𝐴 − 𝜆𝐼|.
Some texts used the form (𝐴 − 𝜆𝐼) which implies ±𝜆𝑛 .
Example 1.
Find the eigenvalues and eigenvectors for the matrix𝐴 =
6 3
[ ].
−2 1

- 193 -
Solution.
Finding the characteristic polynomial
𝜆 0
We know that 𝜆𝐼 = [ ]. Thus
0 𝜆
6 3 𝜆 0 6−𝜆 3
𝐴 − 𝜆𝐼 = [ ]−[ ]=[ ]
−2 1 0 𝜆 −2 1−𝜆
Therefore
𝑝(𝜆) = |𝐴 − 𝜆𝐼|
= (6 − 𝜆)(1 − 𝜆) − 3(−2) =6−
7𝜆 + 𝜆2 + 6
= 𝜆2 − 7𝜆 + 12.

Finding the eigenvalues:

Put the characteristic polynomial equal to 0 and solve to
find that 𝜆 = 3 𝑜𝑟𝜆 = 4.

Finding the eigenvectors:

Going back to the very 1st equation we know that:
𝐴𝒙 = 𝜆𝒙
6 3 𝑥 𝑥
This means that: [ ] [𝑦 ] = 𝜆 [𝑦 ]
−2 1
Multiply out each side:
6𝑥 + 3𝑦 = 𝜆𝑥, −2𝑥 + 𝑦 = 𝜆𝑦
(The 2 eigenvalues we found were λ = 3 andλ = 4)
- 194 -
When 𝝀 = 𝟑 6𝑥 + 3𝑦 = 3𝑥; −2𝑥 + 𝑦 = 3𝑦
Rearranging either of these gives −𝒙 = 𝒚.
The numbers go in the vector the opposite way round
than you would think, i.e. the number in front of𝑥goes in
the𝑦position and vice versa.
1
Therefore, the vector is [ ] and the general
−1
𝑘
Eigenvector when 𝜆 = 3 is [ ].
−𝑘
When 𝝀 = 𝟒: 6𝑥 + 3𝑦 = 4𝑥; −2𝑥 + 𝑦 = 4𝑦
Rearranging either of these gives −2𝒙 = 𝟑𝒚.
Again, the number in front of 𝑥goes in the𝑦position and
vice versa.
3
Therefore, the vector is [ ] and the general
−2
3𝑘
eigenvector when 𝜆 = 4 is [ ].
−2𝑘
6 3
So, for the matrix 𝐴 = [ ]:
−2 1
𝑘
Eigenvalue = 3, associated eigenvector = [ ]
−𝑘
3𝑘
Eigenvalue = 4, associated eigenvector = [ ].◄
−2𝑘

- 195 -
Example 2.
Find the characteristic polynomial, eigenvalues and
1 2
eigenvectors for the matrix𝐴 = [ ].
4 3
Solution.
𝜆 0 1 2 𝜆−1 −2
𝜆𝐼 − 𝐴 = [ ]−[ ]=[ ].
𝜆 0 4 3 −4 𝜆−3
The characteristic polynomial is
𝜆−1 −2
𝑝(𝜆) = | | = (𝜆 − 1)(𝜆 − 3) − 8
−4 𝜆−3
= 𝜆2 − 4𝜆 − 5.
And the eigenvalues are 𝜆 = 5 𝑎𝑛𝑑 𝜆 = −1.
1 2 𝑥 𝑥
The eigenvector for 𝜆 = 5: [ ] [𝑦] = 5 [𝑦]yields the
4 3
equations 𝑥 + 2𝑦 = 5𝑥 and 4𝑥 + 3𝑦 = 𝑦.
Solving, we find that 𝑦 = 2𝑥; therefore the eigenvector is
1
[ ]. Using the same method, we find the eigenvector for
2
1
𝜆 = −1 is [ ].◄
−1

- 196 -
 ♣The eigenvalues for a matrix that is upper or lower
diagonal come from the diagonal. For the upper diagonal
1 1 1 1
0 1 1 1
matrix 𝐶 = [ ],
0 0 1 1
0 0 0 1
𝜆−1 1 1 1
0 𝜆−1 1 1
where (𝜆𝐼 − 𝐴) = [ ],
0 0 𝜆−1 1
0 0 0 𝜆−1
the determinant is (𝜆 − 1)4 , which yields the eigenvalue
𝜆 = 1.
Example 3.
0 2 1
Find the eigenvalues for the matrix 𝐴 = [1 0 0].
0 1 0
Solution.
we found that the characteristic polynomial 𝑝(𝜆) = 𝜆3 −
2𝜆 − 1 could not be factored. We discussed using either
synthetic division or long division to solve. Two of the
1+√5
eigenvalues looked VERY familiar, 𝜆 = = 𝜌 and
2
1
𝜆 = 1 − 𝜌 = − , and the third 𝜆 = −1.◄
𝜌

- 197 -
Theorem 1.
If 𝜆1 , … 𝜆𝑚 are mutually distinct eigenvalues of A with
eigenvectors 𝒙1 , … 𝒙𝑚 , then {𝒙1 , … 𝒙𝑚 } is a linearly
independent set.
Corollary 1.
If a 𝑛 × 𝑛 matrix A has n distinct eigenvalues, then A
has n linearly independent eigenvectors.
Theorem 2.
Let Abe an𝑛 × 𝑛matrix. Then A is invertible if and only
if the number 0 is not an eigenvalue of A.
Example 4:
6 8 4
𝐴 = [−2 −2 −2]
−1 −2 1
What are the eigenvalues and eigenvectors?
Solution. Outlines
1) Find the characteristic polynomial.
When it is zero the solution to it are the only possible
eigenvalues:
𝜆−6 −8 −4
|𝜆𝐼 − 𝐴| = | 2 𝜆−2 2 | = (𝜆 − 1)2 (𝜆 − 2)1
1 2 𝜆−1

- 198 -
has eigenvalues λ = 1, 1, 2.
Solve: 𝐴𝒙 = 1𝒙 and A𝒙=2𝒙
𝐴𝒙 = 1𝒙 has eigenspace of dimension 1.
A𝒙=2𝒙 has eigenspace of dimension 2.
Example 5.
Find the characteristic polynomial, eigenvalues and
eigenvectors for the matrix A, where
−𝟑 𝟏 −𝟏
𝑨 = (−𝟕 𝟓 −𝟏).
−𝟔 𝟔 −𝟐
Solution.
We have
−𝟑 − 𝝀 𝟏 −𝟏
𝑨 − 𝝀𝑰𝟑 = ( −𝟕 𝟓−𝝀 −𝟏 ).
−𝟔 𝟔 −𝟐 − 𝝀
and
|𝑨 − 𝝀𝑰𝟑 | = −𝝀𝟑 + 𝟏𝟐𝝀 + 𝟏𝟔.
In general, it is not easy to find the roots of a polynomial
of degree 3. We try some simple values and hope we will
be lucky. For 𝝀 = −𝟐, you get
|𝑨 − 𝝀𝑰𝟑 | = −(−𝟐)𝟑 + 𝟏𝟐(−𝟐) + 𝟏𝟔 = 𝟎,
so that 𝝀 = −𝟐 is a root of |𝑨 − 𝝀𝑰𝟑 |.
Hence (𝝀 + 𝟐) is a factor of |𝑨 − 𝝀𝑰𝟑 | and |𝑨 − 𝝀𝑰𝟑 | can
be written in the form

- 199 -
 |𝑨 − 𝝀𝑰𝟑 | = (𝝀 + 𝟐)(𝒂𝝀𝟐 + 𝒃𝝀 + 𝒄)
= 𝒂𝝀𝟑 + 𝝀𝟐 (𝒃 + 𝟐𝒂) + 𝝀(𝒄 + 𝟐𝒃) + 𝟐𝒄.
We now have a, band c must satisfy
𝒂 = −𝟏
𝒂 = −𝟏
𝒃 + 𝟐𝒂 = 𝟎
{ , which yields { 𝒃 = 𝟐 .
𝒄 + 𝟐𝒃 = 𝟏𝟐
𝒄=𝟖
𝟐𝒄 = 𝟏𝟔
We are now in position to find all the factors of |𝑨 −
𝝀𝑰𝟑 |:
|𝑨 − 𝝀𝑰𝟑 | = (𝝀 + 𝟐)(−𝝀𝟐 + 𝟐𝝀 + 𝟖)
= −(𝝀 + 𝟐)𝟐 (𝝀 − 𝟒).
Therefore, A has two eigenvalues: 𝝀𝟎 = −𝟐 and 𝝀𝟏 = 𝟒.
To get the eigenvectors associated with 𝝀𝟎 = −𝟐, we
solve (𝑨 + 𝟐𝑰𝟑 )𝒖 = 𝟎.
−𝟏 𝟏 −𝟏 𝒙𝟏 𝟎
(𝑨 + 𝟐𝑰𝟑 )𝒖 = (−𝟕 𝟕 −𝟏) (𝒙𝟐 ) = (𝟎).
−𝟔 𝟔 𝟎 𝒙𝟑 𝟎
−𝒙𝟏 +𝒙𝟐 −𝒙𝟑 = 𝟎
𝒙𝟏 = 𝒙 𝟐
⇔ {−𝟕𝒙𝟏 +𝟕𝒙𝟐 −𝒙𝟑 = 𝟎 ⇔ { 𝒙 = 𝟎 .
𝟑
−𝟔𝒙𝟏 +𝟔𝒙𝟐 =𝟎
Hence the eigenvectors associated with the eigenvalue
𝟏
𝝀𝟎 = −𝟐 are of the form 𝜶 (𝟏).
𝟎

- 200 -
To get the eigenvectors associated with 𝝀𝟏 = 𝟒, we solve
(𝑨 − 𝟒𝑰𝟑 )𝒖 = 𝟎:
−𝟕 𝟏 −𝟏 𝒙𝟏 𝟎
(𝑨 − 𝟒𝑰𝟑 )𝒖 = (−𝟕 𝟏 −𝟏) (𝒙𝟐 ) = (𝟎)
−𝟔 𝟔 −𝟔 𝒙𝟑 𝟎
−𝟕𝒙𝟏 +𝒙𝟐 −𝒙𝟑 = 𝟎
𝒙 =𝟎
⇔ {−𝟕𝒙𝟏 +𝒙𝟐 −𝒙𝟑 = 𝟎 ⇔{ 𝟏 .
𝒙𝟐 = 𝒙𝟑
−𝟔𝒙𝟏 +𝟔𝒙𝟐 −𝟔𝒙𝟑 = 𝟎
Hence the eigenvectors associated with the eigenvalue
𝟎
𝝀𝟏 = 𝟒 are of the form 𝜶 (𝟏).◄
𝟏
Example 5.
Find the characteristic polynomial, eigenvalues and
eigenvectors of the matrix A, where
5 4 2
𝐴 = [4 5 2].
2 2 2
Solution.
The characteristic polynomial is
𝜆 − 5 −4 −2
𝑃(𝜆) = 𝑑𝑒𝑡( 𝜆𝐼 − 𝐴) = | −4 𝜆 − 5 −2 |
−2 −2 𝜆 − 2
( )2( )
= 𝜆 − 1 𝜆 − 10 = 0.⇒ 𝜆 = 1, 1, and 10.
1. As𝜆 = 1,
−4 −4 −2 𝑥1
(1 ⋅ 𝐼 − 𝐴)𝑥 = [−4 −4 −2] [𝑥2 ] = 0.
−2 −2 −1 𝑥3
⇔ 𝑥1 = −𝑠 − 𝑡, 𝑥2 = 𝑠, 𝑥3 = 2𝑡

- 201 -
 𝑥1 −𝑠 − 𝑡 −1 −1
⇔ 𝑥 = [𝑥2 ] = [ 𝑠 ] = 𝑠 [ 1 ] + 𝑡 [ 0 ] , 𝑠, 𝑡 ∈ 𝑅.
𝑥3 2𝑡 0 2
Thus,
−1 −1
= 𝑠 [ 1 ] + 𝑡 [ 0 ] , 𝑠, 𝑡 ∈ 𝑅, 𝑠 ≠ 0 or 𝑡 ≠ 0,
0 2
are the eigenvectors associated with eigenvalue𝜆 = 1.
2. As𝜆 = 10,
5 −4 −2 𝑥1
(10 ⋅ 𝐼 − 𝐴)𝑥 = [−4 5 −2] [𝑥2 ] = 0.
−2 −2 8 𝑥3
⇔ 𝑥1 = 2𝑟, 𝑥2 = 2𝑟, 𝑥3 = 𝑟
𝑥1 2𝑟 2
⇔ 𝑥 = [𝑥2 ] = [2𝑟] = 𝑟 [2] , 𝑟 ∈ 𝑅. Thus,
𝑥3 𝑟 1
2
𝑟 [2] , 𝑟 ∈ 𝑅, 𝑟 ≠ 0,
1
are the eigenvectors associated with eigenvalue𝜆 = 10.◄
Example 5.
Find the eigenvectors of the matrix A, where
0 1 2
𝐴 = [2 3 0].
0 4 5
Solution.
𝜆 −1 −2
𝑓(𝜆) = 𝑑𝑒𝑡( 𝜆𝐼 − 𝐴) = |−2 𝜆 − 3 0 |
0 −4 𝜆−5
2
= (𝜆 − 1) (𝜆 − 6) = 0
⇒ 𝜆 = 1, 1, and 6.

- 202 -
 −1 1 2 𝑥1
3. As 𝜆 = 1,(𝐴 − 1 ⋅ 𝐼)𝑥 = [ 2 2 0] [𝑥2 ] = 0.
0 4 4 𝑥3
𝑥1 1 1
𝑥 = [𝑥2 ] = 𝑡 [−1] , 𝑡 ∈ 𝑅. Thus 𝑡 [−1] , 𝑡 ∈ 𝑅, 𝑡 ≠
𝑥3 1 1

0are the eigenvectors associated with eigenvalue 𝜆 = 1.

4. As𝜆 = 6,

−6 1 2 𝑥1
(𝐴 − 6 ⋅ 𝐼)𝑥 = [ 2 −3 0 ] [𝑥2 ] = 0.
0 4 −1 𝑥3
𝑥1 3 3
𝑥 = [𝑥2 ] = 𝑟 [2] , 𝑟 ∈ 𝑅 or 𝑟 [2] , 𝑟 ∈ 𝑅, 𝑟 ≠ 0.
𝑥3 8 8

are the eigenvectors associated with eigenvalue 𝜆 = 6.◄

- 203 -
 Exercise Set (5.4)
Find the eigenvalues of the matrices in problems 1 - 9.
2 7 3 −2 5 3
1. 𝐴 = ( ) 2. 𝐴 = ( ) 3. 𝐴 = ( )
7 2 1 −1 −4 4
0 0 0 4 0 0
4. 𝐴 = (0 2 5 ) 5. 𝐴 = (0 0 0 )
0 0 −1 1 0 −3
4 −7 0 2
4 0 0
0 3 −4 6
6. 𝐴 = (0 0 0 ) 7. 𝐴 = ( )
0 0 3 −8
1 0 −3
0 0 0 1
5 0 0 0
8 −4 0 0
8. 𝐴 = ( )
0 7 1 0
1 −5 2 1
3 0 0 0 0
−5 1 0 0 0
9. 𝐴 = 3 8 0 0 0 .
0 −7 2 1 0
(−4 1 9 −2 3)

- 204 -
 CHAPTER VI
PARTIAL FRACTIONS &
SEQUENCES

- 205 -
 CHAPTER VI

PARTIAL FRACTIONS & SEQUENCES

1. Partial Fractions Decomposition

In mathematics, we like to undo whatever has happened
to our expressions. For example, we have multiplication
to “undo” division and vice versa. Likewise, the square
root undoes squaring, and logarithms and exponentiation
can undo each other. So, when we add rational
expressions, we need the ability to work backward and
“undo” the sum. When we “undo” the sum of rational
expressions (i.e. when we take a single rational
expression and write is as the sum of other rational
expressions), we are performing partial fraction
decomposition.
Case I. Distinct Roots (Factors):
𝑓(𝑥)
Consider the proper rational function (𝑥−𝑎)(𝑥−𝑏)
where

the degree of𝑓(𝑥 ) is less than the degree of (𝑥 − 𝑎)(𝑥 −

𝑏), and a ≠ b. The partial fraction expansion will be
given by

- 206 -
 𝑓(𝑥) 𝐴 𝐵
(𝑥−𝑎)(𝑥−𝑏)
= + , (1)
𝑥−𝑎 𝑥−𝑏

where A and B must be determined.

If we multiply (1) by (𝑥 − 𝑎), we obtain
𝑓(𝑥) 𝐵 (𝑥−𝑎)
(𝑥−𝑏)
=𝐴+ .
𝑥−𝑏
𝑓(𝑥)
If we evaluate at𝑥 = 𝑎, we obtain that 𝐴 = (𝑥−𝑏)

evaluated at𝑥 = 𝑎.
𝑓(𝑥)
This is the same as covering-up (𝑥 − 𝑎) in (𝑥−𝑎)(𝑥−𝑏) and

evaluating at 𝑥 = 𝑎 in the part of the fraction that is

𝑓(𝑥)
visible; that is(𝑥−𝑏).

To find B, all we have to do is cover-up 𝑥 = 𝑏 in

𝑓(𝑥)
(𝑥−𝑎)(𝑥−𝑏)
and evaluate at 𝑥 = 𝑏.

This is what is known as the cover-up method.

NOTE:
If degree of the numerator is greater than or equal to
the degree of the denominator, long division must be
performed first and then partial fraction expansion can
be applied to the remainder by divisor part.

- 207 -
Example 1.
𝑥+1
Find the partial fraction decomposition of .
𝑥(𝑥−1)(𝑥+2)

Solution.
𝑥+1 𝐴 𝐵 𝐶
= + +
𝑥 (𝑥 − 1)(𝑥 + 2) 𝑥 𝑥−1 𝑥+2
𝑥+1
By the cover up method, 𝐴 = (𝑥−1)(𝑥+2) evaluated at𝑥 =

0. Notice that we have cover-up the x in the denominator,

0+1 1
so 𝐴 = (0−1)(0+2) = − .
2
𝑥+1
By the cover up method, 𝐵 = evaluated at𝑥 = 1.
𝑥(𝑥+2)

Notice that we have cover-up the (𝑥 − 1) in the

1+1 2
denominator, so𝐵 = = .
1(1+2) 3
𝑥+1
By the cover up method, 𝐶 = evaluated at𝑥 = −2.
𝑥(𝑥−1)

Notice that we have cover-up the (𝑥 + 2) in the

−2+1 1
denominator, so 𝐶 = =− .
−2(−2−1) 6

So, the partial fraction decomposition is

𝑥+1 −1/2 2/3 −1/6
= + +
𝑥 (𝑥 − 1)(𝑥 + 2) 𝑥 𝑥−1 𝑥+2
Note. If you solve this problem using the book’s method,
you will need to solve a system of 3 equations.
- 208 -
Exercise 1.
𝑥 2 +2𝑥+5
Find the partial fraction decomposition of .
𝑥 3 +3𝑥 2 −𝑥−3

Case II Repeated Roots (Factors):

𝑓(𝑥)
Consider the proper rational function (𝑥−𝑎)2 (𝑥−𝑏)
,

where (𝑥 − 𝑎)is a repeated root [(𝑥 − 𝑎)2 =

(𝑥 − 𝑎)(𝑥 − 𝑎)]. The partial fraction expansion will be
given by

𝑓(𝑥) 𝐴 𝐵 𝐶
(𝑥−𝑎)2 (𝑥−𝑏)
= 2 + + ,
(𝑥−𝑎) 𝑥−𝑎 𝑥−𝑏

where A, B and C must be determined.

When you have repeated roots, you have to account for
the original factor, and all the lower powers of that factor.
In this case, A and C can be determined using the cover-
up methods (note that those are the factors in the original
function).
Example 2.
𝑥+1
Find the partial fraction decomposition of .
𝑥(𝑥−1)3

- 209 -
Solution.
𝑥+1 𝐴 𝐵 𝐶 𝐷
= + (𝑥−1)3 + (𝑥−1)2 + (𝑥−1) (2)
𝑥(𝑥−1)3 𝑥

A and B can be found using the cover up method (same as

the original factors)
𝑥+1
By the cover up method, 𝐴 = (𝑥−1)3 evaluated at𝑥 = 0,

so 𝐴 = −1
𝑥+1
By the cover up method, 𝐵 = evaluated at𝑥 = 1,
𝑥

so𝐵 = 2.
To find C and D, we have to solve a system of equations.
Since we know A and B, if we multiply (2) times𝑥 (𝑥 −
1)3 , we obtain
𝑥 + 1 = −(𝑥 − 1)3 + 2𝑥 + 𝐶𝑥(𝑥 − 1) + 𝐷𝑥(𝑥 − 1)2 (3)

If we substitute any number other than the ones already

used (say𝑥 = 2) in (3),we obtain
𝐶+𝐷 = 0
If we substitute any other number (say𝑥 = −1) in (3), we
obtain 𝐶 − 2𝐷 = −3
If we solve the system, we get 𝐷 = 1 and 𝐶 = −1
So, the partial fraction decomposition is
𝑥+1 −1 2 −1 1
= + (𝑥−1)3 + (𝑥−1)2 + (𝑥−1).◄
𝑥(𝑥−1)3 𝑥

- 210 -
Note: If you solve this problem using the book’s method,
you will need to solve a system of 4 equations.
Case III Quadratic Roots:
𝑓(𝑥)
Consider the proper rational function (𝑥−𝑎)(𝑥 2 +𝑏𝑥+𝑐)
,

where (𝑥 2 + 𝑏𝑥 + 𝑐) is any quadratic factor with real or

imaginary solutions. The partial fraction expansion will
be given by
𝑓(𝑥) 𝐴 𝐵𝑥+𝐶
(𝑥−𝑎)(𝑥 2 +𝑏𝑥+𝑐)
= + , where A, B and C
(𝑥−𝑎) (𝑥 2 +𝑏𝑥+𝑐)

must be determined.
In this case we can use the cover-up method in the linear
term, and a system of equations find𝐵𝑥 + 𝐶.
Example 3.
𝑥+1
Find the partial fraction decomposition of .
𝑥(𝑥 2 −2𝑥−2)

Solution.
𝑥+1 𝐴 𝐵𝑥+𝐶
= + (4)
𝑥(𝑥 2 −2𝑥−2) 𝑥 𝑥 2 −2𝑥−2
𝑥+1
By the cover up method, 𝐴 = (𝑥 2 evaluated at 𝑥 =
−2𝑥−2)
1
0, so 𝐴 = − .
2

- 211 -
 1
Since we know 𝐴 = − by cover-up, if we multiply (4)
2

times 𝑥 (𝑥 2 − 2𝑥 − 2), we obtain

1
𝑥 + 1 = − (𝑥 2 − 2𝑥 − 2) + 𝑥 (𝐵𝑥 + 𝐶 )(5)
2

At this point, you can substitute any number for x.

If we substitute 𝑥 = 2 in (5) we obtain
2𝐵 + 𝐶 = 1
If we substitute 𝑥 = −1 in (5) we obtain
𝐵 − 𝐶 = 1/2
If we solve the system, we get 𝐵 = 1/2 and 𝐶 = 0.
So, the partial fraction decomposition is
𝑥+1 −1/2 (1/2)𝑥
= + .
𝑥(𝑥 2 −2𝑥−2) 𝑥 𝑥 2 −2𝑥−2

Note: Skip this part. In some cases, this method can be

more complicated than the one shown above. The same
example can be solved by using cover-up in both terms.
𝑥+1 𝐴 𝐵𝑥+𝐶
= 𝑥 + 𝑥 2 −2𝑥−2 (6)
𝑥(𝑥 2 −2𝑥−2)
𝑥+1
By the cover up method, 𝐴 = (𝑥 2 evaluated at𝑥 =
−2𝑥−2)
1
0, so 𝐴 = − .
2

- 212 -
 𝑥+1
By the cover up method, 𝐵𝑥 + 𝐶 = evaluated at𝑥 =
𝑥

1 + √3, one of solutions of (𝑥 2 − 2𝑥 − 2) = 0. So

2+√3
𝐵(1 + √3) + 𝐶 = .
1+√3

If we rearrange and multiply the right-hand side by the

conjugate, we obtain
(2+√3)(1−√3) (−1−√3) 1 √3
(𝐵 + 𝐶 ) + 𝐵√3 = = = + .
(1+√3)(1−√3) −2 2 2

If we equate the 2 sides of the equation

1 √3 1
(𝐵 + 𝐶 ) + 𝐵√3 = + , we obtain(𝐵 + 𝐶 ) = ,
2 2 2
√3
and𝐵√3 = .
2
1
From these two equations, we get 𝐵 = and𝐶 = 0.
2

The partial fraction decomposition of

𝑥+1 −1/2 (1/2)𝑥
= + .
𝑥 (𝑥 2 − 2𝑥 − 2) 𝑥 𝑥 2 − 2𝑥 − 2
Cover-up for this case can become complicated, so in
many cases is better to use the systems of equations
method.

- 213 -
 Exercise Set (6.3)
1) Find the partial fraction decomposition of
2𝑥
4𝑥 2 + 12𝑥 + 9
.
2) Find the partial fraction decomposition of
−10𝑥 2 + 27𝑥 − 14
(𝑥 − 1)3 (𝑥 + 2)
.
3) Find the partial fraction decomposition of
2𝑥 4 +4𝑥 3 −2𝑥 2 +𝑥+7
.
𝑥 3 +2𝑥 2 −𝑥−2

4) Expand into Partial Fractions

Function Answer
1
x 1 1 6 4 1
f( x ) := 2 f( x ) := + −
(x −1) (x +4) 10 x − 1 x + 1 15 x + 4

x3 + 1
g( x ) :=
(x −1) (x −2) g(x) :=
1 11 1 4
h( x ) := h( x ) := − + +
2 x3 + 3 x2 − 2 x 2 x 10 ( x + 2 ) 5 ( 2 x − 1 )

1 1 x
i( x ) := i( x ) := − 2
(x +1) x
2 x x +1

- 214 -
2. SEQUENCES
The "sequence" in mathematics is used to describe
an unending succession of numbers. Some possibilities
are
1 ,2 ,3 ,4 ,…
2 ,4 ,6 ,8 ,…
1 1 1
1, , , , …
2 3 4

1 , −1, 1, −1, …
In each case, the dots are used to suggest that the
sequence continues indefinitely, following the obvious
pattern. The numbers in a sequence are called the terms
of the sequence. The terms may be described according to
the positions they occupy. Thus, a sequence has a first
term, a second term, a third term, and so forth. Because
a sequence continues indefinitely, there is no last term.
The most common way to specify a sequence is to
give a formula for the terms. To illustrate the idea, we
have listed in the table below the terms in the sequence
2, 4, 6, 8, … together with their term numbers:
Term number:1 2 3 4 …
Term 2 4 6 8 …
- 215 -
There is a clear relationship here: each term is twice its
term number. Thus, for each positive integer 𝑛 the nth
term in the sequence 2, 4, 6, 8, … is given by the formula
2 𝑛.This is denoted by writing
2 , 4 , 6 , 8 , … ,2 𝑛 , …
or more compactly in bracket notation as
{2𝑛}+∞
𝑛=1

From the bracket notation, the terms in the sequence can

be generated by successively substituting the integer
values 𝑛 = 1, 2, 3, … into the formula 2 𝑛.
Example 1.
List the first five terms of the sequence{2𝑛 }+∞
𝑛=1

Solution.
Substituting 𝑛 = 1, 2, 3, 4, 5 into the formula2𝑛 yields
21 , 22 , 23 , 24 , 25
or, equivalently 2, 4, 8, 16, 32.◄
Example 2.
Express the following sequences in bracket notation
1 2 3 4
(a) , , , ,…
2 3 4 5
1 1 1 1
(b) , , , ,…
2 4 8 16

- 216 -
 (c) 1, −1, 1, −1, …
1 2 3 4
(d) ,− , ,− ,…
2 3 4 5

(e) 1, 3, 5, 7, …
Solution.
(a) Begin by comparing terms and term numbers:
Term number 1 2 3 4 …
1 2 3 4
Term …
2 3 4 5

In each term, the numerator is the same as the term

number, and the denominator is one greater than the term
𝑛
number. Thus, the nth term is and the sequence may
𝑛+1

be written
𝑛 +∞
{ }
𝑛 + 1 𝑛=1
(b)Observe that the sequence can be rewritten as
1 1 1 1
, , , ,…
2 22 23 24
and construct a table comparing terms and term numbers:
Term number 1 2 3 4…
1 1 1 1
Term …
2 22 23 24

- 217 -
 1
From the table we see that the nth term is so the
2𝑛

sequence can be written as

1 +∞
{ 𝑛}
2 𝑛=1

(c) Observe first that(−1)𝑟 is either 1 or -1 according to

whether r is an even integer or an odd integer. In the
sequence 1, −1, 1, −1, … the odd-numbered terms are
1's and the even-numbered terms are -1's. Thus a
formula for the nth term can be obtained by raising -1
to a power that will be even when n is odd and odd
when n is even. This is accomplished by the formula
(−1)𝑛+1 , so that the sequence can be written as
{(−1)𝑛+1 }+∞
𝑛=1
(d) Combining the results in parts (a) and (c), we may
write this sequence as
𝑛 +∞
{(−1)𝑛+1 }
𝑛 + 1 𝑛=1
(e) Begin by comparing terms and term numbers:
Term number 1 2 3 4 …
Term 1 3 5 7…
From the table we see that each term is one less than
twice the term number. Thus, the nth term is 2𝑛 − 1 and
the sequence can be written as

- 218 -
 {2 𝑛 − 1}+∞
𝑛=1

Frequently we shall want to write down a sequence

without specifying the numerical values of the terms. We
do this by writing
𝑎1 , 𝑎2 , … , 𝑎𝑛 , …
Or in bracket notation
{𝑎𝑛 }+∞
𝑛=1

Or sometimes simply {𝑎𝑛 }

The time has come to formally define the term
"sequence". When we write a sequence such as
2, 4, 6, 8, … , 2𝑛 , …
In bracket notation
(1) {2 𝑛}+∞
𝑛=1

We could rewrite (1) in functional notation as

𝐹(𝑛) = 2𝑛 , 𝑛 = 1, 2, 3, …
From this point of view, the notation
2, 4, 6, 8, … , 2 𝑛, …
represents a listing of the function values
𝑓(1), 𝑓(2), 𝑓(3), … , 𝑓(𝑛), …
This suggests the following definition.

- 219 -
Definition 1.
A sequence (or infinite sequence) is a function whose
domain is the set of positive integers.
Because sequences are functions, we may inquire about
the graph of sequence. For example, the graph of the
𝑛+1 +∞
sequence{ }
2𝑛2 𝑛=1

is the graph of the sequence

𝑛+1
𝑦= ,𝑛 = 1, 2, 3, …
2𝑛2

Because the right side of this equation is defined only for

positive integers values of n, the graph consists of a
succession of isolated points (Figure 1a). This is in
𝑥+1
marked distinction to the graph of 𝑦 = ,𝑥 ≥ 1
2𝑥 2

which is a continuous curve (Figure 1 b).

(Figure 1 a)

- 220 -
y

𝑥+1
𝑦= ,𝑥≥1
2𝑥 2

x
(Figure 1 b).
In Figure 2 we have sketched the graphs of 3sequences:
𝑛 − 1 +∞ 𝑛 − 1 +∞
{ } , {(−1)𝑛+1
( )} , {3}+∞
𝑛=1
𝑛 𝑛=1 𝑛 𝑛=1

Figure 2 (a)

Figure 2 (b)

- 221 -
 Figure 2 (c)

To say that a sequence{𝑎𝑛 }+∞

𝑛=1 approaches a limit L as n

gets large is intended to mean that eventually the terms in

the sequence become arbitrary close to the number L.
Thus, if we choose any positive number 𝜀, the terms in
the sequence willeventually be within 𝜀 units of L.
Geometrically, this means that if we sketch the lines 𝑦 =
𝐿 + 𝜀 and 𝑦 = 𝐿 − 𝜀 , the terms in the sequence will
eventually be trapped within the band between these
lines, and thus be within 𝜀 units of L (Figure 3).

Figure 3

- 222 -
The following definition expresses this idea precisely.
Definition 2.
A sequence{𝑎𝑛 }+∞
𝑛=1 is said to have the limit L if given

any𝜀 > 0 there is a positive integer N such that|𝑎𝑛 −

𝐿| < 𝜀when 𝑛 ≥ 𝑁.If a sequence{𝑎𝑛 }+∞
𝑛=1 has a limit L,

we say that the sequence converges to L and write

lim 𝑎𝑛 = 𝐿
𝑛→+∞

A sequence that does not have a finite limit is said to

diverge.
Example 3.
+∞
𝑛−1
Figure 2 suggests that the sequence {(−1)𝑛+1 { 𝑛 }}
𝑛=1

𝑛−1 +∞
diverge, while the sequences { } and{3}+∞
𝑛=1
𝑛 𝑛=1

converge to 1 and 3, respectively; that is,

𝑛−1
lim = 1 𝑎𝑛𝑑 lim 3 = 1.◄
𝑛→+∞ 𝑛 𝑛→+∞

Theorem 3.
Suppose that the sequences {𝑎𝑛 }and {𝑏𝑛 }converge to
limits 𝐿1 and 𝐿2 , respectively, and c is a constant. Then
(a) lim 𝑐 = 𝑐
𝑛→+∞

- 223 -
(b) lim 𝑐𝑎𝑛 = 𝑐 lim 𝑎𝑛 = 𝑐𝐿1
𝑛→+∞ 𝑛→+∞

(c) lim (𝑎𝑛 + 𝑏𝑛 ) = lim 𝑎𝑛 + lim 𝑏𝑛 = 𝐿1 + 𝐿2

𝑛→+∞ 𝑛→+∞ 𝑛→+∞

(d) lim (𝑎𝑛 − 𝑏𝑛 ) = lim 𝑎𝑛 − lim 𝑏𝑛 = 𝐿1 − 𝐿2

𝑛→+∞ 𝑛→+∞ 𝑛→+∞

(e) lim (𝑎𝑛 . 𝑏𝑛 ) = lim 𝑎𝑛 . lim 𝑏𝑛 = 𝐿1 𝐿2

𝑛→+∞ 𝑛→+∞ 𝑛→+∞

𝑎 lim 𝑎𝑛 𝐿
(f) lim ( 𝑛 ) = 𝑛→+∞ = 1 (𝑖𝑓 𝐿2 ≠ 0).
𝑏
𝑛→+∞ lim 𝑏
𝑛 𝐿
𝑛→+∞ 𝑛 2

Example 4.
In each part, determine if the given sequence converges
or diverges. If it converges, find the limit.
𝑛 +∞ 𝑛 +∞
(a) { } (b){(−1)𝑛+1 }
2𝑛+1 𝑛=1 2 𝑛+1 𝑛=1

1 +∞
(c) {(−1)𝑛+1 } (d) {8 − 2 𝑛}+∞
𝑛=1
𝑛 𝑛=1

𝑛 +∞
(e) { 𝑛} .
𝑒 𝑛=1

Solution.
(a) Dividing numerator and denominator by n yields
𝑛 1 1 1
lim = lim 1 = = .
𝑛→+∞ 2 𝑛+1 𝑛→+∞ (2+ ) 2+0 2
𝑛

𝑛 +∞ 1
Thus{ } converges to .
2 𝑛+1 𝑛=1 2
𝑛 1
(b) From part (a), lim =
𝑛→+∞ 2 𝑛+1 2

- 224 -
Thus, since(−1)𝑛+1 between + 1 and -1, the product
𝑛
(−1)𝑛+1 oscillates between positive and negative
2 𝑛+1
1
values, with the odd-numbered terms approaching and
2
1
the even-numbered terms approaching.− . Therefore,
2

𝑛 +∞
the sequence {(−1)𝑛+1 } approaches no limit. It
2 𝑛+1 𝑛=1

diverges.
1 1
(c) Since lim = 0, the product(−1)𝑛+1 ( )
𝑛→+∞ 𝑛 𝑛

oscillates between positive and negative values, with

the odd-numbered terms approaching 0 through
positive values and the even-numbered terms
approaching 0 through negative values. Thus
1
lim (−1)𝑛+1 = 0.
𝑛→+∞ 𝑛

So the sequence converges to 0.

(d) lim (8 − 2 𝑛) = −∞. So,{8 − 2 𝑛}+∞
𝑛=1 diverges.
𝑛→+∞
𝑛
(e) We want to find lim which is an indeterminate
𝑛→+∞ 𝑒 𝑛

from type ∞/∞. Unfortunately, we cannot apply

L'Hopital's rule directly since𝑒 𝑛 and n are not
differentiable (n assumes only integer values).

- 225 -
 However, we can apply L'Hopital's rule to the related
𝑥
problem lim to obtain
𝑥→+∞ 𝑒 𝑥
𝑥 1
lim = lim =0
𝑥→+∞ 𝑒 𝑥 𝑥→+∞ 𝑒 𝑥
𝑛
We conclude from this that lim = 0since the values
𝑛→+∞ 𝑒 𝑛
𝑛 𝑥
of and are the same when x is a positive integer.◄
𝑒𝑛 𝑒𝑥

Example 5.
𝑛
Show that lim √𝑛 = 1.
𝑛→+∞

Solution.
1 𝑛
By L'Hopital's rule, lim ln 𝑛 = 0.Thus lim √𝑛 =
𝑛→+∞ 𝑛 𝑛→+∞
1
lim 𝑛 𝑛
𝑛→+∞

= lim 𝑒 (1/𝑛) ln 𝑛 = 𝑒 0 = 1.◄

𝑛→+∞

- 226 -
 Exercise Set (6.1)
Show the first five terms of the sequence, determine if the
sequence converges, and if so find the limit. (When
writing out the terms of the sequence, you need not find
numerical values: leave the forms in the first form you
obtain).
+∞ +∞
𝑛 𝑛2
(1) { } (2) { }
𝑛+2 𝑛=1 2 𝑛+1 𝑛=1

1 +∞
(3) {2}+∞
𝑛=1 (4) {ln ( )}
𝑛 𝑛=1

ln 𝑛 +∞ 𝜋 +∞
(5) { } (6) {𝑛 sin }
𝑛 𝑛=1 𝑛 𝑛=1
+∞
(−1)𝑛+1
(7) {1 + (−1)𝑛 }+∞
𝑛=1 (8) { }
𝑛2 𝑛=1
+∞
2 𝑛3 𝑛 +∞
(9) {(−1)𝑛 } (10) { 𝑛 }
𝑛3 +1 𝑛=1 2 𝑛=1

(𝑛+1)(𝑛+2) +∞ 𝜋𝑛 +∞
(11) { } (12){ 𝑛 }
2 𝑛2 𝑛=1 4 𝑛=1
+∞ +∞
1
2 𝑛
(13) {𝑛 } 𝑛 (14) {(1 − ) }
𝑛=1 𝑛 𝑛=1

In Exercise (15)-(22), express the sequence in the

notation{𝑎𝑛 }+∞
𝑛=1 , determine if the sequence converges,

and if so find its limit.

- 227 -
 1 3 5 7 1 2 3
(15) , , , , … (16) 0 , , , ,…
2 4 6 8 22 32 42
1 1 1 1
(17) , , , ,… (18) −1 , 2 , −3 ,4 , −5 , …
3 9 27 81
1 1 1 1 1 1 1
(19)(1 − ) , ( − ) , ( − ) , ( − ),…
2 2 3 3 4 4 5
3 3 3
(20) 3 , , , ,…
2 22 23

(21) (√2 − √3) , (√3 − √4) , (√4 − √5) , …

1 1 1 1
(22) ,− , ,− ,…
35 36 37 38

(23) (a) Let{𝑎𝑛 }be a sequence for which 𝑎1 = 3

and 𝑎𝑛 = 2 𝑎𝑛−1 when 𝑛 ≥ 2. Find the first eight
terms.
(b) Let{𝑎𝑛 } be a sequence for which 𝑎1 = 1,𝑎2 =
1 and 𝑎𝑛 = 𝑎𝑛−1 + 𝑎𝑛−2 when 𝑛 ≥ 3. Find the
first eight terms.
(24) The nth term 𝑎𝑛 of the sequence 1, 2, 1, 4, 1, 6, … is
best written in the form
1 , if 𝑛 𝑖𝑠 𝑜𝑑𝑑
𝑎𝑛 = {
𝑛 , if 𝑛 𝑖𝑠 𝑒𝑣𝑒𝑛
since it would be tedious to find one formula applicable
to all terms. By considering even and odd term

- 228 -
separately, find a formula for the nth term of the given
sequence
1 1 1
(a) 1 , ,3 , ,5 , ,…
23 24 26
1 1 1 1 1 1 1 1
(b)1 , , , , , , , , , …
3 3 5 5 7 7 9 9

(25) Consider the sequence {𝑎𝑛 }+∞

𝑛=1 where
1 2 𝑛
𝑎𝑛 = + + ⋯ +
𝑛2 𝑛2 𝑛2
(a) Write out the first four terms of the
sequence.
(b)Find the limit of the sequence.
[ Hint: Sum up the terms in the formula for𝑎𝑛 ]
(26) Follow the directions in Exercise (25) with
12 22 𝑛2
𝑎𝑛 = 3 + 3 + ⋯ + 3
𝑛 𝑛 𝑛
(27) I
1 +∞
f we accept the fact that the sequence{ } converges to
𝑛 𝑛=1

the limit 𝐿 = 0, then according to Definition 2, for every

𝜀 > 0, there exists an integer N such that |𝑎𝑛 − 𝐿| =
1
|( ) − 0| < 𝜀when 𝑛 ≥ 𝑁. In each part, find the smallest
𝑛

possible value of N for the given value of 𝜀.

- 229 -
 (a) 𝜀 = 0.5 (b) 𝜀 = 0.1 (c) 𝜀 = 0.001
(28) I
𝑛 +∞
f we accept the fact that the sequence{ } converge
𝑛+1 𝑛=1

to the limit 𝐿 = 1, then according to Definition 2 for

every 𝜀 > 0, there exists an integer N such that|𝑎𝑛 −
𝑛
𝐿| = | − 1| < 𝜀 when 𝑛 ≥ 𝑁. In each part, find the
𝑛+1

smallest value of N for the given value of 𝜀.

(a) 𝜀 = 0.25 (b) 𝜀 = 0.1 (c) 𝜀 = 0.001
(29) Prove
1 +∞
(a) The sequence { } converge to 0.
𝑛 𝑛=1

𝑛 +∞
(b) The sequence{ } converge to 1.
𝑛+1 𝑛=1

(30) Consider the sequence {𝑎𝑛 }+∞

𝑛=1 whose nth term is
𝑛−1
1 1
𝑎𝑛 = ∑ 𝑘 .
𝑛
𝑘=0 1 +
𝑛

Show that lim 𝑎𝑛 = ln 2

𝑛→+∞

[Hint: Interpret lim 𝑎𝑛 as a definite integral.]

𝑛→+∞

- 230 -
3. MONOTONE SEQUENCES
Definition 1.
A sequence{𝑎𝑛 } is called
increasing if 𝑎1 < 𝑎2 < 𝑎3 < ⋯ < 𝑎𝑛 < ⋯
nondecreasing if 𝑎1 ≤ 𝑎2 ≤ 𝑎3 ≤ ⋯ ≤ 𝑎𝑛 ≤ ⋯
decreasing if 𝑎1 > 𝑎2 > 𝑎3 > ⋯ > 𝑎𝑛 >
⋯nonincreasing if 𝑎1 ≥ 𝑎2 ≥ 𝑎3 ≥ ⋯ ≥ 𝑎𝑛 ≥ ⋯
A sequence that is either nondecreasing or nonincreasing
is called monotone, and a sequence that is increasing or
decreasing is called strictly monotone. Observe that a
strictly monotone sequence is monotone, but not
conversely.
Example 1.
1 2 3 𝑛
, , ,…, ,… is increasing
2 3 4 𝑛+1
1 1 1
1 , , ,…, ,… is decreasing
2 3 𝑛

1, 1, 2, 2, 3, 3, … is nondecreasing
1 1 1 1
1 ,1 , , , , , … is nonincreasing
2 2 3 3

All four of these sequences are monotone, but the

sequence

- 231 -
 1 1 1 1
1,− , , − , … , (−1)𝑛+1 , …
2 3 4 𝑛
is not. The first and second sequences are strictly
monotone.
Frequently, one can guess whether a sequence is
increasing, decreasing, nondecreasing, or nonincreasing
after writing out some of the initial terms. However, to be
certain that the guess is correct, the kind of analysis
illustrated in the next few examples is needed.
Example 2.
Show that
1 2 3 𝑛
, , ,…, ,…
2 3 4 𝑛+1
is an increasing sequence.
Solution.
Although it is intuitively clear that the sequence is
increasing, intuition cannot replace a definitive proof. To
prove that the sequence is increasing, let
𝑥
𝑓(𝑥 ) =
𝑥+1

- 232 -
So, the nth term in the given sequence is 𝑎𝑛 = 𝑓(𝑛). The
function f is increasing for 𝑥 ≥ 1 since
(𝑥+1)(1)−𝑥(1) 1
𝑓′(𝑥 ) = (𝑥+1)2
= (𝑥+1)2. So 𝑓′(𝑥) > 0 for 𝑥 ≥ 1.

Thus 𝑎𝑛 = 𝑓 (𝑛) < 𝑓(𝑛 + 1) = 𝑎𝑛+1 . This proves that

the given sequence is increasing.
Alternate Solution.
𝑛
Since 𝑎𝑛 = , we have by replacing n by 𝑛 + 1
𝑛+1
𝑛+1 𝑛+1
𝑎𝑛+1 = (𝑛+1)+1 = .
𝑛+2

Thus, for 𝑛 ≥ 1
𝑛 𝑛 + 1 𝑛2 + 2𝑛 − 𝑛2 − 2 𝑛
𝑎𝑛 − 𝑎𝑛+1 = − =
𝑛+1 𝑛+2 (𝑛 + 1)(𝑛 + 2)
1
= − (𝑛+1)(𝑛+2) < 0.

So that𝑎𝑛 < 𝑎𝑛+1 for 𝑛 ≥ 1. This proves that the

sequence is increasing.
●To prove that a sequence {𝑎𝑛 }is decreasing, one can
show either that 𝑓′(𝑥) < 0 for 𝑥 ≥ 1 [where 𝑓(𝑛) =
𝑎𝑛 ] or 𝑎𝑛 − 𝑎𝑛+1 > 0 for 𝑛 ≥ 1.
●If a sequence {𝑎𝑛 }has positive terms, then it is
increasing if

- 233 -
 𝑎𝑛+1
>1 (1a)
𝑎𝑛

for 𝑛 ≥ 1, and it is decreasing if

𝑎𝑛+1
<1 (1b)
𝑎𝑛

for 𝑛 ≥ 1.
Example 3.
Use (1a) to prove that the sequence in Example 2 above is
increasing.
Solution.
As shown in the alternate solution of Example 2,
𝑛 𝑛+1
𝑎𝑛 = 𝑎𝑛𝑑 𝑎𝑛+1 =
𝑛+1 𝑛+2
Thus
(𝑛+1)
𝑎𝑛+1 (𝑛+2) 𝑛2 + 2 𝑛 + 1
= 𝑛 = (2)
𝑎𝑛 𝑛2 + 2 𝑛
(𝑛+1)
𝑎𝑛+1
That is > 1 for 𝑛 ≥ 1.This proves that the sequence
𝑎𝑛

is increasing.◄
●In our subsequent work n! (n factorial)is the product of

the first n positive integers, that is 𝑛 ! = 1 . 2 . 3 … 𝑛 and

0 ! = 1.

- 234 -
Example 4.
Show that the sequence
𝑒 𝑒2 𝑒3 𝑒𝑛
, , ,…, ,…
2! 3! 4! (𝑛 + 1)!

is decreasing

Solution.

𝑒𝑛
We will use (1b). Since 𝑎𝑛 = (𝑛+1)!, it follows on

replacing n by 𝑛 + 1 that

𝑒 𝑛+1 𝑒 𝑛+1
𝑎𝑛+1 = =
[(𝑛 + 1) + 1]! (𝑛 + 2)!

Thus

𝑒 𝑛+1
𝑎𝑛+1 (𝑛+2)! 𝑒 𝑛+1 (𝑛 + 1)! 𝑒
= 𝑒𝑛
= . =
𝑎𝑛 𝑎𝑛 (𝑛 + 2)! 𝑛 + 2
(𝑛+1)!

𝑎𝑛+1
For 𝑛 ≥ 1, 𝑛 + 2 ≥ 3 > 𝑒(≈ 2 .718 … ). So, =
𝑎𝑛
𝑒
< 1 for 𝑛 ≥ 1. This proves that the sequence is
𝑛+2

decreasing. ◄

- 235 -
 Theorem 2.
For a nondecreasing sequence 𝑎1 ≤ 𝑎2 ≤ ⋯ ≤ 𝑎𝑛 ≤ ⋯
there are two possibilities:
(a) There is a constant M such that𝑎𝑛 ≤ 𝑀 for all n, in
which case the sequence converges to a limit L
satisfying 𝐿 ≤ 𝑀.
(b) No such constant exists, in which case
lim 𝑎𝑛 = +∞
𝑛→+∞

Theorem 3.
For a nonincreasing sequence 𝑎1 ≥ 𝑎2 ≥ ⋯ ≥ 𝑎𝑛 ≥ ⋯
there are two possibilities:
(a) There is a constant M such that 𝑎𝑛 ≥ 𝑀 for all n, in
which case the sequence converges to a limit L
satisfying 𝐿 ≤ 𝑀.
(b) No such constant exists, in which case
lim 𝑎𝑛 = −∞.
𝑛→+∞

Axiom 1.

If S is a nonempty set of real numbers, and if there

is some real number that is greater than or equal to every

- 236 -
number in S, then there is a smallest real number that is
greater than or equal to every number in S.

For example, let S be the set of numbers in the

interval (1, 3). It is true that there exists a number u
greater than or equal to every number in S; some
examples are 𝑢 = 10, 𝑢 = 100, and 𝑢 = 3.2. The
smallest number u that is greater than or equal to every
number in S is 𝑢 = 3.

●There is an alternate phrasing of the completeness

axiom that is useful to know. Let us call a number u an
upper bound for a set S if u is greater than or equal to
every number in S.

Axioms 2.

If a nonempty set S of real numbers has an upper

bound, then S has a smallest upper bound.

The smallest upper bound of S is commonly called

the least upper bound of S.

- 237 -
Example 5.
𝑛 +∞
As shown in Examples 2 and 3, the sequence{ } is
𝑛+1 𝑛=1
𝑛
nondecreasing. Since𝑎𝑛 = < 1 , 𝑛 = 1 , 2 , …, the
𝑛+1

terms in the sequence have M = 1 as an upper bound. By

Theorem 2 above the sequence must converge to a
limit 𝐿 ≤ 𝑀 This is indeed the case since
𝑛 1
lim = lim 1 = 1.
𝑛→+∞ 𝑛+1 𝑛→+∞ 1+𝑛

Because the limit of a sequence {𝑎𝑛 }describes the

behavior of the terms as n gets large, one can add or even
delete a finite number of terms in a sequence without
affecting either the convergence or the value of the limit.
That is, the original and the modified sequence will both
converge or both diverge, and in the case of convergence
both will have the same limit. We omit the proof.
Example 6.
5𝑛 +∞
Show that the sequence { }
𝑛! 𝑛=1

converges.

- 238 -
Solution.
It is tedious to determine convergence directly from the
limit
5𝑛
lim
𝑛→+∞ 𝑛!

Thus, we will proceed indirectly. If we let

5𝑛
𝑎𝑛 =
𝑛!
Then
5𝑛+1
𝑎𝑛+1 =
(𝑛 + 1)!
so that
5𝑛+1
𝑎𝑛+1 (𝑛+1)! 5𝑛+1 𝑛! 5
= 5𝑛
= 𝑛 . =
𝑎𝑛 5 (𝑛 + 1)! 𝑛 + 1
𝑛!
𝑎𝑛+1
For 𝑛 = 1, 2, and 3, the value of is greater than 1
𝑎𝑛

so𝑎𝑛+1 > 𝑎𝑛 . Thus 𝑎1 < 𝑎2 < 𝑎3 < 𝑎4 .

𝑎𝑛+1
For 𝑛 = 4 the value of is 1. So𝑎4 = 𝑎5 .
𝑎𝑛
𝑎𝑛+1
For 𝑛 ≥ 5 the value of is less than 1, so
𝑎𝑛

𝑎5 > 𝑎6 > 𝑎7 > 𝑎8 > ⋯

- 239 -
Thus, if we discard the first four terms of the given
sequence (which won't affect convergence) the resulting
sequence will be decreasing. Moreover, each term in the
sequence is positive. So that by Theorem 3, the sequence
converges to some limit that is ≥ 0.
Exercise Set (6.2)
In Exercise 1-6, use differentiation to show that the
sequence is strictly monotone and classify it as increasing
or decreasing
𝑛 +∞ 1 +∞
(1){ } (2) {3 − }
2 𝑛+1 𝑛=1 𝑛 𝑛=1

1 +∞
(3) { } (4) {𝑛 𝑒 −2 𝑛 }+∞
𝑛=1
𝑛+ln 𝑛 𝑛=1

ln(𝑛+2) +∞
(5) { } (6) {tan−1 𝑛}+∞
𝑛=1
𝑛+2 𝑛=1

In exercise (7) - (12), determine if the given sequence

𝑎𝑛+1
{𝑎𝑛 }is monotone by examining . If so, classify it as
𝑎𝑛

increasing, decreasing, nonincreasing, or nondecreasing.

𝑛 +∞ 2𝑛 +∞
(7) { } (8) { }
2 𝑛+1 𝑛=1 𝑛! 𝑛=1

𝑛 +∞ 2𝑛 +∞
(9) { 𝑛} (10) { }
2 𝑛=1 1+2𝑛 𝑛=1

- 240 -
 10 +∞ 𝑛𝑛 +∞
(11) { 𝑛2
} (12) { }
2 𝑛=1 𝑛! 𝑛=1

In Exercise 13 - 18, shows the sequence is monotone and

apply Theorem 2 or 3 above to determine if the sequence
converges.
1 +∞ 2 +∞
(13) {2 + } (14) {4 − }
𝑛 𝑛=1 𝑛 𝑛=1

1 +∞ 4 𝑛−1 +∞
(15) {𝑛 − } (16) { }
𝑛 𝑛=1 5 𝑛+2 𝑛=1

𝑛 +∞ 𝜋 +∞
(17) { 𝑛 } (18) {cos }
5 𝑛=1 2 𝑛 𝑛=1

(19) Find the limit (if it exists) of the sequence

1 1 1 1
(a) 1, −1 , 1 , , , , , …
2 3 4 5
1
(b) − , 0 ,0 ,0 , 1 , 2 , 3 , 4 , …
2

- 241 -
 CHAPTER (VII)
INFINITE SERIES

- 242 -
 CHAPTER (VII)
INFINITE SERIES
1. Infinite Series
Definition 1.
An infinite series is an expression of the form
𝑢1 + 𝑢2 + 𝑢3 + ⋯ + 𝑢𝑘 + ⋯
or in sigma notation
∞

∑ 𝑢𝑘
𝑘=1

The numbers 𝑢1 , 𝑢2 , 𝑢3 , …are called the terms of the

series.
Informally speaking the expression∑∞
𝑘=1 𝑢𝑘 directs us to

obtain the "sum" of the terms 𝑢1 , 𝑢2 , 𝑢3 , …. To carry out

this summation process we proceed as follows:
Let 𝑠𝑛 denote the sum of the first n terms of the series.
Thus
𝑠1 = 𝑢1
𝑠2 = 𝑢1 + 𝑢2
𝑠3 = 𝑢1 + 𝑢2 + 𝑢3
⋮

- 243 -
 𝑛

𝑠𝑛 = 𝑢1 + 𝑢2 + 𝑢3 + ⋯ + 𝑢𝑛 = ∑ 𝑢𝑘
𝑘=1

The number 𝑠𝑛 is called the nth partial sum of the series

and the sequence {𝑠𝑛 }+∞
𝑛=1 is called the sequence of

partial sums.
Example 1.
For the infinite series
3 3 3 3
+ 2+ 3+ 4+⋯
10 10 10 10
the partial sums are
3
𝑠1 =
10
3 3 33
𝑠2 = + 2=
10 10 100
3 3 3 333
𝑠3 = + 2+ 3=
10 10 10 1000
3 3 3 3 3333
𝑠4 = + 2+ 3+ 4=
10 10 10 10 10000
⋮

- 244 -
Definition 2.
Let {𝑠𝑛 } be the sequence of partial sums of the series
∑∞
𝑘=1 𝑢𝑘 . If the sequence {𝑠𝑛 }converges to a limit S, then

the series is said to converge and S is called the sum of

the series. We denote this by writing 𝑆 = ∑∞
𝑘=1 𝑢𝑘 .

If the sequence of partial sums of a series diverges, then

the series is said to diverge. A divergent series has no
sum.
Example 2.
Determine if the series
3 3 3 3
+ 2 + 3 +⋯+ 𝑘 + ⋯
10 10 10 10
converges or diverges. If it converges find the sum.
Solution.
Let us verify that this is indeed the case. The nth partial
sum is
3 3 3
𝑠𝑛 = + 2 +⋯+ (1)
10 10 10𝑛

The problem of calculating lim 𝑠𝑛 is complicated

𝑛→+∞

by the fact that the number of terms in (1) changes with n.

For purposes of calculation, it is desirable to rewrite (1)
in closed form as follows:
- 245 -
 1
We multiply both sides of (1) by to obtain
10
1 3 3 3 3
𝑠𝑛 = 2 + 3 + ⋯+ 𝑛 + (2)
10 10 10 10 10𝑛+1

and then subtract (2) from (1) to obtain

1 3 3
𝑠𝑛 − 𝑠𝑛 = − .
10 10 10𝑛+1

Or
9 3 1
𝑠𝑛 = (1 − 10𝑛 ).
10 10

Or
1 1
𝑠𝑛 = (1 − 𝑛 ) (3)
3 10
1
Since → 0 𝑎𝑠 𝑛 → +∞, it follows from (3) that
10𝑛
1
𝑆 = lim 𝑠𝑛 = .
𝑛→+∞ 3

Thus,
1 3 3 3 3
= + + + ⋯+ + ⋯◄
3 10 102 103 10𝑛

Example 3.
Determine the series
1 − 1 + 1 − 1 + …
converges or diverges. If it converges, find its sum.

- 246 -
Solution.
The partial sums are 𝑠1 = 1 , 𝑠2 = 1 − 1 = 0 , 𝑠3 = 1 −
1 + 1 = 1, 𝑠4 = 1 − 1 + 1 − 1 = 0, and so forth. Thus
the sequence of partial sums is
1, 0, 1, 0, 1, 0, …
Since this is a divergent sequence, the given series
diverges and consequently has no sum.◄
●The series in Example 2 and 3 are examples of
geometric series. A geometric series is one of the forms
𝑎 + 𝑎 𝑟 + 𝑎𝑟 2 + ⋯ + 𝑎𝑟 𝑘−1 + ⋯ (𝑎 ≠ 0)
where each term is obtained by multiplying the previous
one by a constant r. the multiplier r is called the ratio for
the series. Some examples of geometric series are:
1 + 2 + 4 + 8 + ⋯ + 2𝑘−1 + ⋯ (𝑎 = 1 , 𝑟 = 2);
3 3 3 1
3+ + +⋯+ + ⋯ (𝑎 = 3 , 𝑟 = );
10 102 10𝑘−1 10
1 1 1 1 1 1
2
− + − ⋯ + (−1)𝑘+1
4 8 2𝑘
+ ⋯ (𝑎 =
2
,𝑟 = − )
2
;
1+ 1+⋯+1 +⋯ (𝑎 = 1 , 𝑟 = 1);
1 − 1 + 1 − ⋯ + (−1)𝑘+1 + ⋯ (𝑎 = 1 , 𝑟 = −1).
The following theorem is the fundamental result on
convergence of geometric series

- 247 -
Theorem 3.
A geometric series
𝑎 + 𝑎𝑟 + 𝑎𝑟 2 + ⋯ + 𝑎𝑟 𝑘−1 + ⋯ (𝑎 ≠ 0)
converges if |𝑟| < 1 and diverges if |𝑟| ≥ 1. When the
series converges the sum is
𝑎
= 𝑎 + 𝑎𝑟 + 𝑎𝑟 2 + ⋯ + 𝑎𝑟 𝑘−1 + ⋯
1−𝑟
Proof.
Let us treat the case|𝑟| = 1 first. If 𝑟 = 1, then the
series is
𝑎 + 𝑎 + ⋯ + 𝑎 + ⋯.
So that the nth partial sum is 𝑠𝑛 = 𝑛𝑎 and lim 𝑠𝑛 =
𝑛→+∞

lim 𝑛 𝑎 = ±∞ ( the sign depending on whether a is

𝑛→+∞

positive or negative). This proves divergence.

If 𝑟 = −1, the series is𝑎 − 𝑎 + 𝑎 − 𝑎 + ⋯.
So the sequence of partial sums is
𝑎, 0, 𝑎, 0, 𝑎, 0, …
which diverges.
Now let us consider the case where|𝑟| ≠ 1. The nth
partial sum of the series is
𝑠𝑛 = 𝑎 + 𝑎 𝑟 + 𝑎𝑟 2 + ⋯ + 𝑎𝑟 𝑛−1 (4)

- 248 -
Multiplying both sides of (4) by r yields
𝑟 𝑠𝑛 = 𝑎 𝑟 + 𝑎𝑟 2 + ⋯ + 𝑎𝑟 𝑛−1 + 𝑎𝑟 𝑛 (5)
and subtracting (5) form (4) gives
𝑠𝑛 − 𝑟𝑠𝑛 = 𝑎 − 𝑎𝑟 𝑛
or (1 − 𝑟)𝑠𝑛 = 𝑎 − 𝑎𝑟 𝑛 (6)
Since 𝑟 ≠ 1in the case we are considering, this can be
rewritten as
𝑎−𝑎𝑟 𝑛 𝑎 𝑎𝑟 𝑛
𝑠𝑛 = = − (7)
1−𝑟 1−𝑟 1−𝑟

If|𝑟| < 1, then lim 𝑟 𝑛 = 0.So that{𝑠𝑛 }converges. From

𝑛→+∞
𝑎
(7) lim 𝑠𝑛 = .
𝑛→+∞ 1−𝑟

If|𝑟| < 1 then either 𝑟 > 1 or 𝑟 < −1.

In the case 𝑟 > 1, lim 𝑟 𝑛 = +∞, and in the case 𝑟 <
𝑛→+∞

−1, 𝑟 𝑛 oscillates between positive and negative values

that grow in magnitude, so{𝑠𝑛 } diverges.◄
Example 4.
5 5 5
The series5 + + + ⋯+ +⋯
4 42 4 𝑘−1
1
is a geometric series with 𝑎 = 5 and 𝑟 = .
4

- 249 -
 1
Since |𝑟| = < 1, the series converges and the sum is
4
𝑎 5 20
= 1 = ◄
1−𝑟 1− 3
4

Example 5.
Find the rational number represented by repeating
decimal 0.784784784 …
Solution.
We can write
0.784784784 … = 0.784 + 0.000784 +
0.000000784 + …
so the given decimal is the sum of a geometric series with
𝑎 = 0.784and 𝑟 = 0.001. Thus
𝑎 0.784 0.784 784
0.784784784…= = = = .◄
1−𝑟 1−0.001 0.999 999

Example 6.
Determine if the series
∞
1 1 1 1
∑ = + + +⋯
𝑘(𝑘 + 1) 1.2 2.3 3.4
𝑘=1

Converges or diverges. If it converges, find the sum.

Solution.
The nth sum of the series is

- 250 -
 𝑛
1 1 1 1 1
𝑠𝑛 = ∑ = + + + ⋯+
𝑘(𝑘 + 1) 1.2 2.3 3.4 𝑛(𝑛 + 1)
𝑘=1

To calculate lim 𝑠𝑛 we will rewrite𝑠𝑛 in closed form.

𝑛→+∞

This may be accomplished by using the method of partial

fractions to obtain (verify):
1 1 1
= −
𝑘(𝑘+1) 𝑘 𝑘+1

from which it follows that

1 1
𝑠𝑛 = ∑𝑛𝑘=1 ( − )
𝑘 𝑘+1
1 1 1 1 1 1 1
= (1 − ) + ( − ) + ( − ) + ⋯ + ( − )
2 2 3 3 4 𝑛 𝑛+1
1 1 1 1 1 1 1
= 1 + (− + ) + (− − ) + ⋯ + (− + ) − .
2 2 3 3 𝑛 𝑛 𝑛+1

The sum above is an example of a telescoping sum,

which means that each term cancels part of the next term,
thereby collapsing the sum (like a folding telescope) into
only two terms. After the cancellation, (8) can be written
1
as 𝑠𝑛 = 1 −
𝑛+1
1
So lim 𝑠𝑛 = lim (1 − ) = 1.
𝑛→+∞ 𝑛+1
𝑛→+∞
1
Therefore 1 = ∑∞
𝑘=1 .◄
𝑘(𝑘+1)

- 251 -
Example 7.
One of the most famous and important of all diverging
series is the harmonic series
∞
1 1 1 1
∑ =1+ + + +⋯
𝑘 2 3 4
𝑘=1

However, the divergence will become apparent when we

examine the partial sums in detail. Because the terms in
the series are all positive, the partial sums
1 1 1
𝑠1 = 1, 𝑠2 = 1 + , 𝑠3 = 1 + + ,…
2 2 3

form an increasing sequence

𝑠1 < 𝑠2 < 𝑠3 < ⋯ < 𝑠𝑛 < ⋯
Thus, by Theorem 2 in Section 2 of Chapter VI we can
prove divergence by demonstrating that there is no
constant M that is greater than or equal to every partial
sum. To this end, we will consider some selected sums,
namely 𝑠2 , 𝑠4 , 𝑠8 , 𝑠16 , 𝑠32 , …. Note that the subscripts
are successive powers of 2, so that these are the partial
sums of the form𝑠2𝑛 . These partial sums satisfy the
inequalities
1 1 1 2
𝑠2 = 1 + > + =
2 2 2 2
- 252 -
 1 1 1 1 1 3
𝑠4 = 𝑠2 + + > 𝑠2 + ( + ) = 𝑠2 + >
3 4 4 4 2 2
1 1 1 1 1 1 1 1
𝑠8 = 𝑠4 + + + + > 𝑠4 + ( + + + )
5 6 7 8 8 8 8 8
1 4
= 𝑠4 + >
2 2
1 1 1 1 1 1 1 1
𝑠16 = 𝑠8 + + + + + + + +
9 10 11 12 13 14 15 16
1 1 1 1 1 1 1 1
> 𝑠8 + ( + + + + + + + )
16 16 16 16 16 16 16 16
1 5
= 𝑠8 + >
2 2
⋮
𝑛+1
𝑠 2𝑛 > .
2

Now if M is any constant, we can certainly find a positive

(𝑛+1)
integer n such that > 𝑀. But for this n
2
𝑛+1
𝑠2𝑛 > >𝑀
2
So no constant M is greater than or equal to every partial
sum of the harmonic series. This proves divergence.◄

- 253 -
 Exercise Set (7.1)

(1) In each part, find the first four partial sums, find a
closed form for the nth partial sum; determine if this
series converges, and if so give the sum.
2 1
(a) ∑∞
𝑘=1 (b) ∑∞
𝑘=1
5𝑘−1 (𝑘+1)(𝑘+2)

2𝑘−1
(c) ∑∞
𝑘=1 4

In Exercise (2) - (16), determine if the series converges or

diverges. If it converges, find the sum.
1 3 𝑘−1
(2) ∑∞
𝑘=1 𝑘 ; (3) ∑∞
𝑘=1 (− 4) ;
5

2 𝑘+2 7
(4) ∑∞
𝑘=1 (3) ; (5) ∑∞
𝑘=1(−1)
𝑘−1
;
6𝑘−1

3 𝑘+1
(6) ∑∞
𝑘=1 4
𝑘−1
; (7) ∑∞
𝑘=1 (− 2) ;
1 1 1
(8) ∑∞
𝑘=1 ( − ); (9) ∑∞
𝑘=1 ;
𝑘+3 𝑘+4 (𝑘+2)(𝑘+3)
1 1 1
(10) ∑∞
𝑘=1 ( − ); (11) ∑∞
𝑘=1 ;
2𝑘 2𝑘+1 9𝑘 2 +3𝑘−2

1 4𝑘+2
(12) ∑∞
𝑘=1 ; (13) ∑∞
𝑘=1 ;
𝑘 2 −1 7𝑘−1

𝑒 𝑘−1 1 𝑘
(14) ∑∞
𝑘=1 (𝜋) ; (15) ∑∞
𝑘=1 (− 2) ;

- 254 -
 5
(16) ∑∞
𝑘=1 ;
𝑘−2

In Exercise (17) - (22), express the repeating decimal as a

fraction.
(17) 0.4444 … ; (18) 0.9999. ..;
(19) 5.373737 …; (20) 0.159159159 …;
(21) 0.782178217821 …; (22) 0.451141414 …
(23) Find a closed form for the nth partial sum of the
series
1 2 3 𝑛
ln + ln + ln + ⋯ + ln +⋯
2 3 4 𝑛+1
and determine if the series converges.
(24) A ball is dropped from a height of 10 meters. Each
time it strikes the ground it bounces vertically to a
height that is three fourths of the previous height.
Find the total distance the ball will travel if it is
allowed to bounce indefinitely.
1
(25) Show ∑∞
𝑘=1 In (1 − ) = −In 2
𝑘2

√𝑘+1−√𝑘
(26) Show ∑∞
𝑘=1 =1
√𝑘 2 +𝑘

(27) Use geometric series to show :

1
(a) ∑∞ 𝑘 𝑘
𝑘=1(−1) 𝑥 = if − 1 < 𝑥 < 1
1+𝑥

- 255 -
 1
(b) ∑∞ 𝑘
𝑘=1(𝑥 − 3) = if 2 < 𝑥 < 4
4−𝑥
1
(c) ∑∞ 𝑘 2𝑘
𝑘=1(−1) 𝑥 = if −1 < 𝑥 < 1.
1+𝑥 2

(28) Prove the following decimal equality assuming𝑎𝑛 ≠

9
0 . 𝑎1 𝑎2 … 𝑎𝑛 9999 … = 0. 𝑎1 𝑎2 … (𝑎𝑛 + 1) 0000 …

- 256 -
2. CONVERGENCE; THE INTEGRAL TEST
Theorem 1.
If the series∑ 𝑢𝑘 converges, then lim 𝑢𝑘 = 0.
𝑘→+∞

Proof.
The term 𝑢𝑘 can be written
𝑢𝑘 = 𝑠𝑘 − 𝑠𝑘−1 (1)
where 𝑠𝑘 is the sum of the first k terms and 𝑠𝑘−1 is the sum
of the first k-1 terms. If S denotes the sum of the series,
then lim 𝑠𝑘 = 𝑆 and since(𝑘 − 1) → +∞ 𝑎𝑠 𝑘 → +∞ we
𝑘→+∞

also have lim 𝑠𝑘−1 = 𝑆. Thus from (1)

𝑘→+∞

lim 𝑢𝑘 = lim (𝑠𝑘 − 𝑠𝑘−1 ) = 𝑆 − 𝑆 = 0.

𝑘→+∞ 𝑘→+∞

Theorem 2 (The Divergence test).

If lim 𝑢𝑘 ≠ 0, then the series ∑ 𝑢𝑘 diverges.
𝑘→+∞

Proof.
It is just an alternate phrasing of the above theorem and
needs no additional proof.
Example 1.
The series
∞
𝑘 1 2 3 𝑘
∑ = + + +⋯+ +⋯
𝑘+1 2 3 4 𝑘+1
𝑘=1
- 257 -
 𝑘 1
diverges since lim = lim 1 = 1 ≠ 0.◄
𝑘→+∞ 𝑘+1 𝑘→+∞ 1+𝑘

WARNING:
The converse of Theorem 1 is false. To prove that a series
converges it does not suffice to show that lim 𝑢𝑘 = 0.
𝑘→+∞

Since this property may hold for divergent as well as

convergent series. For example, the kth term of the
divergent harmonic series 1 + 1/2 + … + 1/𝑘 +
…tends to zero as 𝑘 → +∞ , and the kth term of the
1 1 1
convergent geometric series + 2 +⋯+ + ⋯ tends
2 2 2𝑘

to zero as 𝑘 → +∞.
Theorem 3 (The proof left as exercise).
(a) If∑ 𝑢𝑘 𝑎𝑛𝑑 ∑ 𝑣𝑘 are convergent series, then∑(𝑢𝑘 +
𝑣𝑘 ) are convergent series and the sums of these series
are related by ∑(𝑢𝑘 ± 𝑣𝑘 ) = ∑∞ ∞
𝑘=1 𝑢𝑘 ± ∑𝑘=1 𝑣𝑘 .

(b) If c is a nonzero constant then series∑ 𝑢𝑘 and

∑ 𝑐 𝑢𝑘 both converge or both diverge. In the case of
convergence, the sums are related by
∑∞ ∞
𝑘=1 𝑐 𝑢𝑘 = 𝑐 ∑𝑘=1 𝑢𝑘 .

- 258 -
(c) Convergence or divergence is unaffected by deleting a
finite number of terms from the beginning of a series;
that is, for any positive integer K, the series
∑∞
𝑘=1 𝑢𝑘 = 𝑢1 + 𝑢2 + 𝑢3 + ⋯

and
∑∞
𝑘=𝐾 𝑢𝑘 = 𝑢𝑘 + 𝑢𝑘+1 + 𝑢𝑘+2 + ⋯

both converge or both diverge.

Example 2.
3 2
Find the sum of the series ∑∞
𝑘=1 ( − ).
4𝑘 5𝑘−1

Solution.
3 3 3 3
The series ∑∞
k=1 = + + + ⋯is a convergent
4k 4 42 43
3 1
geometric series(𝑎 = ,𝑟 = ) and the series
4 4
2 2 2 2
∑∞
𝑘=1 =2+ + + + ⋯is also a convergent
5𝑘−1 5 52 53
1
geometric series(𝑎 = 2 , 𝑟 = ).Thus by part (a) of
5

Theorem 3 above the given series converges and

∞ ∞ ∞
3 2 3 2
∑ ( 𝑘 − 𝑘−1 ) = ∑ 𝑘 − ∑ 𝑘−1
4 5 4 5
𝑘=1 𝑘=1 𝑘=1
3
2 3
= 4
1 − 1 = − .◄
1−4 1−5 2

- 259 -
Example 3.
5 5 5 5
The series∑∞
𝑘=1 = 5 + + + ⋯ + + ⋯
𝑘 2 3 𝑘

diverges by part (b) of Theorem 3 above since

5 1
∑∞ ∞
𝑘=1 𝑘 = ∑𝑘=1 5 (𝑘 ).

So each term is a constant times the corresponding term

of the divergent harmonic series.◄
Example 4.
1 1 1 1
The series ∑∞
𝑘=1 = + + +⋯
𝑘 10 11 12

diverges by part (c) of Theorem 3 above, since this series

results by deleting the first nine terms from the divergent

harmonic series.

Theorem 4.
If∑ 𝑎𝑘 is a series with positive terms, and if there is a
constant 𝑀 such that 𝑠𝑛 = 𝑎1 + 𝑎2 + ⋯ + 𝑎𝑛 ≤ 𝑀for
every n, then the series converges and the sum
𝑆 satisfies 𝑆 ≤ 𝑀. If no such 𝑀 exists then the series
diverges.

1
●If we have a series with positive terms, say ∑∞
𝑘=1 𝑘2

- 260 -
 +∞ 1
and if we form the improper integral∫1 𝑑𝑥 whose
𝑥2

integrand is obtained by replacing the summation index k

by x, then there is a relationship between convergence of
the series and convergence of the improper integral.
Theorem 5 (the Integral test).
Let∑ 𝑎𝑘 be a series with positive terms, and let 𝑓(𝑥) be
the function that results when k is replaced by x in the
formula 𝑎𝑘 . If f is decreasing and continuous for 𝑥 ≥ 1,
+∞
then∑∞
𝑘=1 𝑎𝑘 and ∫1 𝑓(𝑥 )𝑑𝑥 both converge or both
diverge.

- 261 -
Example 5.
1
Determine if ∑∞
𝑘=1 converges or diverges.
𝑘2

Solution.
1
If we replace k by x in the formula 𝑎𝑘 = we obtain the
𝑘2
1
function𝑓 (𝑥 ) = which satisfies the hypotheses of the
𝑥2

integral test. (Verify.) Since

+∞ ℓ
1 𝑑𝑥 1ℓ
∫ 2 𝑑𝑥 = lim ∫ 2 = lim [− ]
𝑥 ℓ→+∞ 𝑥 ℓ→+∞ 𝑥1
1 1
1
= lim [1 − ] = 1,
ℓ→+∞ ℓ

the integral converges and consequently the series

converges.◄
Example 6.
The integral test provides another way to demonstrate
1
divergence of the harmonic series∑∞
𝑘=1 . If we replace k
𝑘
1
by x in the formula for 𝑎𝑘 = we obtain the function
𝑘
1
𝑓 (𝑥 ) = which satisfies the hypotheses of the integral
𝑥

test. (Verify.) Since

- 262 -
 +∞ ℓ
1 1
∫ 𝑑𝑥 = lim ∫ 𝑑𝑥 = lim [In ℓ − In 1] = +∞
𝑥 ℓ→+∞ 𝑥 ℓ→+∞
1 1

the integral and the series diverge.◄

Example 7.
Determine if the series
1 2 3 𝑘
+ 4 + 9 + ⋯ + 𝑘2 + ⋯
𝑒 𝑒 𝑒 𝑒
converges or diverges.
Solution.
𝑘
If we replace k by x in the formula for 2 , we obtain the
𝑒𝑘
𝑥 2
function𝑓 (𝑥 ) = 2 = 𝑥 𝑒 −𝑥
𝑒𝑥

For 𝑥 ≥ 1 , this function has positive values and is

continuous. Moreover, for 𝑥 ≥ 1 the derivative
2 2 2
𝑓 ′ (𝑥 ) = 𝑒 −𝑥 − 2𝑥 2 𝑒 −𝑥 = 𝑒 −𝑥 (1 − 2𝑥 2 )
is negative, so that f is decreasing for 𝑥 ≥ 1 Thus, the
hypotheses of the integral test are met. But
+∞ ℓ 1 ℓ
−𝑥 2 −𝑥 2 −𝑥 2
∫1 𝑥 𝑒 𝑑𝑥 = lim ∫1 𝑥 𝑒 𝑑𝑥 = lim [− 𝑒
2
]
ℓ→+∞ ℓ→+∞ 1
1 2 1
= (− ) lim [𝑒 −ℓ − 𝑒 −1 ] =
2 ℓ→+∞ 2𝑒

The improper integral and the series converge. ◄

- 263 -
Remark.

If the summation index in a series∑ 𝑎𝑘 does not begin

with 𝑘 = 1,a variation of the integral test may still apply.
+∞
It can be shown that lim 𝑎𝑘 and ∫𝑘 𝑓(𝑥 )𝑑 𝑥 both
𝑘=𝐾

converge or both diverge provided the hypotheses of

Theorem 5 hold for 𝑥 ≥ 𝐾.
♣The harmonic series and the series in Example 5 are
special cases of a class of series called p-series or hyper
harmonic series. A p-series is an infinite series of the
form
∞
1 1 1 1
∑ = 1 + + + ⋯ + +⋯
𝑘𝑝 2𝑝 3𝑝 𝑘𝑝
𝑘=1

where 𝑝 > 0.
Examples of p-series are
∞
1 1 1 1
∑ = 1 + + + ⋯+ + ⋯ (𝑝 = 1)
𝑘 2 3 𝑘
𝑘=1
∞
1 1 1 1
∑ = 1 + + + ⋯ + +⋯ (𝑝 = 2)
𝑘2 22 32 𝑘2
𝑘=1

- 264 -
 ∞
1 1 1 1 1
∑ =1+ + +⋯+ + ⋯ (𝑝 = )
√𝑘 √2 √3 √𝑘 2
𝑘=1

The following theorem tells when a p-series converges.

Theorem 6 (Convergence of p-series).
∞
1 1 1 1
∑ = 1 + + + ⋯ + +⋯
𝑘𝑝 2𝑝 3𝑝 𝑘𝑝
k=1

converge if 𝑝 > 1 and diverge if 0 < 𝑝 ≤ 1.

Proof.
To establish this result when 𝑝 ≠ 1, we will use the
integral test.
+∞ ℓ
1
∫ 𝑝 𝑑𝑥 = lim ∫ 𝑥 −𝑝 𝑑𝑥
𝑥 ℓ→+∞
1 1
ℓ
𝑥 ℓ−𝑝 ℓ1−𝑝 1
= lim [ ] = lim [ − ]
ℓ→+∞ 1 − 𝑝 ℓ→+∞ 1 − 𝑝 1 − 𝑝
1

For 𝑝 > 1, 1 − 𝑝 < 0 andℓ1−𝑝 → 0 𝑎𝑠 ℓ → +∞.So

the integral and the series converge. For 0 < 𝑝 <
1, 1 − 𝑝 > 0 andℓ1−𝑝 → +∞ 𝑎𝑠 ℓ → +∞, so the
integral and the series diverge. The case 𝑝 = 1 is the
harmonic series, which was previously shown to
diverge.◄
- 265 -
Example 8.
1 1 1
1 + 3 + 3 + ⋯+ 3 + ⋯
√2 √3 √𝑘
1
diverge since it is a p-series with 𝑝 = < 1.◄
3

Exercise Set (7.2)

In Exercise 1-4, find the sum of the series.
1 1 1 1
(1) ∑∞
𝑘=1 [ 𝑘 + 𝑘] (2) ∑∞
𝑘=1 [ 𝑘 − ]
2 4 5 𝑘(𝑘+1)
1 7 7 6
(3) ∑∞
𝑘=2 [ − 𝑘−1 ] (4) ∑∞
𝑘=1 [ 𝑘 + (𝑘+3)(𝑘+4)]
𝑘 2 −1 10 3

In Exercise (5) - (26), determine if the series converges or

diverges.
1 3
(5) ∑∞
𝑘=1 (6) ∑∞
𝑘=1
𝑘+6 5𝑘
1 𝑘
(7) ∑∞
𝑘=1 (8) ∑∞
𝑘=1
5 𝑘+2 1+𝑘 2
1 1
(9) ∑∞
𝑘=1 (10) ∑∞
𝑘=1 3
1+9 𝑘 2
(4+2 𝑘)2
1 1
(11) ∑∞
𝑘=1 (12) ∑∞
𝑘=1 𝑘
√𝑘+5 √𝑒
1 𝑙𝑛 𝑘
(13) ∑∞
𝑘=1 3 (14) ∑∞
𝑘=3
√2 𝑘−1 𝑘
𝑘 2
(15) ∑∞
𝑘=1 (16) ∑∞
𝑘=1 𝑘 𝑒
−𝑘
𝑙𝑛 (𝑘+1)

1 𝑘 2 +1
(17) ∑∞
𝑘=1 (𝑘+1)[𝑙𝑛 (18) ∑∞
𝑘=1
(𝑘+1)]2 𝑘 2 +3

- 266 -
 1 𝑘 1
(19) ∑∞
𝑘=1 (1 + 𝑘 ) (20) ∑∞
𝑘=1 √𝑘 2 +1
tan−1 𝑘
(21) ∑∞
𝑘=1 (22) ∑∞ 2
𝑘=1 sech 𝑘
1+𝑘 2
1 3
(23) ∑∞ 2 2
𝑘=1 𝑘 sin (𝑘 ) (24) ∑∞ 2 −𝑘
𝑘=1 𝑘 𝑒

1
(25) Prove∑∞
𝑘=2 converges if 𝑝 > 1 and
𝑘 (𝑙𝑛 𝑘)𝑝

diverges if 𝑝 ≥ 1.
1
(28) Prove ∑∞
𝑘=3 converges if 𝑝 >
𝑘(ln 𝑘)[ln (ln 𝑘)]𝑝

1 and diverges if 𝑝 ≤ 1.
(29) Prove : If
∑ 𝑢𝑘 converges and∑ 𝑣𝑘 diverges, then ∑(𝑢𝑘 + 𝑣𝑘 )
diverges and∑(𝑢𝑘 − 𝑣𝑘 ) diverges.
(30) Find examples to show that∑(𝑢𝑘 +
𝑣𝑘 )and∑(𝑢𝑘 − 𝑣𝑘 ) may converge or may diverge
if∑ 𝑢𝑘 and∑ 𝑣𝑘 both diverge.
(31) With the help of Exercise (29) determine if the
given series converges or diverges.
∞ 2 𝑘−1 1 ∞ 𝑘2 1
(a)∑𝑘=1 [( ) + ](b) ∑𝑘=1 [ 2 + ]
3 𝑘 1+𝑘 𝑘(𝑘+1)

1 1 1 1
(c) ∑∞
𝑘=1 [ + 3 ](d) ∑∞
𝑘=2 [ 2 − ].
3 𝑘+2 𝑘(𝑙𝑛 𝑘) 𝑘2
𝑘2

- 267 -
3. ADDITIONAL CONVERGENCE TESTS
In this section we will develop some additional
convergence tests for series with positive terms.
Theorem 1 (The Comparison Test).
Let∑ 𝑎𝑘 and∑ 𝑏𝑘 be series with positive terms and
suppose𝑎1 ≤ 𝑏1 , 𝑎2 ≤ 𝑏2 , … , 𝑎𝑘 ≤ 𝑏𝑘 , …
(a) If the "bigger series" ∑ 𝑏𝑘 converges, then the
"smallest series" ∑ 𝑎𝑘 also converges.
(b) If the "smaller series” ∑ 𝑎𝑘 diverges, then the
"bigger series"∑ 𝑏𝑘 also diverges.
Theorem 2 (The Ratio Test).
Let∑ 𝑢𝑘 be a series with positive terms and suppose
𝑢𝑘+1
lim =𝜌
𝑘→+∞ 𝑢𝑘

(a) If 𝜌 < 1, the series converges.

(b) If 𝜌 > 1 or𝜌 = +∞ , the series diverges.
(c) If 𝜌 = 1, the series may converge or
diverge, so that another test must be tried.
Proof.
●The proof of (a) and (b) are omitted.
(c) The series

- 268 -
 ∞ ∞
1 1
∑ 𝑎𝑛𝑑 ∑
𝑘 𝑘2
𝑘=1 𝑘=1

both have 𝜌= 1 (verify). Since the first is the divergent

harmonic series and the second is a convergent p-series,
the ratio test does not distinguish between convergence
and divergence when 𝜌 = 1.◄
Example 1.
The series
∞
1
∑
𝑘!
𝑘=1

converges by the ration test since

1
𝑢𝑘+1 (𝑘+1) 1
𝜌 = lim = lim 1 == lim =0
𝑘→+∞ 𝑢𝑘 𝑘→+∞ 𝑘→+∞ 𝑘 + 1
𝑘!

so that 𝜌 < 1.◄

Example 2.
𝑘
The series ∑∞
𝑘=1 𝑘 converges by the ration test since
2

𝑢𝑘+1 𝑘 + 1 2𝑘 1 𝑘+1 1
𝜌 = lim = lim 𝑘+1 . = lim =
𝑘→+∞ 𝑢𝑘 𝑘→+∞ 2 𝑘 2 𝑘→+∞ 𝑘 2
so that 𝜌 < 1.◄

- 269 -
Example 3.
𝑘𝑘
The series ∑∞
𝑘=1 diverges by the ratio test since
𝑘!
𝑢𝑘+1 (𝑘 + 1)𝑘+1 𝑘 !
𝜌 = lim = lim . 𝑘
𝑘→+∞ 𝑢𝑘 𝑘→+∞ (𝑘 + 1)! 𝑘
(𝑘 + 1)𝑘 1 𝑘
= lim = lim (1 + ) = 𝑒
𝑘→+∞ 𝑘𝑘 𝑘→+∞ 𝑘
Since 𝜌 = 𝑒 > 1, the series diverges.◄
Example 4.
Determine whether the series
1 1 1 1
1+ + + +⋯+ +⋯
3 5 7 2𝑘−1
converges or diverges.
Solution.
The ratio test is of no help since
𝑢𝑘+1 1 2𝑘−1
𝜌 = lim = lim .
𝑘→+∞ 𝑢𝑘 𝑘→+∞ 2 (𝑘 + 1) − 1 1
2𝑘−1
= lim =1
𝑘→+∞ 2 𝑘 + 1

However, the integral test proves that the series diverges

since
+∞ ℓ
𝑑𝑥 𝑑𝑥
∫ = lim ∫
2𝑥 − 1 ℓ→+∞ 2𝑥 − 1
1 1

- 270 -
 1 ℓ
= lim 𝑙𝑛 (2𝑥 − 1)] = +∞.◄
ℓ→+∞ 2 1

Example 5.
The series
2! 4! 6! (2 𝑘)!
+ 2 + 3 + ⋯+ 𝑘 + ⋯
4 4 4 4
diverges since
𝑢𝑘+1 [2 (𝑘 + 1)]! 4𝑘
𝜌 = lim = lim .
𝑘→+∞ 𝑢𝑘 𝑘→+∞ 4𝑘+1 (2 𝑘)!
(2 𝑘 + 2)! 1 1
= lim ( . ) = lim (2 𝑘 + 1)(2 𝑘 + 2)
𝑘→+∞ (2 𝑘)! 4 4 𝑘→+∞
= +∞.◄
Theorem 3 (The Root Test).
Let∑ 𝑢𝑘 be a series with positive terms and suppose
𝜌 = lim 𝑘√𝑢𝑘 = lim (𝑢𝑘 )1/𝑘
𝑘→+∞ 𝑘→+∞

(a) If 𝜌 < 1, the series converges.

(b) If 𝜌 > 1 or 𝜌 = ∞, the series diverges.
(c) If 𝜌 = 1, the series may converge or
diverge, so that another test must be tried.

- 271 -
Example 6.
4𝑘−5 𝑘
The series ∑∞
𝑘=1 ( ) diverges by the root test since
2 𝑘+1
4 𝑘−5
𝜌 = lim 𝑢𝑘 1/𝑘 = lim = 2 > 1.◄
𝑘→+∞ 𝑘→+∞ 2 𝑘+1

Example 7.
1
The series ∑∞
𝑘=1 (𝑙𝑛 converges by the root test, since
(𝑘+1)𝑘
1
lim (𝑢𝑢 )1/𝑘 = lim = 0 < 1.◄
𝑘→+∞ 𝑘→+∞ 𝐼𝑛 (𝑘+1)

Remark. Until now we have written most of our infinite

series in the form
∞

∑ 𝑢𝑘 (5)
𝑘=1

with the summation index beginning at 1. If summation

index begins at some other integer, it is always possible
to rewrite the series in form (5). Thus, for example, the
series
∞
2𝑘 22 23
∑ =1+2+ + +⋯ (6)
𝑘! 2! 3!
𝑘=0

can be written as
∞
2𝑘−1 22 23
∑ =1+2+ + +⋯ (7)
(𝑘 − 1)! 2! 3!
𝑘=1
- 272 -
However, for purpose of applying the convergence tests,
it is not necessary that the series have form (5). For
example, we can apply the ratio test to (6) without
converting to the more complicated form (7). Doing so
yields
𝑢𝑘+1 2𝑘+1 𝑘! 2
𝜌 = lim = lim . 𝑘 = lim =0
𝑘→+∞ 𝑢𝑘 𝑘→+∞ (𝑘 + 1)! 2 𝑘→+∞ 𝑘 + 1

which shows the series converges since 𝜌 < 1.◄

Exercise Set (7.3)
In Exercise (1) - (6) apply the ratio test. According to the
test, does the series converge, does the series diverge, or
are the results inconclusive?
3𝑘 4𝑘
(1) ∑∞
𝑘=1 (2) ∑∞
𝑘=1
𝑘! 𝑘2

1 1 𝑘
(3) ∑∞
𝑘=2 (4) ∑∞
𝑘=1 𝑘 (2)
5𝑘
𝑘! 𝑘
(5) ∑∞
𝑘=1 (6) ∑∞
𝑘=1
𝑘3 𝑘 2 +1

In Exercise (7) - (10), apply the root test. According to

the test, does the series converge, does the series diverge,
or are the results inconclusive?
3 𝑘+2 𝑘 𝑘 𝑘
(7) ∑∞
𝑘=1 ( ) (8) ∑∞
𝑘=1 ( )
2 𝑘−1 100

- 273 -
 𝑘 1
(9) ∑∞
𝑘=1 (10) ∑∞
𝑘=1
5𝑘 𝑘2

In Exercise (11) - (32), use any appropriate test to

determine if the series converges, or not.
2𝑘 1
(11) ∑∞
𝑘=1 (12) ∑∞
𝑘=1
𝑘3 𝑘2
7𝑘 1
(13) ∑∞
𝑘=0 (14) ∑∞
𝑘=1
𝑘! 2 𝑘+1
𝑘2 𝑘! 10𝑘
(15) ∑∞
𝑘=1 (16) ∑∞
𝑘=1
5𝑘 3𝑘
𝑘2
(17) ∑∞ 50 −𝑘
𝑘=1 𝑘 𝑒 (18) ∑∞
𝑘=1 𝑘 3 +1

2 𝑘
(19) ∑∞
𝑘=1 𝑘 (3) (20) ∑∞
𝑘=1 𝑘
𝑘

1 2 𝑘
(21) ∑∞
𝑘=2 𝑘 𝐼𝑛 𝑘 (22) ∑∞
𝑘=1 𝑘 3 +1

4 𝑘 (𝑘 !)2 2𝑘
(23) ∑∞
𝑘=1 ( ) (24) ∑∞
𝑘=1 (2
7 𝑘−1 𝑘+2)!

(𝑘 !)2 1
(25) ∑∞
𝑘=0 (2 (26) ∑∞
𝑘=1
𝑘)! 𝑘 2 +25

1 𝑘𝑘
(27) ∑∞
𝑘=1 (28) ∑∞
𝑘=1
1+√𝑘 𝑘!
𝐼𝑛 𝑘 𝑘!
(29) ∑∞
𝑘=1 (30) ∑∞
𝑘=1 2
𝑒𝑘 𝑒𝑘
2
(𝑘+4)! 𝑘 𝑘
(31) ∑∞
𝑘=0 4 !𝑘 ! 4 𝑘 (32) ∞
∑𝑘=1 ( )
𝑘+1

(33) Determine if the following series converges:

- 274 -
 1 .2 1 .2 .3 1 .2 .3 .4
1+ + + +⋯
1 .3 1 .3 .5 1 .3 .5 .7
𝑎𝑘
(34) For which positive values of a does ∑∞
𝑘=1 𝑘𝑎

converges?
(35) (a) Show lim (𝑙𝑛 𝑘)1/𝑘 = 1 = 1
𝑘→+∞

[Hint ∶ Let 𝑦 = (𝑙𝑛 𝑥 )1/𝑥 and lim ln 𝑦]

𝑥→+∞

(b) Use the result in (a) and the root test to show
𝑙𝑛 𝑘
that∑∞
𝑘=1 converges.
3𝑘

(c) Show that the series converges using the ratio test.
𝑘!
(36) Prove: lim =0
𝑘→+∞ 𝑘 𝑘

[Hint: Exploit Theorem 1 in Section 2]

𝑎𝑘
(37) Prove: lim = 0 for every real number a. [See
𝑘→+∞ 𝑘 !

Hint to Exercise (36) ]

- 275 -
4. Applying the comparison test
Informal Principle 1.
Constants added or subtracted in the formula for𝑢𝑘 may
usually be deleted without affecting the convergence or
divergence behavior of the series.
Example 1.
Use the above principle to help guess whether the
following series converge or diverge.
∞ ∞ ∞
1 1 1
(𝑎) ∑ 𝑘 (𝑏) ∑ (𝑐) ∑
2 +1 √𝑘 − 2 1 3
𝑘=1 𝑘=5 𝑘=1 (𝑘 + )
2

Solution.
(a) Deleting the constant 1 suggest that
∞ ∞
1 1
∑ 𝑘 behaves like ∑ 𝑘
2 +1 2
𝑘=1 𝑘=1

The modified series is a convergent geometric series, so

the given series is likely to converge.
(b) Deleting the -2 suggest that
∞ ∞
1 1
∑ behaves like ∑
𝑘=5
√𝑘 − 2 𝑘=5
√𝑘

- 276 -
 1
The modified series is a divergent p-series(𝑝 = ) , so the
2

given series is likely to diverge.

1
(c) Deleting the suggests that
2
∞ ∞
1 1
∑ behaves like ∑
1 3 𝑘3
𝑘=1 (𝑘 + ) 𝑘=1
2

The modified series is a convergent p-series (𝑝 = 3), so

the given series is likely to converge.◄

Informal Principle 2.

If a polynomial in k appears as a factor in the numerator

or denominator of 𝑢𝑘 , all but the highest power of k in the
polynomial may usually be deleted without affecting the
convergence or divergence behavior of the series.
Example 2.
Use the above principle to help guess whether the
following series converge or diverge.
∞ ∞
1 6 𝑘 4 − 2𝑘 3 + 1
(𝑎) ∑ (𝑏) ∑ 5
3
√𝑘 + 2 𝑘 𝑘 + 𝑘2 − 2 𝑘
𝑘=1 𝑘=1

Solution.
(a) Deleting the term 2 k suggest that
- 277 -
 ∞ ∞ ∞
1 1 1
∑ behaves like ∑ =∑
√𝑘 3 + 2 𝑘 √𝑘 3 𝑘 3/2
𝑘=1 𝑘=1 𝑘=1

Since the modified series is a convergent p-series(𝑝 =

2
3
), the given series is likely to converge.

(b)Deleting all but the highest power of k in the

numerator and also in the denominator suggests that
∞ ∞ ∞
6 𝑘 4 − 2𝑘 3 + 1 6 𝑘4 1
∑ 5 behaves like ∑ = 6 ∑
𝑘 + 𝑘2 − 2 𝑘 𝑘5 𝑘
𝑘=1 𝑘=1 𝑘=1

Since the modified series is a constant times the divergent

harmonic series, the given series is likely to diverge.
♠To prove∑ 𝑎𝑘 converge by the comparison test, we must
find a convergent series ∑ 𝑏𝑘 such that
𝑎𝑘 ≤ 𝑏𝑘
for all k, frequently,𝑏𝑘 is derived from the formula for𝑎𝑘
by either increasing the numerator of𝑎𝑘 , or decreasing the
denominator of 𝑎𝑘 , or both.
Example 3.
Use the comparison test to determine whether

- 278 -
 ∞
1
∑
2 𝑘2 + 𝑘
𝑘=1

converges or diverges
Solution.
Using Principle 2, the series behaves like the series
∞ ∞
1 1 1
∑ = ∑
2 𝑘2 2 𝑘2
𝑘=1 𝑘=1

which is a constant times a convergent p-series. Thus the

given series is likely to converge. To prove the
convergence, observe that when we discard the k from the
1
denominator of(2 , the denominator decrease and the
𝑘 2 +𝑘)

ratio increases, so that

1 1
<
2 𝑘2 + 𝑘 2 𝑘2
for 𝑘 = 1, 2, … since
∞ ∞
1 1 1
∑ 2= ∑ 2
2𝑘 2 𝑘
𝑘=1 𝑘=1
1
converges, so does ∑∞
𝑘=1 by the comparison test.
2 𝑘 2 +𝑘

- 279 -
Example 4.
Use the comparison test to determine whether
∞
1
∑
2 𝑘2 − 𝑘
𝑘=1

converges or diverges.
Solution.
Using Principle 2, the series behaves like the convergent
series
∞ ∞
1 1 1
∑ = ∑
2 𝑘2 2 𝑘2
𝑘=1 𝑘=1

Thus the given series is likely to converge. However, if

1
we discard the k from the denominator of , the
(2 𝑘 2 −𝑘)

denominator increases and the ratio decreases, so that

1 1
>
2 𝑘2 − 𝑘 2 𝑘2
Unfortunately, this inequality is in the wrong direction to

prove convergence of the given series. A different

approach is needed; we must do something to decrease

the denominator, not increase it. We accomplish this by

- 280 -
 1 1 1
replacing k by𝑘 2 to obtain ≤ = .
2𝑘 2 −𝑘 2 𝑘 2 −𝑘 2 𝑘2

1
Since∑∞
𝑘=1 is a convergent p-series, the given series
𝑘2

converges by the comparison test.◄

♠To prove a series∑ 𝑎𝑘 diverges by the comparison test,

we must produce a divergent series∑ 𝑏𝑘 of positive terms

such that 𝑎𝑘 ≥ 𝑏𝑘 for all k.

Example 5.
Use the comparison test to determine whether
∞
1
∑ 1
𝑘=1 𝑘 − 4

converges or diverges.
Solution.
Using Principle 1, the series behaves like the divergent
1
harmonic series ∑∞
𝑘=1 𝑘

Thus, the given series is likely to diverge. Since

1 1
1> 𝑘
for 𝑘 = 1 , 2 , …
𝑘−
4

- 281 -
 1
and since∑∞
𝑘=1 diverges, the given series diverges by the
𝑘

comparison test.◄
Example 6.
Use the comparison test to determine whether
1
∑∞
𝑘=1 converges or diverges
√𝑘+5

Solution.
Using Principle 1, the series behaves like the divergent
1
p-series∑∞
𝑘=1 √𝑘

Thus the given series is likely to diverge. For𝑘 ≥ 25 we

1 1 1
have = ≥
√𝑘+5 √25+√𝑘 2 √𝑘
1
and since∑∞
𝑘=25 2 √𝑘
1
diverges (why?) the series∑∞
𝑘=1 diverges by the
√𝑘+5

comparison test; consequently, the given series diverges

by Theorem 3(a) in Section 2.
Theorem 3 (Comparison Test).
Let∑ 𝑎𝑘 and ∑ 𝑏𝑘 be series with positive terms and
𝑎𝑘
suppose 𝜌 = lim
𝑘→+∞ 𝑏𝑘

if 𝜌 is finite and 𝜌 > 0, then the series both converge or

both diverge.
- 282 -
Proof.
The proof will be completed if we can show that∑ 𝑏𝑘
converges when∑ 𝑎𝑘 converges and conversely. To this
end, let
𝜌
𝜀= (1)
2

So that 𝜀 > 0 because 𝜌 > 0 by hypothesis. From the

𝑎𝑘
assumption that𝜌 = lim .It follows that for k
𝑘→+∞ 𝑏𝑘
𝑎𝑘
sufficiently large; say 𝑘 ≥ 𝐾 the ratio will be
𝑏𝑘
𝑎𝑘
within𝜀units of 𝜌. Thus 𝜌 − 𝜀 < < 𝜌 + 𝜀 when 𝑘≥
𝑏𝑘

𝐾.
1 𝑎𝑘 3
Or on substituting (1) 𝜌 < < 𝜌 when 𝑘 ≥ 𝐾.
2 𝑏𝑘 2
1 3
Or 𝜌 𝑏𝑘 < 𝑎𝑘 < 𝜌 𝑏𝑘 when 𝑘 ≥ 𝐾 (2)
2 2

If∑∞
𝑘=1 𝑎𝑘 converges, then∑∞
𝑘=𝐾 𝑎𝑘 converges by
Theorem3 (c) in Section 2. It follows from the
comparison test and left-hand inequality in
1
(2),that∑∞ ∞
𝑘=𝐾 𝜌 𝑏𝑘 converges. Thus∑𝑘=𝐾 𝑏𝑘 converges
2

(Theorem3 (b) in Section 2), and ∑ 𝑏𝑘 converges

(Theorem3 (c) in Section 2).

- 283 -
 3
Conversely, if∑∞ ∞
𝑘=1 𝑏𝑘 converges, ∑𝑘=𝐾 𝜌 𝑏𝑘 converges
2

(Theorem3 (b) and (c) in Section 2) so

that∑∞
𝑘=𝐾 𝑎𝑘 converges by the comparison test and the

right-hand inequality in (2). Thus∑∞

𝑘=1 𝑎𝑘 converges.◄

Example 7.
Use the limit comparison test to determine if
3𝑘 3 −2𝑘 2 +4
∑∞
𝑘=1 converges or diverges.
𝑘 5 −𝑘 3 +2

Solution.
From Principle2, the series behave like
3𝑘 3 3
∑∞
𝑘=1 = ∑∞
𝑘=1 (3)
𝑘5 𝑘2

which converges since it is a constant time a convergent

p-series. Thus, the given series is likely to converge. To
substantiate this, we apply the limit comparison test to
series (3) and the given series. We obtain
3 𝑘 3 −2 𝑘 2 +4
𝑘 5 −𝑘 3 +2 3 𝑘 5 − 2𝑘 4 + 4𝑘 2
𝜌 = lim 3 = lim =1
𝑘→+∞ 𝑘→+∞ 3 𝑘 5 − 3𝑘 3 + 6
𝑘2

Since 𝜌 > 0, the given series converges because series

(3) converges.◄

- 284 -
 Exercise Set (7.4)
In Exercise (1)-(6) prove that the series converges by the
comparison test.
1 2 1
(1) ∑∞
𝑘=1 (2) ∑∞
𝑘=1 (3) ∑∞
𝑘=1 .
3𝑘 +5 𝑘 4 +𝑘 5𝑘 2 −𝑘
𝑘 2𝑘 −1 5 sin2 𝑘
(4)∑∞
𝑘=1 (5)∑∞
𝑘=1 (6)∑∞
𝑘=1 .
8 𝑘 2 +𝑘 2 −1 3𝑘 +2 𝑘 𝑘!

In Exercises (7)-(12), prove that the series diverges by the

comparison test
3 1 9
(7)∑∞
𝑘=1 1 ;(8) ∑∞
𝑘=1 ; (9) ∑∞
𝑘=1 ;
𝑘−4 √𝑘+8 √𝑘+1
𝑘+1 𝑘 4/3 𝑘 −1/2
(10)∑∞
𝑘=1 𝑘 2 −𝑘;
∞
(11)∑𝑘=1 2 ∞
;(12)∑𝑘=1 ;
8 𝑘 +5 𝑘+1 2+sin2 𝑘
In exercises (13)-(18), use the limit comparison test to
determine if the series converges or diverges.
4𝑘 2 −2 𝑘+6 1
(13) ∑∞
𝑘=1 (14) ∑∞
𝑘=1
8 𝑘 7 +𝑘−8 9 𝑘+6
5 𝑘 (𝑘+3)
(15) ∑∞
𝑘=1 (16) ∑∞
𝑘=1 (𝑘+1)(𝑘+2)(𝑘+5)
3𝑘 +1
1 1
(17) ∑∞
𝑘=1 3 (18) ∑∞
𝑘=1 (2 𝑘+3)17
√8 𝑘 2 −3𝑘

In Exercise (19)-(33), use any method to determine if the

series converges or diverges. In some cases you may have
to use tests from earlier sections.
1 1
(19) ∑∞
𝑘=1 (20) ∑∞
𝑘=1 (3+𝑘)2/5
𝑘 3 +2 𝑘+1
1 In 𝑘
(21) ∑∞
𝑘=1 (22) ∑∞
𝑘=1
9 𝑘−2 𝑘

- 285 -
 √𝑘 4
(23) ∑∞
𝑘=1 (24) ∑∞
𝑘=1
𝑘 2 +1 2+𝑘3𝑘
1 2+(−1)𝑘
(25) ∑∞
𝑘=1 (26) ∑∞
𝑘=1
√𝑘(𝑘+1) 5𝑘

2+√𝑘 4+|cos 𝑘|
(27) ∑∞
𝑘=1 (𝑘+1)3 (28) ∑∞
𝑘=1
−1 𝑘3

1 √𝑘In 𝑘
(29) ∑∞
𝑘=1 (30) ∑∞
𝑘=1
4+2𝑘 𝑘 3 +1

∞ tan−1 𝑘 ∞ 5𝑘 +𝑘
(31) ∑𝑘=1 2 (32) ∑𝑘=1
𝑘 𝑘 !+3
In 𝑘
(33) ∑∞
𝑘=1 𝑘 √𝑘

(34) Use the limit comparison test to show

𝜋
that∑∞
k=1 sin (𝑘 ) diverges.

𝜋
[Hint : Compare with the series∑∞
𝑘=1 ]] 𝑘

(35) Use the comparison test to determine

In 𝑘
if∑∞
𝑘=1 converges or diverges.
𝑘2
1
(36) Determine if∑∞
𝑘=1 (In converges or diverges.
𝑘)2

(37) Let a, b and p be positive constants. For which

1
values of p does the series∑∞
𝑘=1 (𝑎+𝑏 converge?
𝑘)𝑝

- 286 -
(38) Show that𝑘 𝑘 ≥ 𝑘! and use this result to prove that
the series ∑∞
𝑘=1 𝑘
−𝑘
converges by the comparison
test.Prove convergence using the root test.
(39) Use the limit comparison test to investigation
(𝑘+1)2
convergence of∑∞
𝑘=1 (𝑘+2)! .

(40) Use the limit comparison test to investigate

1 1 1
convergence of the series 1 + + + + ⋯
3 5 7
1
(41) Prove that∑∞
𝑘=1 converges by comparison with a
𝑘!

suitable geometric series.

(42) Let∑ 𝑎𝑘 and ∑ 𝑏𝑘 be series with positive terms Prove:
𝑎
(a) If lim ( 𝑘 ) = 0 and ∑ bk converges, then
k→+∞ 𝑏 𝑘

∑ 𝑎𝑘 converges.
𝑎
(b) If lim ( 𝑘 ) + ∞ and ∑ bk diverges, then
k→+∞ 𝑏 𝑘

∑ 𝑎𝑘 diverges.

- 287 -
5. ALTERNATING SERIES; CONDITIONAL
CONVERGENCE
In this section we discuss series containing negative
terms.
Of special importance are series whose terms are
alternately positive and negative. These are called
alternating series. Such series have one of two possible
forms :
𝑎1 − 𝑎2 + 𝑎3 − 𝑎4 + ⋯ + (−1)𝑘+1 𝑎𝑘 + ⋯
or
−𝑎1 + 𝑎2 − 𝑎3 + 𝑎4 − ⋯ + (−1)𝑘 𝑎𝑘 + ⋯
where the 𝑎𝑘 ′𝑠 are all positive.
The following theorem is the key result on convergence
of alternating series.
Theorem 1 (Alternating Series Test).
An alternating series
𝑎1 − 𝑎2 + 𝑎3 − 𝑎4 + ⋯ + (−1)𝑘+1 𝑎𝑘 + ⋯
or
−𝑎1 + 𝑎2 − 𝑎3 + 𝑎4 − ⋯ + (−1)𝑘 𝑎𝑘 + ⋯
converges if the following two conditions are satisfied :
(𝑎)𝑎1 ≥ 𝑎2 ≥ 𝑎2 ≥ ⋯ ≥ 𝑎𝑘 ≥ ⋯

- 288 -
(𝑏) lim 𝑎𝑘 = 0
𝑘→+∞

♠To prove this result, we will need the following fact

about sequences.
If the even-numbered terms of a sequence tend toward a
limit L, and if the odd-numbered terms of the sequence
tend toward the same limit L, then the entire sequence
tend toward the limit L.
This result should be intuitively obvious; we omit the
proof.
Proof Theorem1.
We will consider the case
𝑎1 − 𝑎2 + 𝑎3 − 𝑎4 + ⋯ + (−1)𝑘+1 𝑎𝑘 + ⋯
The other case is left as an exercise. Consider the partial
sums
𝑠2 = 𝑎1 − 𝑎2
𝑠4 = (𝑎1 − 𝑎2 ) + (𝑎3 − 𝑎4 )
𝑠6 = (𝑎1 − 𝑎2 ) + (𝑎3 − 𝑎4 ) + (𝑎5 − 𝑎6 )
𝑠8 = (𝑎1 − 𝑎2 ) + (𝑎3 − 𝑎4 ) + (𝑎5 − 𝑎6 ) + (𝑎7 − 𝑎8 )
⋮
By hypothesis of (a), each of the difference appearing in
the parentheses is nonnegative so that
- 289 -
 𝑠2 ≤ 𝑠4 ≤ 𝑠6 ≤ 𝑠8 ≤ ⋯
Moreover, the terms in this sequence are all less than or
equal to 𝑎1 since we can write
𝑠2 = 𝑎1 − 𝑎2
𝑠4 = 𝑎1 − (𝑎2 − 𝑎3 ) − 𝑎4
𝑠6 = 𝑎1 − (𝑎2 − 𝑎3 ) + (𝑎4 − 𝑎5 ) − 𝑎6
𝑠8 = 𝑎1− (𝑎2 − 𝑎3 ) + (𝑎4 − 𝑎5 ) + (𝑎6 − 𝑎7 ) − 𝑎8
⋮
Thus the sequence
𝑠2 , 𝑠4 , 𝑠8 , … , 𝑠2 𝑛 , …
converges to some limit S, by Theorem 2 in Section 2 of
Chapter (VI). That is,
lim 𝑠2 𝑛 = 𝑆 (1)
𝑛→+∞

Now consider the sequence

𝑠1 , 𝑠3 , 𝑠5 , … , 𝑠2 𝑛−1 , …
We can write
𝑠2𝑛−1 = 𝑠2𝑛 − 𝑎2𝑛 (2)
But2𝑛 → +∞ 𝑎𝑠 𝑛 → +∞ . So lim 𝑎2𝑛 = 0by
𝑛→+∞

hypothesis (b). Thus, from (1) and (2)

lim 𝑠2𝑛−1 = lim 𝑠2𝑛 − lim 𝑎2𝑛 = 𝑆 − 0 = 𝑆 (3)
𝑛→+∞ 𝑛→+∞ 𝑛→+∞

- 290 -
Form (1) and (3), the sequence 𝑠1 , 𝑠2 , 𝑠3 , … , 𝑠𝑛 , …
converges to S, so the series converges.◄
Example 1.
The series
1 1 1 1
1− + − + ⋯ + (−1)𝑘+1 + ⋯
2 3 4 𝑘
is called the alternating harmonic series. Since
1 1
𝑎𝑘 = > = 𝑎𝑘+1
𝑘 𝑘+1
and
1
lim 𝑎𝑘 = lim =0
𝑘→+∞ 𝑘→+∞ 𝑘

this series converges by the alternating series test.◄

Example 2.
Determine whether the alternating series
∞
𝑘+3
∑ (−1)𝑘+1
𝑘(𝑘 + 1)
𝑘=1

converges or diverges.
Solution.
Requirement (b) of the alternating series test is satisfied
since

- 291 -
 1 3
𝑘+3 + 2
𝑘 𝑘
lim 𝑎𝑘 = lim = lim =0
𝑘→+∞ 𝑘→+∞ 𝑘(𝑘 + 1) 𝑘→+∞ 1 + 1
𝑘

To see if requirement (a) is met, we must determine if the

sequence
𝑘 + 3 +∞
{𝑎𝑘 }+∞
𝑘=1 ={ }
𝑘 (𝑘 + 1) 𝑘=1
is nonincreasing. Since
𝑎𝑘+1 𝑘+4 𝑘(𝑘 + 1) 𝑘 2 + 4𝑘
= . = 2
𝑎𝑘 (𝑘 + 1)(𝑘 + 2) 𝑘 + 3 𝑘 +5𝑘+6
𝑘 2 +4 𝑘 𝑘 2 +4 𝑘
= (𝑘 2 = < 1.
+4 𝑘)+(𝑘+6) 𝑘 2 +5 𝑘+6

We have 𝑎𝑘 > 𝑎𝑘+1 the series converges by the

alternating series test.◄
Theorem 2.
If an alternating series satisfies the hypotheses of the
alternating series test, and if the sum S of the series is
approximated by the nth partial sum 𝑠𝑛 , then the absolute
value of the error is less than 𝑎𝑛+1 .
Example 3.
As shown in Example 1, the alternating harmonic series
1 1 1 1
1− + − + ⋯ + (−1)𝑘+1 + ⋯
2 3 4 𝑘
- 292 -
satisfies the hypotheses of the alternating series test. If we
approximate the sum of the series by
1 1 1 7
1− + − =
2 3 4 12
1
then the absolute value of the error is less than = 0.2 ,
5

and if we approximate the sum by

1 1 1 1 1 1 319
1− + − + − + = , then the absolute value
2 3 4 5 6 7 420
1
of the error is less than = 0,125.◄
8

Remark 1.
1 1 1 1 1 1
The series 1 − − 2 + 3 + 4 − 5 − + ⋯does not fit
2 2 2 2 2 26

in any categories studied so far-it has mixed signs, but is

not alternating. We now develop some convergence tests
that can be applied to such series.
Definition 3.
A series∑∞
𝑘=1 𝑢𝑘 = 𝑢1 + 𝑢2 + ⋯ + 𝑢𝑘 + ⋯is said to
converge absolutely if the series of absolute values
∞

∑ |𝑢𝑘 | = |𝑢1 | + |𝑢2 | + ⋯ + |𝑢𝑘 | + ⋯

𝑘=1

converges.

- 293 -
Example 4.
The series
1 1 1 1 1 1
1− − 2+ 3+ 4− 5− 6+⋯
2 2 2 2 2 2
converges absolutely since the series of absolute values
1 1 1 1 1 1
1+ + 2+ 3+ 4− 5− 6+⋯
2 2 2 2 2 2
is a convergent geometric series. On the other hand, the
alternating harmonic series
1 1 1 1
1− + − + −⋯
2 3 4 5
does not converge absolutely since the series of absolute
values
1 1 1 1
1+ + + + +⋯
2 3 4 5
diverges.◄
Theorem 4.
If the series
∞

∑ |𝑢𝑘 | = |𝑢1 | + |𝑢2 | + ⋯ + |𝑢𝑘 | + ⋯

𝑘=1

converges, then so does the series

- 294 -
 ∞

∑ 𝑢𝑘 = 𝑢1 + 𝑢2 + ⋯ + 𝑢𝑘 + ⋯
𝑘=1

In other words, if a series converges absolutely, then it

converges.
Proof.
Our proof is based on a trick. We will show that the series
∞

∑ (𝑢𝑘 + |𝑢𝑘 |) (2)

𝑘=1

converges. Since∑|𝑢𝑘 | is assumed to converge, it will

then follow from Theorem 3 (a) in Section 2 in this
Chapter that∑ 𝑢𝑘 converges, since
∞ ∞

∑ 𝑢𝑘 = ∑ (𝑢𝑘 + |𝑢𝑘 |) − |𝑢𝑘 |]

𝑘=1 𝑘=1

For all k, the value of 𝑢𝑘 + |𝑢𝑘 | is either 0 or 2|𝑢𝑘 |,

depending on whether𝑢𝑘 is negative or not. Thus for all
values of k,
0 ≤ 𝑢𝑘 + |𝑢𝑘 | ≤ 2|𝑢𝑘 | (3)
But∑ 2|𝑢𝑘 |is a convergent series since it is a constant
times the convergent series|𝑢𝑘 |. Thus from (3), series (2)
converges by the comparison test.◄

- 295 -
Example 5.
In Example (4) we showed that
1 1 1 1 1 1
1− − 2+ 3+ 4− 5− 6+⋯
2 2 2 2 2 2
converges absolutely. It follows from Theorem 4 above
that the series converges.◄
Example 6.
cos 𝑘
Show that the series ∑∞
𝑘=1 converges.
𝑘2

Solution.
cos 𝑘 1 cos 𝑘
Since |cos 𝑘| ≤ 1for all k, | |≤ 2 . Thus
∑∞
𝑘=1 | |
𝑘2 𝑘 𝑘2

converges by the comparison test, and consequently

cos 𝑘
∑∞
𝑘=1 converges.◄
𝑘2

♠If∑|𝑢𝑘 | diverges, no conclusion can be drawn about the

convergence or divergence of∑ 𝑢𝑘 . For example, consider
the two series
1 1 1 1
1− + − + ⋯ + (−1)𝑘+1 + ⋯ (4)
2 3 4 𝑘
1 1 1 1
−1 − − − − ⋯ − − ⋯ (5)
2 3 4 𝑘
Series (4), the alternating harmonic series, converges;
while series (5), being a constant time the harmonic

- 296 -
series, diverges. Yet in each case the series of absolute
values is
1 1 1
1+ + + ⋯+ + ⋯
2 3 𝑘
which diverges. A series such as (4), which is convergent,
but not absolutely convergent, is called conditionally
convergent.
The following version of the ratio test is useful for
investigating absolute convergence.
Theorem 5 (The Ratio Test for Absolute
Convergence).
Let∑ 𝑢𝑘 be a series with nonzero terms and suppose
|𝑢𝑘+1 |
lim =𝜌
𝑘→+∞ |𝑢𝑘 |

(a) If 𝜌 < 1, the series∑ 𝑢𝑘 converges absolutely.

(b) If 𝜌 > 1 or if 𝜌 = +∞ , then the series∑ 𝑢𝑘 diverges.
(c) If 𝜌 = 1, no conclusion about converges can be
drawn.
Example 7.
2𝑘
The series ∑∞
𝑘=1(−1)
𝑘
converges absolutely since
𝑘!

- 297 -
 |𝑢𝑘+1 |
𝜌 = lim
𝑘→+∞ |𝑢𝑘 |

2𝑘+1 𝑘! 2
= lim . = lim = 0 < 1.◄
𝑘→+∞ (𝑘+1)! 2𝑘 𝑘→+∞ 𝑘+1

Example 8.
We proved earlier (Theorem 3 in Section 1 in this
Chapter) that a geometric series
𝑎 + 𝑎𝑟 + 𝑎𝑟 2 + ⋯ + 𝑎𝑟 𝑘−1 + ⋯
converges if|𝑟| < 1 and diverges if |𝑟| ≥ 1. However, a
stronger statement can be made - the series converges
absolutely if|𝑟| < 1. This follows from Theorem 5 in
Section 4 above since
|𝑢𝑘+1 | |𝑎𝑟 𝑘 |
𝜌 = lim = lim = lim |𝑟| = |𝑟|
𝑘→+∞ |𝑢𝑘 | 𝑘→+∞ |𝑎𝑟 𝑘−1 | 𝑘→+∞

so that 𝜌 < 1 if |𝑟| < 1.

The following review is included as a ready
reference to convergence tests.

- 298 -
 Review of Convergence Tests
NAME STATEMENT COMMENTS
Divergence test If lim 𝑢𝑘 ≠ 0 then∑ 𝑢𝑘 diverges If lim 𝑢𝑘 =
𝑘→+∞ 𝑘→+∞

0 , ∑ 𝑢𝑘 may or may
not converge.
Integral test Let∑ 𝑢𝑘 be a series with positive Use this test when
terms and let 𝑓(𝑥) be the function 𝑓(𝑥) is easy to
that results when k is replaced by x integrate.
in the formula for 𝑢𝑘 . If f is
decreasing and continuous for 𝑥 ≥
+∞
1, then ∑∞
𝑘−1 𝑢𝑘 𝑎𝑛𝑑 ∫1 𝑓(𝑥)𝑑𝑥
both converges or both diverges.
Comparison test Let∑ 𝑎𝑘 and∑ 𝑏𝑘 be series with Use this test as a last
positive terms such that reason easier to apply.
𝑎1 ≤ 𝑏1 , 𝑎2 ≤ 𝑏2 , … , 𝑎𝑘 ≤ 𝑏𝑘 , …
If
∑ 𝑏𝑘 converges, then∑ 𝑎𝑘
converges, and if∑ 𝑎𝑘 diverges,
then∑ 𝑏𝑘 diverges
Ratio test let∑ 𝑢𝑘 be a series with positive Try this test when 𝑢𝑘
𝑢𝑘+1
terms and let lim =𝜌 involves factorials or
𝑘→+∞ 𝑢𝑘
kth powers.
(a) Series converges if 𝜌 < 1.
(b) Series diverges if 𝜌 > 1 or
𝜌 = +∞
(c) No conclusion if 𝜌 = 1

- 299 -
 NAME STATEMENT COMMENTS
Root test Let be∑ 𝑢𝑘 a series with positive Try this test when 𝑢𝑘
terms such that 𝜌 = lim 𝑘√𝑢𝑘 involves kth power.
𝑘→+∞

(a) Series converges if 𝜌 < 1.

(b) Series diverges if 𝜌 > 1.
or.𝜌 = +∞
(c) No conclusion if 𝜌 = 1.
Limit let∑ 𝑎𝑘 and∑ 𝑏𝑘 be series with This is easier to apply
comparison test positive terms such that than the comparison
𝑎𝑘 test, but still requires
𝜌 = lim
𝑘→+∞ 𝑏𝑘
some still in choosing
if 0 < 𝜌 < +∞, then both series
the series∑ 𝑏𝑘 for
converges, or both diverges.
comparison.
Alternating The series 𝑎1 − 𝑎2 + 𝑎3 − 𝑎4 + ⋯ This test applies only
series test and −𝑎1 + 𝑎2 − 𝑎2 + 𝑎4 − ⋯ no alternating series.
converges if 𝑎1 ≥ 𝑎2 ≥ 𝑎3 ≥ ⋯
lim 𝑎𝑘 = 0
𝑘→+∞

Ratio test for Let∑ 𝑢𝑘 be a series with nonzero The series need not
absolute |𝑢𝑘+1 | have positive terms
terms such that𝜌 = lim
𝑘→+∞ |𝑢𝑘 |
convergence and need not be
(a) Series converges absolutely
alternating to use this
if 𝜌 < 1
test
(b) Series diverges if 𝜌 > 1
or𝜌 = +∞
(c ) No conclusion if 𝜌 = 1.

- 300 -
 Exercise Set (7.5)
In Exercise (1)-(6), use the alternating series test to
determine if the series converges or diverges
(−1)𝑘+1 𝑘
(1) ∑∞
𝑘=1 2 𝑘+1 (2) ∑∞
𝑘=1(−1)
𝑘+1
3𝑘
𝑘+1 𝑘+4
(3) ∑∞
𝑘=1(−1)
𝑘+1
(4) ∑∞
𝑘=1(−1)
𝑘+1
3 𝑘+1 𝑘 2 +𝑘
𝐼𝑛 𝑘
(5) ∑∞
𝑘=1(−1)
𝑘+1 −𝑘
𝑒 (6) ∑∞
𝑘=3(−1)
𝑘
𝑘

In Exercise (7)-(12), use ratio test for absolute

convergence to determine if the series converges
absolutely or diverges.
3 𝑘 2𝑘
(7) ∑∞
𝑘=1 (− 5) (8) ∑∞
𝑘=1(−1)
𝑘+1
𝑘!

3𝑘 𝑘
(9) ∑∞
𝑘=1(−1)
𝑘+1
(10) ∑∞ 𝑘
𝑘=1(−1) (5𝑘 )
𝑘2

∞ 𝑘3 𝑘+1 𝑘
𝑘
(11) 𝑘
∑𝑘=1(−1) ( 𝑘 ) (12) ∑∞
𝑘=1( −1 )
𝑒 𝑘!

In Exercise (13)-(30), classify the series as:absolutely

convergent, conditionally convergent, or divergent.
(−1)𝑘+1 (−1) 𝑘+1
(13) ∑∞
𝑘=1 3 𝑘 (14) ∑∞
𝑘=1 𝑘 4/3

(−4)𝑘 (−1) 𝑘+1

(15) ∑∞
𝑘=1 𝑘 2 (16) ∑∞
𝑘=1 𝑘!
cos 𝑘𝜋 𝑘
(−1) 𝐼𝑛 𝑘
(17) ∑∞
𝑘=1 𝑘 (18) ∑∞
𝑘=3 𝑘

- 301 -
 𝑘+2 (−1)𝑘+1
(19) ∑∞
𝑘=1(−1)
𝑘+1
( ) (20) ∑∞
𝑘=1
3 𝑘−1 𝑘 2 +1

𝑘+2 (−1)𝑘+1 𝑘 2
(21) ∑∞
𝑘=1(−1)
𝑘+1
(22) ∑∞
𝑘=1
𝑘 (𝑘+3) 𝑘 3 +1
𝑘𝜋 sin 𝑘
(23) ∑∞
𝑘=1 sin (24) ∑∞
𝑘=1
𝑘3 𝑘3
(−1)𝑘 (−1)𝑘
(25) ∑∞
𝑘=2 (26) ∑∞
𝑘=1
𝑘 𝐼𝑛 𝑘 √𝑘 (𝑘+1)

1 (−1)𝑘+1
(27) ∑∞
𝑘=2 (− ) (28) ∑∞
𝑘=1
𝐼𝑛 𝑘 √𝑘+1+√𝑘
(−1)𝑘 (𝑘 2 +1) 𝑘 cos 𝑘𝜋
(29) ∑∞
𝑘=2 (30) ∑∞
𝑘=1
𝑘 3 +2 𝑘 2 +1

In Exercise 31-34, the given series satisfies the

hypotheses of the alternating series test. For the stated
value of n, estimate the error that results if the sum of the
series is approximately by the nth partial sum.
(−1)𝑘+1
(31) ∑∞
𝑘=1 ; 𝑛 = 7.
𝑘
(−1)𝑘+1
(32) ∑∞
𝑘=1 ; 𝑛 =5.
𝑘!
(−1)𝑘+1
(33) ∑∞
𝑘=1 ; 𝑛 = 99 .
√𝑘
(−1)𝑘+1
(34) ∑∞
𝑘=1 (𝑘+1) ; 𝑛 =3.
𝐼𝑛 (𝑘+1)

In Exercise (35)-(38), the given series satisfies the

hypotheses of the alternating series test. Find the smallest

- 302 -
value of n for which the nth partial sum approximates the
sum of the series to the stated accuracy.
(−1)𝑘+1
(35) ∑∞
𝑘=1 ; |𝑒𝑟𝑟𝑜𝑟| < 0.0001 .
𝑘
(−1)𝑘+1
(36) ∑∞
𝑘=1 ; |𝑒𝑟𝑟𝑜𝑟| < 0.00001
𝑘!
(−1)𝑘+1
(37) ∑∞
𝑘=1 ; |𝑒𝑟𝑟𝑜𝑟| < 0.005 .
√𝑘
(−1)𝑘+1
(38) ∑∞
𝑘=1 (𝑘+1) ; |𝑒𝑟𝑟𝑜𝑟| < 0.1.
𝐼𝑛 (𝑘+1)

(39) Prove: If∑ 𝑎𝑘 converges absolutely, then∑ 𝑎𝑘2

converges.

- 303 -
6. POWER SERIES
If 𝑐0 , 𝑐1 , 𝑐2 , … are constants and x is a variable, then a
series of the form
∞

∑ 𝑐𝑘 𝑥 𝑘 = 𝑐0 + 𝑐1 𝑥 + 𝑐2 𝑥 2 + ⋯ + 𝑐𝑘 𝑥 𝑘 + ⋯
𝑘=0

is called a power series in x. some examples are

∞

∑ 𝑥𝑘 = 1 + 𝑥 + 𝑥2 + 𝑥3 + ⋯
𝑘=0
∞
𝑥𝑘 𝑥2 𝑥3
∑ =1+𝑥+ + +⋯
𝑘! 2! 3!
𝑘=0
∞
𝑥 𝑘+1 𝑥2 𝑥3 𝑥4
∑ (−1 )𝑘 =𝑥− + −
𝑘+1 2 3 4
𝑘=0
∞
𝑥2 𝑘 𝑥2 𝑥4 𝑥6
∑ (−1 )𝑘 =1− + − +⋯
(2 𝑘) ! 2! 4! 6!
𝑘=0
∞
𝑥 2 𝑘+1 𝑥3 𝑥5 𝑥7
∑ (−1 )𝑘 =𝑥− + − +⋯
(2 𝑘 + 1) ! 3! 5! 7!
𝑘=0

A fundamental problem. For what values of x does a

given power series,∑ 𝑐𝑘 𝑥 𝑘 , converge?
Theorem 1.
For any power series in x, exactly of the following is true:
- 304 -
(a) The series converges only for 𝑥 = 0.
(b) The series converges absolutely for all x.
(c)The series converges absolutely for all x in some finite
open interval (−𝑅, 𝑅), and diverges if 𝑥 < −𝑅 𝑜𝑟 𝑥 >
𝑅. At 𝑥 = 𝑅 and 𝑥 = −𝑅 the series may converge
absolutely, converge conditionally, or diverge, depending
on the particular series.

In case (c), where the power series converges absolutely

for |𝑥 | < 𝑅 and diverges for|𝑥 | > 𝑅, we call R the radius
of convergence. In case (a), where the series converges
only for 𝑥 = 0, we define the radius of convergence to
be 𝑅 = 0; and in case (b), where the series converges
absolutely for all x, we define the radius of convergence
to be 𝑅 = −∞.
The set of all values of x for which a power series
converges is called the interval of convergence.
Example 1.
Find the interval of convergence and radius of
convergence of the power series
- 305 -
 ∞

∑ 𝑥𝑘 = 1 + 𝑥 + 𝑥2 + ⋯ + 𝑥𝑘 + ⋯
𝑘=0

Solution.
For every x, the given series is a geometric series with
ratio 𝑟 = 𝑥.By Example 8 of Section 5 in this chapter,
the series converges absolutely if −1 < 𝑥 < 1 and
diverges if|𝑥 | ≥ 1. So, the interval of convergence is
(−1, 1) and the radius of convergence is 𝑅 = 1.
Example 2.
Find the interval of convergence and radius of
𝑥𝑘
convergence of ∑∞
𝑘=0 .
𝑘!

Solution.
We will apply the ratio test for absolute convergence. for
every real number x,
𝑢𝑘+1 𝑥 𝑘+1 𝑘!
𝜌 = lim | | = lim | . 𝑘|
𝑘→+∞ 𝑢𝑘 𝑘→+∞ (𝑘 + 1) ! 𝑥

𝑥
= lim | |=0
𝑘→+∞ 𝑘+1
Since 𝜌 < 1 for all x, the series converges absolutely for
all x. Thus the interval of convergence is(−∞ , +∞) and
the radius of convergence is 𝑅 = +∞ .◄
- 306 -
REMARK.
There is a useful byproduct of Example 2. Since
∞
𝑥𝑘
∑
𝑘!
𝑘=0

converges for all x, Theorem 4.1 implies that for all

values of x
𝑥𝑘
lim (1)
𝑘→+∞ 𝑘 !

We will need this result later.

Example 3.
Find the interval and radius of converges of
∞

∑ 𝑘 ! 𝑥𝑘
𝑘=1

Solution.
If 𝑥 = 0, the series has only one nonzero term and
therefore converges. If 𝑥 ≠ 0, the ratio test yields
𝑢𝑘+1 (𝑘 + 1) ! 𝑥 𝑘+1
𝜌 = lim | | = lim | |
𝑘→+∞ 𝑢𝑘 𝑘→+∞ 𝑘 ! 𝑥𝑘
= lim |(𝑘 + 1) 𝑥 | = +∞
𝑘→+∞

Therefore, the series converges if 𝑥 = 0, but diverges

for all other x. consequently, the interval of convergence
- 307 -
is the single point 𝑥 = 0 and the radius of convergence
is 𝑅 = 0.◄
Example 4.
Find the interval of convergence and radius of
convergence of
∞
(−1)𝑘 𝑥 𝑘
∑ 𝑘
3 (𝑘 + 1)
𝑘=0

Solution.
Since|(−1)𝑘 | = |(−1)𝑘+1 | = 1 we obtain
𝑢𝑘+1 𝑥 +1 3𝑘 (𝑘 + 1)
𝜌 = lim | | = lim | 𝑘+1 . |
𝑘→+∞ 𝑢𝑘 𝑘→+∞ 3 (𝑘 + 2) 𝑥𝑘
|𝑥 | 𝑘 + 1
= lim [ .( )]
𝑘→+∞ 3 𝑘+2
|𝑥| 1+1/𝑘 |𝑥|
= lim ( )= .
3 𝑘→+∞ 1+2/𝑘 3

The ratio test for absolute convergence implies that the

series converges absolutely if |𝑥 | < 3and diverges if
|𝑥 | > 3. The ratio test fails when |𝑥 | = 3.
So the cases 𝑥 = −3 and 𝑥 = 3 need separate analyses.
Substituting 𝑥 = −3 in the given series yields

- 308 -
 ∞ ∞ ∞
(−1)𝑘 (−3)𝑘 (−1)𝑘 (−1)𝑘 3𝑘 1
∑ 𝑘 =∑ = ∑
3 (𝑘 + 1) 3𝑘 (𝑘 + 1) 𝑘+1
𝑘=0 𝑘=0 𝑘=0

which is the divergent harmonic series

1 1 1
1 + + + + ⋯.
2 3 4

Substituting 𝑥 = 3 in the given series yields

∞ ∞
(−1)𝑘 3𝑘 (−1)𝑘
∑ 𝑘 =∑
3 (𝑘 + 1) 𝑘+1
𝑘=0 𝑘=0

which is the conditionally convergent alternating

1 1 1
harmonic series .1 − + + − ⋯Thus the interval of
2 3 4

convergence for the given series is (-3, 3] and the radius

of convergence is 𝑅 = 3.◄
Example 5.
Find the interval and radius of convergence of the series
∞
𝑥 2𝑘
∑ (−1 )𝑘
(2 𝑘)!
𝑘=0

Solution.
Since|(−1)𝑘 | = |(−1)𝑘+1 | = 1, we have
𝑢𝑘+1 𝑥 2(𝑘+1) (2 𝑘) !
𝜌 = lim | | = lim | . 2𝑘 |
𝑘→+∞ 𝑢𝑘 𝑘→+∞ [2 (𝑘 + 1)] ! 𝑥

- 309 -
 𝑥 2 𝑘+2 (2 𝑘) !
= lim | |
𝑘→+∞ (2 𝑘 + 2) ! 𝑥 2 𝑘

𝑥2
= lim | |
𝑘→+∞ (2 𝑘 + 2)(2 𝑘 + 1)

1
= 𝑥 2 lim
𝑘→+∞ (2 𝑘 + 2)(2 𝑘 + 1)

= 𝑥 2 .0 = 0
Thus 𝜌 < 1 for all x, which means that the interval
of convergence is(−∞ , +∞) and the radius of
convergence is 𝑅 = +∞.◄
Theorem 2.
For a power series∑ 𝑐𝑘 (𝑥 − 𝑎)𝑘 , exactly one of the
following is true:
(a) The series converges only for 𝑥 = 𝑎.
(b) The series converges absolutely for all x.
(c) The series converges absolutely for all x in some
finite open interval (𝑎 − 𝑅, 𝑎 + 𝑅) and diverges if 𝑥 <
𝑎 − 𝑅 𝑜𝑟 𝑥 > 𝑎 + R. At the points 𝑥 = 𝑎 − 𝑅 and
𝑥 = 𝑎 + 𝑅, the series may converge absolutely,
converge conditionally, or diverge, depending on the
particular series.

- 310 -
In case (a), (b) and (c) of Theorem 8.2 the series is said to
have radius of convergence, 0, +∞, and R respectively.
The set of all values of x for which the series converges is
called the interval of convergence.
Example 6.
Find the interval of convergence and radius of
convergence of the series
∞
(𝑥 − 5)𝑘
∑
𝑘2
𝑘=1

Solution.
We apply the ratio test for absolute convergence.
𝑢𝑘+1
𝜌 = lim | |
𝑘→+∞ 𝑢𝑘

(𝑥 − 5)𝑘+1 𝑘2
= lim | . |
𝑘→+∞ (𝑘 + 1)2 (𝑥 − 5)𝑘

𝑘 2
= lim [|𝑥 − 5| ( ) ]
𝑘→+∞ 𝑘+1

- 311 -
 2
1
= |𝑥 − 5| lim ( 1)
𝑘→+∞ 1+
𝑘

= |𝑥 − 5|.
Thus, the series converges absolutely if|𝑥 − 5| < 1, or
−1 < 𝑥 − 5 < 1, or 4 < 𝑥 < 6. The series diverges
if
𝑥 < 4 or 𝑥 > 6.
To determine the convergence behavior at end
points 𝑥 = 4 and 𝑥 = 6, we substitute these values in
the given series, if 𝑥 = 6 the series becomes
1 𝑘 1 1 1 1
∑∞
𝑘=1 𝑘 2 = ∑∞
𝑘=1 =1+ + + +⋯
𝑘2 22 32 42

which is a convergent p-series (𝑝 = 2).

If 𝑥 = 4 the series becomes
(−1)𝑘 1 1 1
∑∞
𝑘=1 = −1 + − + −⋯
𝑘2 22 32 42

Since this series converges absolutely, the interval of

converges for the given series is [4, 6]. The radius of
convergence is 𝑅 = 1.

- 312 -
 Exercise Set (7.6)
In Exercise (1)-(24), find the radius of convergence and
the interval of convergence.
∞ 𝑥𝑘
(1) ∑𝑘=0 ; (2) ∑∞ 𝑘 𝑘
𝑘=0 3 𝑥 ;
𝑘+1
(−1) 𝑥 𝑘 𝑘 𝑘!
(3) ∑∞
𝑘=0 ; (4) ∑∞
𝑘=0 𝑥𝑘;
𝑘! 2𝑘
5𝑘 𝑥𝑘
(5) ∑∞
𝑘=1
𝑘
2𝑥 ; (6) ∑∞
𝑘=1 ;
𝑘 𝑙𝑛 𝑘
𝑥𝑘 (−2)𝑘 𝑥 𝑘+1
(7) ∑∞
𝑘=1 ; (8) ∑∞
𝑘=0 ;
𝑘 (𝑘+1) 𝑘+1

𝑥𝑘 (−1)𝑘 𝑥 2𝑘
(9) ∑∞
𝑘=1(−1)
𝑘−1
; (10) ∑∞
𝑘=0 ;
√𝑘 (2𝑘) !

𝑥 2𝑘+1 𝑥 3𝑘
(11) ∑∞ 𝑘
𝑘=0(−1) (2𝑘+1) ; (12) ∑∞
𝑘=0(−1)
𝑘
;
! 𝑘 3/2

3 𝑘 𝑥𝑘
(13) ∑∞ 𝑘
𝑘=0 𝑘 ! 𝑥 ; (14) ∞
∑𝑘=2(−1) 𝑘+1
;
𝑘 (𝑙𝑛 𝑘)2

𝑥𝑘 (𝑥−3)𝑘
(15) ∑∞
𝑘=0 2; (16) ∑∞
𝑘=0 ;
1+𝑘 2𝑘
(𝑥+1)𝑘 (𝑥−4)𝑘
(17) ∑∞
𝑘=1(−1)
𝑘+1
; (18)∑∞ 𝑘
𝑘=0(−1) (𝑘+1)2 ;
𝑘

3 𝑘 (2𝑘+1) !
(19) ∑∞ 𝑘 ∞
𝑘=0 (4) (𝑥 + 5) ; (20) ∑𝑘=1 (𝑥 − 2)𝑘 ;
𝑘3
(𝑥+1)2𝑘+1 (𝐼𝑛 𝑘)(𝑥−3)𝑘
(21) ∑∞
𝑘=1(−1)
𝑘
; (22) ∑∞
𝑘=1 ;
𝑘 2 +4 𝑘
𝜋𝑘 (𝑥−1)2 𝑘 (2 𝑥−3)𝑘
(23) ∑∞
𝑘=0 ; (24) ∑∞
𝑘=1 .
(2 𝑘+1) ! 4 2𝑘

- 313 -
(25) Use the root test to find the interval of
𝑥𝑘
convergence of ∑∞
𝑘=2 .
(𝑙𝑛 𝑘)𝑘

(25) Find the radius of convergence of

∞
1.2.3 … 𝑘
∑ (−1)𝑘 𝑥 2 𝑘+1
1.3.5 … (2𝑘 − 1)
𝑘=1
(𝑥−𝑎) 𝑘
(26) Find the interval of convergence of∑∞
𝑘=0 𝑏𝑘

where 𝑏 > 0.
(27) Find the radius of convergence of the power
(𝑝𝑘)!
series ∑∞
𝑘=0 (𝑘 𝑥 𝑘 , where p is a positive integer.
!)𝑝

(28) Find the radius of convergence of the power

(𝑘+𝑝)!
series ∑∞
𝑘=0 𝑥 𝑘 , where p and q are positive
𝑘!(𝑘+𝑞)!

integers.
(29) Prove: If lim |𝑐𝑘 |1/𝑘 = 𝐿 , where 𝐿 ≠ 0, then
𝑘→+∞

the power series ∑∞ 𝑘

𝑘=0 𝑐𝑘 𝑥 has radius of

convergence 1/𝐿.

- 314 -
7. TAYLOR AND MACLAURIN SERIES
In this section we will study the approximation of
function by polynomials and introduce an important class
of power series.
Suppose we are interested in approximating a
function f by a polynomial
𝑝(𝑥 ) = 𝑐0 + 𝑐1 𝑥 + ⋯ + 𝑐𝑛 𝑥 𝑛 (1)
over an interval centered at 𝑥 = 0. Because 𝑝(𝑥) has
𝑛 + 1 coefficients, it seems reasonable that we will be
able to impose 𝑛 + 1 conditions on this polynomial. We
will assume that the first n derivatives of f exist at 𝑥 =
0, and we will choose these 𝑛 + 1 conditions to be:
𝑓(0) = 𝑝(0), 𝑓 ′ (0) = 𝑝′ (0), 𝑓 ′′ (0) = 𝑝′′ (0), …,
𝑓 (𝑛) (0) = 𝑝(𝑛) (0) (2)
These conditions require that the value of 𝑝(𝑥) and its

first n derivatives match the value of 𝑓(𝑥) and its first n

derivatives at 𝑥 = 0. By forcing this high degree of

"match" at 𝑥 = 0, it is reasonable to hope that 𝑓(𝑥)and

- 315 -
𝑝(𝑥) will remain close over some interval (possibly

small) centered at 𝑥 = 0. Since

𝑝(𝑥 ) = 𝑐0 + 𝑐1 𝑥 + 𝑐2 𝑥 2 + 𝑐3 𝑥 3 + ⋯ + 𝑐𝑛 𝑥 𝑛
𝑝′ (𝑥 ) = 𝑐1 + 2𝑐2 𝑥 + 3𝑐3 𝑥 2 + ⋯ + 𝑛𝑐𝑛 𝑥 𝑛−1
𝑝′′ (𝑥 ) = 2𝑐2 + 3.2𝑐3 𝑥 + ⋯ + 𝑛(𝑛 − 1)𝑐𝑛 𝑥 𝑛−2
𝑝′′′ (𝑥 ) = 3.2𝑐3 + ⋯ + 𝑛(𝑛 − 1)(𝑛 − 2)𝑐𝑛 𝑥 𝑛−3
⋮
𝑝(𝑛) (𝑥 ) = 𝑛(𝑛 − 1)(𝑛 − 2) … (1)𝑐𝑛
we obtain on substituting 𝑥 = 0
𝜌(0) = 𝑐0
𝑝′ (0) = 𝑐1
𝑝′′ (0) = 2𝑐2 = 2! 𝑐2
𝑝′′′ (0) = 3 .2𝑐2 = 3! 𝑐3
⋮
𝑝(𝑛) (0) = 𝑛(𝑛 − 1)(𝑛 − 2) … 𝑐𝑛 = 𝑛! 𝑐𝑛
Thus, from (2),
𝑓(0) = 𝑐0
⋮
𝑓 ′ (0) = 𝑐1
𝑓 ′′ (0) = 2! 𝑐2

- 316 -
 𝑓 ′′′ (0) = 3! 𝑐3
⋮
𝑓 (𝑛) (0) = 𝑛 ! 𝑐𝑛
So

′(
𝑓 ′′ (0) 𝑓 ′′′ (0)
𝑐0 = 𝑓(0), 𝑐1 = 𝑓 0), 𝑐2 = , 𝑐3 = , … , 𝑐𝑛
2! 3!
𝑓 (𝑛) (0)
=
𝑛!
Substituting these values in (1) yields a polynomial,
called the nth Maclaurin polynomial for f.

Definition 1.
If f can be differentiated n times at 0, then we define the
nth Maclaurin polynomial for f to be
𝑓 ′ (0) 𝑓 ′′ (0) 2 𝑓 ′′′ (0) 3
𝑝𝑛 (𝑥 ) = 𝑓(0) + 𝑥+ 𝑥 + 𝑥 +⋯
1! 2! 3!
𝑓 (𝑛) (0) 𝑛
+ 𝑥
𝑛!
(3)
This polynomial has the property that its value and the
values of its first n derivatives match the value of 𝑓(𝑥)
and its first n derivatives when 𝑥 = 0.

- 317 -
Example 1.
Find the Maclaurin polynomials 𝑝0 , 𝑝1 , 𝑝2 , 𝑝3 and𝑝𝑛 for
𝑒𝑥.
Solution.
Let 𝑓(𝑥) = 𝑒 𝑥 .
Thus 𝑓 ′ (𝑥 ) = 𝑓 ′′ (𝑥 ) = ⋯ = 𝑓 (𝑛) (𝑥 ) = 𝑒 𝑥 .
and 𝑓(0) = 𝑓 ′ (0) = 𝑓 ′′ (0) = ⋯ = 𝑓 (𝑛) = 𝑒 0 = 1.
Therefore
𝑝0 (𝑥 ) = 𝑓 (0) = 1;
𝑝1 (𝑥 ) = 𝑓 (0) + 𝑓 ′ (0) = 1 + 𝑥;
𝑓 ′′ (0) 𝑥2
𝑝2 (𝑥 ) = 𝑓 (0) + 𝑓 ′ (0)𝑥 + 𝑥2 = 1 + 𝑥 + =1+
2! 2!
1
𝑥 + 𝑥 2;
2

′(
𝑓 ′′ (0) 2 𝑓 ′′′ (0) 3
𝑝3 (𝑥 ) = 𝑓 (0) + 𝑓 0) + 𝑥 + 𝑥
2! 3!
𝑥2 𝑥3 1 1
=1+𝑥+ + = 1 + 𝑥 + 𝑥2 + 𝑥3;
2! 3! 2 6

′(
𝑓 ′′ (0) 2 𝑓 (𝑛) (0) 𝑛
𝑝𝑛 (𝑥 ) = 𝑓(0) + 𝑓 0) + 𝑥 +⋯+ 𝑥
2! 𝑛!
𝑥2 𝑥𝑛
=1+𝑥+ +⋯+ ;
2! 𝑛!

- 318 -
 In the figure we have sketched the graphs of 𝑒 𝑥 and
its Maclaurin polynomials of degree 1, 2, and 3. Note that
the graphs of 𝑒 𝑥 and𝑝3 (𝑥) are virtually indistinguishable
over the interval from −0.5 𝑡𝑜 + 0.5.
Example 2.
Find the nth Maclaurin polynomial for 𝑙𝑛 (𝑥 + 1).
Solution.
Let 𝑓(𝑥) = 𝑙𝑛 (𝑥 + 1) and arrange the computations as
follows:
𝑓(𝑥 ) = ln (𝑥 + 1) ⟹ 𝑓(0) = ln 1 = 0;
1
𝑓 ′ (𝑥 ) = ⟹ 𝑓 ′ (0) = 1;
𝑥+1
1
𝑓 ′′ (𝑥 ) = − ⟹ 𝑓 ′′ (0) = −1;
(𝑥+1)2

- 319 -
 2
𝑓 ′′′ (𝑥 ) = (𝑥+1)3 ⟹ 𝑓 ′′′ (0) = 2;
3 .2
𝑓 4 (𝑥) = (𝑥+1)4 ⟹ 𝑓 (4) (0) = −3 !;
4 .3 .2
𝑓 (5) (𝑥) = (𝑥+1)5 ⟹ 𝑓 (5) (0) = 4 !;

⋮ ⋮
(𝑛 − 1) !
𝑓 (𝑛) = (−1)𝑛+1 ⟹
(𝑥 + 1)𝑛
𝑓 (𝑛) (0) = (−1)𝑛+1 (𝑛 − 1) !.
Substituting these values in (3) yields
𝑥2 𝑥3 𝑥𝑛
𝑝𝑛 (𝑥 ) = 𝑥 − + − ⋯ + (−1)𝑛+1 .◄
2 3 𝑛

Example 3.
In the Maclaurin polynomials for sin 𝑥, only the odd
powers of x appear explicitly. To see this, let 𝑓(𝑥) = sin
x, thus
𝑓 (𝑥 ) = sin 𝑥 ⟹ 𝑓(0) = 0
𝑓′(𝑥) = cos 𝑥 ⟹ 𝑓′(0) = 1
𝑓"(𝑥) = − sin 𝑥 ⟹ 𝑓"(0) = 0
𝑓 ′′′ (𝑥 ) = − cos 𝑥 ⟹ 𝑓 ′′′ (0) = −1
Since𝑓 (4) (𝑥) = sin 𝑥 = 𝑓(𝑥), the pattern 0, 1, 0, −1
will repeat over and over as we evaluate successive

- 320 -
 derivatives at 0. Therefore, the successive Maclaurin
polynomials for sin x are
𝑝1 (𝑥) = 0 + 𝑥 = 𝑥 ;
𝑝2 (𝑥) = 0 + 𝑥 + 0 = 𝑥 ;
𝑥3 𝑥3
𝑝3 (𝑥 ) = 0 + 𝑥 + 0 − =𝑥− ;
3! 3!
𝑥3 𝑥3
𝑝4 (𝑥 ) = 0 + 𝑥 + 0 − +0=𝑥− ;
3! 3!
𝑥3 𝑥5 𝑥3 𝑥5
𝑝5 (𝑥 ) = 0 + 𝑥 + 0 − +0+ =𝑥− + ;
3! 5! 3! 5!
𝑥3 𝑥5 𝑥3 𝑥5
𝑝6 (𝑥 ) = 0 + 𝑥 + 0 − +0+ +0=𝑥− + ;
3! 5! 3! 5!

𝑥3 𝑥5 𝑥7
𝑝7 (𝑥 ) = 0 + 𝑥 + 0 − + 0 + + 0 −
3! 5! 7!
𝑥3 𝑥5 𝑥7
=𝑥− + −
3! 5! 7!
In general, the Maclaurin polynomials for sin x are
𝑥3 𝑥5 𝑥7
𝑝2𝑛+1 (𝑥 ) = 𝑝2𝑛+2 (𝑥 ) = 𝑥 − + − + ⋯+
3! 5! 7!
𝑥 2𝑛+1
(−1)𝑛 , (𝑛 = 0, 1, 2, … ).◄
(2𝑛+1)!

Example 4.
The successive Maclaurin polynomials for cos 𝑥 are
𝑝0 (𝑥 ) = 𝑝1 (𝑥 ) = 1;

- 321 -
 𝑥2
𝑝2 (𝑥 ) = 𝑝3 (𝑥 ) = 1 − ;
2!
𝑥2 𝑥4
𝑝4 (𝑥 ) = 𝑝5 (𝑥 ) = 1 − =𝑥− ;
2! 4!
𝑥2 𝑥4 𝑥6
𝑝6 (𝑥 ) = 𝑝7 (𝑥 ) = 1 − + − ;
2! 4! 6!
𝑥2 𝑥4 𝑥6 𝑥8
𝑝8 (𝑥 ) = 𝑝9 (𝑥 ) = 1 − + − + .
2! 4! 6! 8!

In general, the Maclaurin polynomials for cos 𝑥 are

𝑥2 𝑥4 𝑥 2𝑛
𝑝2𝑛 (𝑥 ) = 𝑝2𝑛+1 (𝑥 ) = 1 − + − ⋯ + (−1)𝑛 (2𝑛)!,
2! 4!

(𝑛 = 0, 1, 2, … ).◄
Definition 2.
If f can be differentiated n times at a, then we define the
nth Taylor polynomials for f about 𝑥 = 𝑎 to be

′(
𝑓 ′′ (𝑎)
𝑝𝑛 (𝑥 ) = 𝑓(𝑎) + 𝑓 𝑎)(𝑥 − 𝑎) + (𝑥 − 𝑎)2
2!
𝑓 ′′′ (𝑎) 𝑓 𝑛 (𝑎)
+ (𝑥 − 𝑎)3 + ⋯ + (𝑥 − 𝑎)𝑛 (5)
3! 𝑛!

Observe that the Taylor polynomials include the

Maclaurin polynomials as a special case (let 𝑎 = 0).
Example 5.
Find the Taylor polynomials 𝑝1 (𝑥 ), 𝑝2 (𝑥 ), and 𝑝3 (𝑥) for
sin x about 𝑥 = 𝜋/3.

- 322 -
 Solution.
Let 𝑓(𝑥 ) = sin 𝑥 . Thus
𝜋 √3
𝑓(𝑥 ) = sin 𝑥 ⇒ 𝑓(𝜋/3) = sin = ;
3 2
𝜋 1
𝑓 ′ (𝑥 ) = cos 𝑥 ⇒ 𝑓(𝜋/3) = cos = ;
3 2
𝜋 √3
𝑓 ′′ (𝑥 ) = − sin 𝑥 ⇒ 𝑓 ′′ (𝜋/3) = − sin = − ;
3 2
𝜋 1
𝑓 ′′′ (𝑥 ) = − cos 𝑥 ⇒ 𝑓 ′′′ (𝜋/3) = − cos = − .
3 2

Substituting in (5) with 𝑎 = 𝜋/3 yields

𝜋 √3 1 𝜋
𝑝1 (𝑥 ) = 𝑓(𝜋/3) + 𝑓 ′ (𝜋/3) (𝑥 − ) = + (𝑥 − ).
3 2 2 3
𝜋
𝑝2 (𝑥 ) = 𝑓(𝜋/3) + 𝑓 ′ (𝜋/3) (𝑥 − )
3
𝑓 ′′ (𝜋/3) 𝜋 2
+ (𝑥 − )
2! 3
√3 1 𝜋 √3 𝜋 2
= + (𝑥 − ) − (𝑥 − ) .
2 2 3 2.2 ! 3
𝜋
𝜋 𝜋 𝜋 𝑓 ′′ ( ) 𝜋 2
′ 3
𝑝3 (𝑥 ) = 𝑓 ( ) + 𝑓 ( ) (𝑥 − ) + (𝑥 − )
3 3 3 2! 3
𝑓 ′′′ (𝜋/3) 𝜋 3
+ (𝑥 − )
3! 3
√3 1 𝜋 √3 𝜋 2 1 𝜋 3
= + (𝑥 − ) − (𝑥 − ) − (𝑥 − 3 ) .
2 2 3 2 .2 ! 3 2 .3 !

- 323 -
 Frequently, it is convenient to express the defining
formula for the Taylor polynomials in sigma notation. To
do this, we use the notation 𝑓 (𝑘) (𝑎) to denote the kth
derivative of f at 𝑥 = 𝑎, and we make the added
convention that 𝑓 (0) (𝑎) denotes 𝑓(𝑎). This enables us to
write
𝑓 (𝑘) (𝑎)
∑𝑛𝑥=0 (𝑥 − 𝑎)𝑘 = 𝑓(𝑎) + 𝑓 ′ (𝑎)(𝑥 − 𝑎)
𝑘!

𝑓 ′′ (𝑎) 𝑓 (𝑛) (𝑎)

+ (𝑥 − 𝑎 )2 +⋯+ (𝑥 − 𝑎)𝑛 .
2! 𝑛!

In particular, the nth Maclaurin polynomial for 𝑓(𝑥) may

be written
(𝑘)
𝑓 (𝑎) 𝑘 𝑓 ′′ (0)
∑𝑛𝑘=0 𝑥 = 𝑓(0) + 𝑓 0) + ′(
𝑥2 + ⋯ +
𝑘! 2!
𝑓 (𝑛) (0)
𝑥 𝑛 .◄
𝑛!

Definition 3.
If f derivatives of all orders at a, then we define the
Taylor Series for f about 𝑥 = 𝑎 to be
∞
𝑓 (𝑘) (𝑎)
∑ (𝑥 − 𝑎)𝑘 = 𝑓(𝑎) + 𝑓 ′ (𝑎)(𝑥 − 𝑎)
𝑘!
𝑘=0
𝑓 ′′ (𝑎) 𝑓 (𝑘) (𝑎)
+ (𝑥 − 𝑎)2 + ⋯ + (𝑥 − 𝑎)𝑘 + ⋯ (6)
2! 𝑘!

- 324 -
Definition 4.
If f has derivatives of all orders at 0, then we define the
Maclaurin series for f to be
∞
𝑓 (𝑘) (0) 𝑘
∑ 𝑥
𝑘!
𝑘=0

𝑓 ′′ (0) 2
′(
= 𝑓(0) + 𝑓 0)𝑥 + 𝑥 +⋯
2!
𝑓 (𝑘) (0) 𝑘
+ 𝑥 + ⋯ (7)
𝑘!
Observe that the Maclaurin series for f is just the Taylor
series for f about 𝑎 = 0.
Example 6.
In Example 1, we found the nth Maclaurin polynomial
for 𝑒 𝑥 to be
𝑥 𝑘 𝑥2 𝑥𝑛
∑∞
𝑘=0 𝑘 ! =1+𝑥+ + ⋯+
2! 𝑛!

Thus, the Maclaurin series for 𝑒 𝑥 is

𝑥 𝑘 𝑥2 𝑥3 𝑥𝑛
∑∞
𝑘=0 𝑘 ! =1+𝑥+ + + ⋯+ +⋯
2! 3! 𝑛!

Similarly, from Example 3 the Maclaurin series for sin x

is
∞ 𝑘 𝑥 2 𝑘+1 𝑥3 𝑥5 𝑥7
∑𝑘=0(−1) =𝑥− + − +⋯
(2𝑘+1) ! 3! 5! 7!

- 325 -
and from Example 4 the Maclaurin series for cos x is
∞
𝑥2 𝑘 𝑥2 𝑥4 𝑥6
∑ (−1 )𝑘 =1− + − +⋯
(2𝑘) ! 2! 4! 6!
𝑘=0

Example 7.
Find the Taylor series about x = 1 for 1/x.
Solution.
Let 𝑓(𝑥) = 1 / 𝑥 so that
1
𝑓(𝑥 ) = ⇒ 𝑓(1) = 1;
𝑥
1
𝑓 ′ (𝑥 ) = − ⇒ 𝑓 ′ (1) = −1;
𝑥2
2
𝑓 ′′ (𝑥 ) = ⇒ 𝑓 ′′ (1) = 2 !;
𝑥3
3 .2
𝑓 ′′′ (𝑥 ) = − ⇒ 𝑓 ′′′ (1) = −3 !;
𝑥4
4 .3 .2
𝑓 (4) (𝑥 ) = ⇒ 𝑓 (4) (1) = 4 !;
𝑥5

⋮ ⋮
𝑘!
𝑓 (𝑘) (𝑥 ) = (−1)𝑘 ⇒ 𝑓 (𝑘) (1) = (−1)𝑘 𝑘 !;
𝑥 𝑘+1

⋮ ⋮
Thus, substituting in (6) with 𝑎 = 1 yields
(−1)𝑘 𝑘 !
∑∞
𝑘=0 (𝑥 − 1)𝑘 = ∑∞ 𝑘
𝑘=0(−1) (𝑥 − 1)
𝑘
𝑘!

= 1 − (𝑥 − 1) + (𝑥 − 1)2 − (𝑥 − 1)3 + ⋯◄

- 326 -
 Exercise Set (7.7)
In Exercise (1)-(12), find the fourth Maclaurin
polynomial (𝑛 = 4) for the given function.
1
(1) 𝑒 −2𝑥 ; (2) ;
1+𝑥

(3) sin 2𝑥;(4) 𝑒 𝑥 cos 𝑥;

(5) tan 𝑥;(6) 𝑥 3 − 𝑥 2 + 2𝑥 + 1;
(7) 𝑥 𝑒 𝑥 ; (8) tan−1 𝑥;
(9) sec 𝑥; (10)√1 + 𝑥;
(11) ln (3 + 2 𝑥); (12) sinh 𝑥.
In Exercise (13)–(22), find the third Taylor polynomial
(𝑛 = 3) about 𝑥 = 𝑎 for the given function.
(13) 𝑒 𝑥 , 𝑎 = 1; (14) ln 𝑥 , 𝑎 = 1;
(15) √𝑥 , 𝑎 = 4; (16) 𝑥 4 + 𝑥 − 3 , 𝑎 = −2;
𝜋 𝜋
(17) cos 𝑥 , 𝑎 = ; (18) tan 𝑥 , 𝑎 = ;
4 3
1 𝜋
(19) sin 𝜋 𝑥 , 𝑎 = − ; (20) csc 𝑥 , 𝑎 = ;
3 2

(22) tan−1 𝑥 , 𝑎 = 1; (23) cosh 𝑥 , 𝑎 = ln 2;

In Exercise (23)–(31), find the Maclaurin series for the
given function. Express your answer in sigma notation.
1
(23) 𝑒 −𝑥 ; (24) 𝑒 𝑎 𝑥 ; (25) ; (26) 𝑥𝑒 𝑥 ;
1+𝑥

- 327 -
 𝑥
(27) ln(1 + 𝑥);(28)sin 𝜋 𝑥; (29)cos ( );
2

(30) sinh 𝑥; (31) cosh 𝑥.

In Exercise (32)–(39), find the Taylor series about 𝑥 = 𝑎
for the given function. Express your answer in sigma
notation.
1 1
(32) , 𝑎 = 3; (33) , 𝑎 = −1;
𝑥 𝑥

(34) 𝑒 𝑘 , 𝑎 = 2; (35) ln 𝑥 , 𝑎 = 1;
𝜋 1
(36) cos 𝑥 , 𝑎 = ; (37) sin 𝜋 𝑥 , 𝑎 = ;
2 2
1
(38) , 𝑎 = 3; (39) sinh 𝑥 𝑥 , 𝑎 = ln 4.
𝑥+2
𝑓 (𝑛) (𝑎)
(40) Prove that the value of𝑝𝑛 (𝑥 ) = ∑𝑛𝑘=0 (𝑥 −
𝑘!

𝑎)𝑘 and its first n derivatives match the value of 𝑓(𝑥)

and its first n derivatives at 𝑥 = 𝑎.

- 328 -
 CHAPTER (VIII)
DIFFERENTIAL EQUATION

- 329 -
 CHAPTER (VIII)
DIFFERENTIAL EQUATION

1. Introduction
The words differential and equations clearly indicate
solving some kind of equation involving derivatives.
Differential equations are interesting and important
because they express relationships involving rates of
change. Such relationships form the basis for developing
ideas and studying phenomena in the sciences,
economics, engineering, finance, medicine and in short
without any exaggeration every aspect of human
knowledge.
2. Definitions and Terminology
Definition 1.

An equation containing the derivatives of one or more

dependent variable, with respect to one or more

independent variables, is said to be a differential

equation (DE).

- 330 -
Definition 2.

A differential equation is said to be an ordinary

differential equation (ODE) if it contains only ordinary
derivatives of one or more dependent variables with
respect to a single independent variable.
Definition 3.
An equation involving the partial derivatives of one or
more dependent variables of two or more independent
variables is called a partial differential equation (PDE).
Example 1.
𝑑𝑦 𝑑2 𝑦
+ 10𝑦 = 𝑒 𝑥 , + 10𝑦 = sin 𝑥
𝑑𝑥 𝑑𝑥 2
𝑑2 𝑦 𝑑𝑦 𝑑𝑥 𝑑𝑦
− + 6𝑦 = 0, + = 2𝑥 + 𝑦
𝑑𝑥 2 𝑑𝑥 𝑑𝑡 𝑑𝑡
are examples of ordinary differential equation.
Example 2.
𝜕2 𝑢 𝜕2 𝑢 𝜕𝑦 𝜕2 𝑢 𝜕2 𝑢 𝜕2 𝑢
= 2, = 2, =−
𝜕𝑡 2 𝜕𝑥 𝜕𝑥 𝜕𝑡 𝜕𝑥 2 𝜕𝑦 2

𝜕2 𝑢 𝜕2 𝑢 𝜕𝑢 𝜕𝑢 𝜕𝑣
= −2 and =−
𝜕𝑥 2 𝜕𝑡 2 𝜕𝑡 𝜕𝑦 𝜕𝑥

are examples of partial differential equation.

- 331 -
Definition 4.
The order of a differential equation (ODE or PDE) is the
order of the highest derivative in the equation.
Example 3.
(i) Order of the differential equation
𝑑2 𝑦 𝑑𝑦 4
+ 5 ( ) − 4𝑦 = 𝑒 𝑥 is 2.
𝑑𝑥 2 𝑑𝑥

(ii) Order of the differential equation

𝑑3 𝑦 𝑑2 𝑦
+ − 4𝑦 = 𝑒 −𝑥 is 3.
𝑑𝑥 3 𝑑𝑥 2

Definition 5.
The degree of a differential equation is the degree of the
highest order derivative in the equation.
Example 4.
(i) The degree of ODE
𝑑𝑦 𝑑𝑦 2
𝑦 − √𝑥 ( ) − 5 = 0 is 2.
𝑑𝑥 𝑑𝑥

(ii) The degree of ODE

𝑑2 𝑦 𝑑𝑦 2
− ( ) + 6𝑦 + 10 = 0is 1.
𝑑𝑥 2 𝑑𝑥

Remarks 1.
(i) Very often notation 𝑦 ′ , 𝑦 ′′ , . . . , 𝑦 (𝑛) are respectively
𝑑𝑦 𝑑2 𝑦 𝑑3 𝑦 𝑑𝑛 𝑦
used for , 2, 3,…,
𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑛
- 332 -
(ii) In symbols we can express an nth order ordinary
differential equation in one dependent variable by the
general form
𝐹(𝑥, 𝑦, 𝑦 ′ , 𝑦 ′′ , . . . , 𝑦 (𝑛) ) = 0 (1)
where F is a real-valued function of 𝑛 + 2 variables
𝑥, 𝑦, 𝑦 ′ , 𝑦 ′′ , . . . , 𝑦 (𝑛) .
Definition 6.
An n th-order ordinary differential equation is said to be
linear in y if it can be written in the form
𝑎𝑛 (𝑥)𝑦 (𝑛) + 𝑎𝑛−1 (𝑥)𝑦 (𝑛−1) + . . . + 𝑎1 (𝑥)𝑦’ + 𝑎0 (𝑥)𝑦
= 𝑓(𝑥)
where𝑎0 , 𝑎1 , 𝑎2 , … , 𝑎𝑛 and f are functions of x on some
interval, and 𝑎𝑛 (𝑥)  0 on that interval. The functions
𝑎𝑘 (𝑥), 𝑘 = 0, 1, 2, . . . . , 𝑛 are called the coefficient
functions. A differential equation that is not linear is
called non-linear.
Example 5.
(i)𝑦 ′′ = 4𝑦 ′ + 3𝑦 − 𝑥 4 and 𝑥𝑦′′ + 𝑦𝑒 𝑥 + 6 = 0
are linear differential equations.
𝑑3 𝑦 𝑑𝑦
(ii) (𝑦 − 𝑥)𝑑𝑥 + 4𝑥𝑑𝑦 = 0, 𝑥 3 − 4𝑥 + 12𝑦 = 𝑒 𝑥
𝑑𝑥 3 𝑑𝑥

- 333 -
 𝑦’‘ − 4𝑦’ + 𝑦 = 0
are linear differential equations.
𝑥 𝑑2 𝑦
(iii) (1 + 𝑦)𝑦’ + 2𝑦 = 𝑒 , + cos 𝑦 = 0,
𝑑𝑥 2
𝑑3 𝑦
+ 𝑦 2 = 0are non-linear.
𝑑𝑥 3

Remark 2.
An ordinary differential equation is linear if the following
conditions are satisfied.
(i) The unknown function and its derivatives occur in the
first degree only.
(ii) There are no products involving either the unknown
function and its derivatives or two or more derivates.
(iii) There are no transcendental functions involving the
unknown function or any of its derivatives.
Definition 7.
(i) A solution or a general solution of an nth-order
differential equation of the form (1) on an interval 𝐼 =
[𝑎, 𝑏] = {𝑥ℝ: 𝑎 ≤ 𝑥 ≤ 𝑏} is any function possessing
all the necessary derivatives, which when substituted
for𝑥, 𝑦, 𝑦 ′ , 𝑦 ′′ , . . . , 𝑦 (𝑛) , reduces the differential equation
to an identity. In other words, an unknown function is a

- 334 -
solution of a differential equation if it satisfies the
equation.
(ii) A solution of a differential equation of order n will
have n independent arbitrary constants. Any solution
obtained by assigning particular numerical values to some
or all of the arbitrary constants is called a particular
solution.
(iii) A solution of a differential equation that is not
obtainable from a general solution by assigning particular
numerical values is called a singular solution.
(iv) A real function 𝑦 = (𝑥) is called an explicit
solution of the differential equation
𝐹(𝑥, 𝑦, 𝑦 ′ , 𝑦 ′′ , . . . , 𝑦 (𝑛) ) = 0on [𝑎, 𝑏]
if𝐹(𝑥,  (𝑥), ’(𝑥), . . . , (𝑛) (𝑥) ) = 0 on [𝑎, 𝑏].
(v) A relation 𝑔(𝑥, 𝑦) = 0 is called an implicit solution
of the differential equation 𝐹(𝑥, 𝑦, 𝑦 ′ , 𝑦 ′′ , . . . , 𝑦 (𝑛) ) =
0on [𝑎, 𝑏] if 𝑔(𝑥, 𝑦) = 0 defines at least one real
function f on [𝑎, 𝑏] such that 𝑦 = 𝑓(𝑥) is an explicit
solution on this interval.
We now illustrate these concepts through the following
examples:
- 335 -
Example 6.
(i) 𝑦 = 𝑐1 𝑒 𝑥+𝑐2 is a solution of the equation
𝑦’’ − 𝑦 = 0.
This ODE is of order 2 and so its solution involves 2
arbitrary constants c1 and c2. It is clear that 𝑦 ′ = 𝑐1 𝑒 𝑥+𝑐2 ,
𝑦′′ = 𝑐1 𝑒 𝑥+𝑐2 and so 𝑐1 𝑒 𝑥+𝑐2 − 𝑐1 𝑒 𝑥+𝑐2 = 0.
Hence𝑦 = 𝑐1 𝑒 𝑥+𝑐2 is a general solution or simply a
solution.
(ii)𝑦 = 𝑐𝑒 2𝑥 is a solution of ODE𝑦’ − 2𝑦 = 0,
because 𝑦′ = 2𝑐𝑒 2𝑥 and 𝑦 = 𝑐𝑒 2𝑥 satisfy the ODE. Since
given ODE is of order 1, solution contains only one
constant.
1
(iii) 𝑦 = 𝑐𝑥 + 𝑐 2 is a solution of the equation
2
1
(𝑦’)2 + 𝑥𝑦’ – 𝑦 = 0.
2

To verify the validity, we note that 𝑦’ = 𝑐, and therefore

1 2 1
(𝑐) + 𝑐𝑥 − (𝑐𝑥 + 𝑐 2 ) = 0
2 2
(iv)𝑦 = 𝑐1 𝑒 2𝑥 + 𝑐2 𝑒 −𝑥 is a general solution of the
differential equation 𝑦’’ − 𝑦’ − 2𝑦 = 0 of order 2.

- 336 -
To check the validity, we compute 𝑦’ and 𝑦’’ and put
values in these equation.
𝑦 ′ = 2𝑐1 𝑒 2𝑥 − 𝑐2 𝑒 −𝑥 , 𝑦′′ = 4𝑐1 𝑒 2𝑥 + 𝑐2 𝑒 −𝑥
L.H.S. of the given ODE is =
(4𝑐1 𝑒 2𝑥 + 𝑐2 𝑒 −𝑥 ) − (2𝑐1 𝑒 2𝑥 − 𝑐2 𝑒 −𝑥 )
− 2(𝑐1 𝑒 2𝑥 + 𝑐2 𝑒 −𝑥 ) = 0
Example 7.
(i) Choosing𝑐 = 1 we get a particular solution of
differential equation considered in Example 6(iii).
(ii) For 𝑐1 = 1 we get a particular solution of differential
equation in Example 6(i) that is,𝑦 = 𝑒 𝑥+𝑐2 is a particular
solution of𝑦′′ − 𝑦 = 0.
Example 8.
1
(i) 𝑦 = − 𝑥 2 is a singular solution of differential equation
2

in Example 6(iii).
(ii)𝑦 = 0 is a singular solution of 𝑦’ = 𝑥𝑦1/2 .
Verification.
1
(i) 𝑦 = − 𝑥 2 is not obtainable from the general solution𝑦 =
2
1
𝑐𝑥 + 𝑐 2 . However, it is a solution of the given differential
2

equation, can be checked as follows:𝑦’ = −𝑥.

- 337 -
By putting values of y and 𝑦’ into the RHS of the
equation we get
1 1
(−𝑥 )2 + 𝑥 (−𝑥 ) − (− 𝑥 2 ) = 0
2 2
1
(ii) The solution of this equation is 𝑦 = 𝑥 2 + 𝑐.
4

For 𝑐 = 0, we do not get the solution 𝑦 = 0.

Therefore, the solution 𝑦 = 0 of the equation is not
obtainable from the general solution. Hence 𝑦 = 0 is a
singular solution. ◄
Example 9.
(i) 𝑦 = sin 4𝑥 is an explicit solution of𝑦’’ + 16𝑦 = 0 for
all real x.
(ii) 𝑦 = 𝑐1 𝑒 𝑥 + 𝑐2 𝑒 −𝑥 is an explicit solution of the
equation 𝑦’’ − 𝑦 = 0.
Verification.
(i) 𝑦’ = −4 cos 4𝑥, 𝑦’’ = −16 sin 4𝑥.
Putting the value of y and 𝑦’’ in terms of x into the RHS
of the equation we get – 16 sin 4𝑥 + 16 sin 4𝑥 =
0.Hence equation is satisfied for 𝑦 = sin 4𝑥. Therefore
𝑦 = sin4𝑥 is an explicit solution of the given equation.
(ii) 𝑦′ = 𝑐1 𝑒 𝑥 − 𝑐2 𝑒 −𝑥 , 𝑦′′ = 𝑐1 𝑒 𝑥 + 𝑐2 𝑒 −𝑥 .
- 338 -
Put values of y and 𝑦’’ in the RHS of the given equation
to get (𝑐1 𝑒 𝑥 + 𝑐2 𝑒 −𝑥 ) − (𝑐1 𝑒 𝑥 + 𝑐2 𝑒 −𝑥 ) = 0.◄
Example 10.
(i) The relation 𝑥 2 + 𝑦 2 = 4 is an implicit solution of
𝑑𝑦 𝑥
the differential equation = − on the interval (−2,2).
𝑑𝑥 𝑦

(ii) The relation 𝑦 2 + 𝑥 − 4 = 0 is an implicit solution

of 2𝑦𝑦’ + 1 = 0 on the interval (−, 4).
Verification.
(i) By implicit differentiation of the relation
𝑑𝑦 𝑑𝑦 𝑥
𝑥 2 + 𝑦 2 = 4 we get 2𝑥 + 2𝑦 = 0 or =−
𝑑𝑥 𝑑𝑥 𝑦

Further,𝑦1 = √4 − 𝑥 2 and 𝑦2 = −√4 − 𝑥 2

2
satisfying the relation 𝑦 = 4 − 𝑥 2 and 𝑦 = ±√4 − 𝑥 2 .
𝑑𝑦 𝑥
and are solutions of the differential equation =− .
𝑑𝑥 𝑦
1 1 𝑥
It is clear that𝑦′1 = 2
(−2𝑥 ) = −
2 √4−𝑥 𝑦1
1 1 𝑥
And𝑦′2 = − 2
(−2𝑥 ) = − .
2 √4−𝑥 𝑦2

(ii) Differentiating 𝑦 2 + 𝑥 − 4 = 0 with respect to x, we

obtain 2𝑦𝑦’ + 1 = 0, which is the given differential
equation. Hence 𝑦 2 + 𝑥 − 4 = 0is an implicit solution

- 339 -
if it defines a real function on (−, 4). Solving the
equation 𝑦 2 + 𝑥 − 4 = 0for y, we get𝑦 =
±√4 − 𝑥.Since both𝑦1 = √4 − 𝑥 and𝑦2 = −√4 − 𝑥
and their derivatives are functions defined for all x in the
interval (−, 4), we conclude that 𝑦 2 + 𝑥 − 4 = 0 is
an implicit solution on this interval. ◄
Remark 3.
It is very pertinent to note that a relation 𝑔(𝑥, 𝑦) = 0
can reduce a differential to an identity without
constituting an implicit solution of the differential
equation. For example 𝑥 2 + 𝑦 2 + 1 = 0satisfies𝑦𝑦’ +
𝑥 = 0, but it is not an implicit solution as it does not
define a real-valued function. This is clear from the
solution of the equation 𝑥 2 + 𝑦 2 + 1 = 0 or𝑦 =
±√4 − 𝑥 2 , imaginary number.The relation 𝑥 2 + 𝑦 2 +
1 = 0is called a formal solution of 𝑦𝑦’ + 𝑥 = 0. That is
it appears to be a solution. Very often we look for a
formal solution rather than an implicit solution.◄

- 340 -
♠ Formation of a differential Equation
Let us consider an equation containing n arbitrary
constants. Then by differentiating it successively n times
we get n more equations containing n arbitrary constants
and derivatives. Now by eliminating n arbitrary constants
from the above (𝑛 + 1) equations and obtaining an
equation which involves derivatives up to the nth order,
we get a differential equation of order n. The concept of
obtaining differential equations from a family of curves is
illustrated in following examples.
Example 11.
Find the differential equation of the family curves𝑦 =
𝑐𝑒 2𝑥 .
Solution.
Given 𝑦 = 𝑐𝑒 2𝑥 . Differentiating we get
𝑦 ′ = 2𝑐𝑒 2𝑥 = 2𝑦
or
𝑦’ − 2𝑦 = 0 (3)
Thus, arbitrary constant c is eliminated and equation (3)
is the required equation of the family of curves given by
equation (2).◄

- 341 -
Example 12.
Find the differential equation of the family of curves
𝑦 = 𝑐1 cos 𝑥 + 𝑐2 sin 𝑥 (4)
Solution.
Differentiating (4) twice we get
𝑦 ′ = −𝑐1 sin 𝑥 + 𝑐2 cos 𝑥 (5)
𝑦 ′′ = − 𝑐1 cos 𝑥 − 𝑐2 sin 𝑥 (6)
c1 and c2 can be eliminated from (4) and (6) and we
obtain the different equation
𝑦’’ + 𝑦 = 0 (7)
Therefore (7) is the differential equation of the family of
curves given by (4).◄
Example 13.
Find the differential equation which has
𝑦 = 𝑐1 𝑒 𝑥 + 𝑐2 𝑒 −𝑥 + 3𝑥
as its general solution.
Solution.
Differentiate the given expression twice
𝑦 = 𝑐1 𝑒 𝑥 + 𝑐2 𝑒 −𝑥 + 3𝑥(8)
𝑦 / = 𝑐1 𝑒 𝑥 − 𝑐2 𝑒 −𝑥 + 3(9)
𝑦 // = 𝑐1 𝑒 𝑥 + 𝑐2 𝑒 −𝑥 (10)
- 342 -
Eliminate 𝑐1 and 𝑐2 by subtracting (10) from (8) to obtain
𝑦 − 𝑦 // = 3𝑥 gives which is the desired differential
equation.Note that the desired differential equation is free
from the arbitrary constants.◄
3. Initial-Value and Boundary-Value Problems
A general solution of an nth order ODE contains n
arbitrary constants. To obtain a particular solution, we are
required to specify n conditions on solution function and
its derivatives and thereby expect to find values of n
arbitrary constants. There are two well known methods
for specifying auxiliary conditions. One is called initial
conditions and other is said to be boundary conditions.
It may be observed that an ODEdoes not have solution or
unique solution. However, by imposing initial and
boundary conditions uniqueness can be ensured for
certain classes of differential equations.
Definition 1. (initial-Value Problem)
If the auxiliary conditions for a given differential
equation relate to a single x value, the conditions are
called initial conditions. The differential equation with its
initial conditions is called an initial-value problem.

- 343 -
Definition 2. (boundary-value problem)
If the auxiliary conditions for a given differential
equation relate to two or more x values, the conditions are
called boundary conditions or boundary values. The
differential equation with its boundary conditions is
called boundary-value problem.

Example 1.

- 344 -
(i) 𝑦′ + 𝑦 = 3, 𝑦(0) = 2 is a first-order initial value
problem. Order of initial value problem is nothing but
order of the given equation. 𝑦(0) = 2 is an initial
condition.
(ii) 𝑦’’ + 2𝑦 = 0, 𝑦(1) = 2, 𝑦’(1) = −3 is a second-
order initial value problem. Initial conditions are 𝑦(1) =
2 and 𝑦’(1) = −3. Values of function 𝑦(𝑥) and its
derivative are specified for value 𝑥 = 1.
(iii) 𝑦’’ − 𝑦’ + 𝑦 = 𝑥 3 , 𝑦(0) = 4, 𝑦’(1) = −2 is a
second-order boundary-value problem.
Boundary conditions are specified at two points namely
𝑥 = 0 and 𝑥 = 1.
One may specify boundary conditions for different values
of x say 𝑥 = 2 and 𝑥 = 5. In this case the boundary-
value problem is
𝑦’’ − 𝑦’ + 𝑦 = 𝑥 3 , 𝑦(2) = 4, 𝑦’(5) = −2.◄
● The following questions are quite pertinent as boundary
value and initial value problems represent important
phenomena in nature:

- 345 -
Problem 1.
When does a solution exist? That is, does an initial-value
problem or a boundary value problem necessarily have a
solution?
Problem 2.
Is a known solution unique? That is, is there only one
solution of an initial-value problem or a boundary-value
problem?
The following theorem states that under the specified
conditions, a first-order initial-value problem has a
unique solution.
Theorem 1.
𝜕𝑓
Let f and 𝑓𝑦 ( ) be continuous functions of x and y in
𝜕𝑦

some rectangle R of the xy-plane, and let (𝑥∘ , 𝑦∘ ) be a

point in that rectangle. Then on some interval centred at
𝑥∘ there is a unique solution 𝑦 = (𝑥) of the initial value
problem:
𝑑𝑦
= 𝑓(𝑥, 𝑦), 𝑦(𝑥∘ ) = 𝑦∘
𝑑𝑥
Figure 1 Geometrical Illustration of Theorem 1.

- 346 -
Example 2.
(i) 𝑦 = 3 𝑒 𝑥 is a solution of the initial-value
problem𝑦’ = 𝑦, 𝑦(0) = 3
This means that the solution of the differential equation
𝑦’ = 𝑦passes through the point (0, 3).
(ii) Find a solution of the initial-value problem 𝑦’ = 𝑦,
𝑦(1) = −2. That is, find a solution of differential
equation 𝑦’ = 𝑦 which passes through the point (1, −2).
Verification.
(i) Let 𝑦 = 𝑐 𝑒 𝑥 , where c is an arbitrary constant.
Then 𝑦′ = 𝑐 𝑒 𝑥 = 𝑦.Thus, 𝑦 = 𝑐 𝑒 𝑥 is a general
solution of the given equation 𝑦’ = 𝑦.
By applying initial condition, we get 3 = 𝑦(0) =
𝑐 𝑒 0 = 𝑐 or c = 3. Therefore 𝑦 = 3 𝑒 𝑥 is a solution of
the given initial value problem.
(ii) As seen in part (i) 𝑦 = 𝑐 𝑒 𝑥 is a solution of the
given equation. By imposing given initial condition we
get −2 = 𝑦(1) = 𝑐 𝑒 1 or 𝑐 = −2/𝑒.
2
Therefore𝑦 = − 𝑒 𝑥 = −2𝑒 𝑥−1 is a solution of the
𝑒

initial-value problem.◄

- 347 -
Example 3.
𝑑𝑦
= 𝑥𝑦1/2 , 𝑦(0) = 0
𝑑𝑥
has at least two solutions, namely 𝑦 = 0 and
𝑦 = 𝑥 4 /16.
Example 4.
(i) Does a solution of the boundary value problem 𝑦’’ +
𝑦 = 0, 𝑦(0) = 0, 𝑦() = 2 exist?
(ii) Show that the boundary value problem
𝑦’’ + 𝑦 = 0, 𝑦(0) = 0, 𝑦() = 0
Has infinitely many solutions.
Solution.
(i)𝑦 = 𝑐1 cos 𝑥 + 𝑐2 sin 𝑥 is a solution of the
differential equation 𝑦’’ + 𝑦 = 0.Using given boundary
conditions in𝑦 = 𝑐1 cos 𝑥 + 𝑐2 sin 𝑥, we get
0 = 𝑐1 cos 0 + 𝑐2 sin 0
and
2 = 𝑐1 cos 𝜋 + 𝑐2 sin 𝜋
The first equation yields 𝑐1 = 0 and the second yields
𝑐1 = −2 which is absurd, hence no solution exists.

- 348 -
(ii) The boundary values yield
0 = 𝑐1 cos 0 + 𝑐2 sin 0
and
0 = 𝑐1 cos 𝜋 + 𝑐2 sin 𝜋
Both of these equations lead to the fact that 𝑐1 = 0. The
constant c2 is not assigned a value and therefore takes
arbitrary values. Thus, there are infinitely many solutions
represented by 𝑦 = 𝑐2 sin 𝑥.◄
Example 5.
Examine existence and uniqueness of a solution of the
following initial-value problems:
𝑦
(i) 𝑦 ′ = , 𝑦(2) = 1;
𝑥
𝑦
(ii) 𝑦 ′ = , 𝑦(0) = 3;
𝑥

(iii) 𝑦 ′ = −𝑥, 𝑦(0) = 2.

Solution.
(i) We examine whether conditions of Theorem 1 in this
section are satisfied. To check, we observe that
𝑦 𝜕𝑓 1
𝑓(𝑥, 𝑦) = and (𝑥, 𝑦) = .
𝑥 𝜕𝑦 𝑥

Both functions are continuous except at 𝑥 = 0.

- 349 -
 𝜕𝑓
Hence f and satisfy the conditions of Theorem 1 in any
𝜕𝑦

rectangle R that does not contain any part of the y -

axis(𝑥 = 0). Since the point (2, 1) is not on the y-axis,
1
there is a unique solution. One can check that 𝑦 = 𝑥 is
2

the only solution.

𝜕𝑓
(ii) In this problem neither f nor in continuous at 𝑥 =
𝜕𝑦

0, which means that (0, 3) cannot be included in any

𝜕𝑓
rectangle R where f and are continuous.
𝜕𝑦

Hence, we cannot conclude anything from Theorem 1.

However, it can be verified that 𝑦 = 𝑐𝑥 is a general
𝑦
solution of𝑦’ = , but that a particular solution cannot be
𝑥

found whose graph passes through the point (0, 3).

(iii) It can be seen that conditions of Theorem 1 are
satisfied. Therefore, the initial-value problem has a
unique solution. ◄

- 350 -
4. Solutions of First Order Differential Equations
The systemic development of techniques for solving
differential equations logically begins with the equations
of the first order and first degree.
Equations of this type can in general, be written as
𝑑𝑦
= 𝑓(𝑥, 𝑦), (1)
𝑑𝑥

where𝑓(𝑥, 𝑦) is a given function. However, despite the

apparent simplicity of this equation, analytic solutions
are usually possible only when 𝑓 (𝑥, 𝑦) has simple forms.
Type (I) Exact
If you are asked to solve the Differential Equation
𝑑𝑦
= cos 𝑥 + 3𝑥 2
𝑑𝑥
you can integrate immediately giving
𝑦 = sin 𝑥 + 𝑥 3 + 𝐴.
In the same way the following second order Differential
Equation can be solved directly by integration:
𝑑2 𝑦
2
= 3𝑒 3𝑥 + 4𝑥 − 𝑠𝑖𝑛 𝑥
𝑑𝑥
𝑑𝑦
= 𝑒 3𝑥 + 2𝑥 2 + 𝑐𝑜𝑠 𝑥 + 𝐴
𝑑𝑥

- 351 -
 𝑒 3𝑥 2𝑥 3
𝑦= + + 𝑠𝑖𝑛 𝑥 + 𝐴𝑥 + 𝐵
3 3
Note that integrating twice has introduced the 2 arbitrary
constants A and B.◄

𝒅𝒏 𝒚
♣ If you have a DE of the form = 𝒇(𝒙)then you can,
𝒅𝒙𝒏

in theory, integrate n times to find the solution. This

assumes that you can integrate at each step (remember
that you cannot integrate every function algebraically).
Your solution will contain n arbitrary constants.

Type (II)Variables separable

If 𝑓(𝑥, 𝑦) = 𝑓(𝑥)𝑔(𝑦)where𝑓(𝑥) and𝑔(𝑥) are
respectively functions of x only and y only, then, (1)
becomes
𝑑𝑦
= 𝑓(𝑥)𝑔(𝑦)
𝑑𝑥
Since the variables x and yare now separable, we have
𝑑𝑦
∫ = ∫ 𝑓(𝑥)𝑑𝑥
𝑔(𝑦)
which expresses y implicitly in terms of x.

- 352 -
Example 1.
𝑑 𝑦 𝑦+1
Solve the equation =
𝑑𝑥 𝑥−1

Solution.
Rewriting the given equation, we get:
𝑑𝑦 𝑑𝑥
∫ =∫
𝑦+1 𝑥−1
or 𝑙𝑜𝑔𝑒 ( 𝑦 + 1) = 𝑙𝑜𝑔𝑒 ( 𝑥 − 1) + 𝑙𝑜𝑔𝑒 𝐶
where 𝐶is an arbitrary constant.
𝑦+1
Hence = 𝐶 is the general solution.◄
𝑥−1

Exercise.
Given the boundary condition that 𝑦 = 1 at 𝑥 = 0.
Use the expression for the general solution in the above
example to work out the particular solution.
[Solution𝑦 = 2(1 − 𝑥) − 1)].
Example 2.
Solve the equation
𝑑𝑦 2𝑥 + 1
=
𝑑𝑥 𝑦
Solution.
Rewriting the given equation we obtain
𝑦𝑑𝑦 = (2𝑥 + 1)𝑑𝑥
- 353 -
Or,∫ 𝑦𝑑𝑦 = ∫(2𝑥 + 1)𝑑𝑥.
1
Hence 𝑦 2 = 𝑥 2 + 𝑥 + 𝑐,where c is an arbitrary
2

constant. Thus 𝑦 = √2𝑥 2 + 2𝑥 + 𝑐is the general

solution.◄
Example 3.
Find the general solution of the equation
𝑑𝑦
( 𝑥 2 + 1) − 𝑥𝑦 = 0
𝑑𝑥
Solution.
𝑑𝑦 𝑥𝑦
This gives = (𝑥 2 .
𝑑𝑥 +1)
𝑑𝑦 𝑥𝑑𝑥
Or, = (𝑥 2 .
𝑦 +1)

𝑑𝑦 𝑥𝑑𝑥
Integrating ∫ = ∫ (𝑥 2 .
𝑦 +1)

1
ln|𝑦| = ln|𝑥 2 + 1| + 𝐴
2

And simplify, putting ln 𝐵 = 𝐴

ln|𝑦| = ln √𝑥 2 + 1 + ln 𝐵 or ln|𝑦| = ln(𝐵√𝑥 2 + 1).

Then 𝑦 = 𝐵√𝑥 2 + 1 is the general solution.◄

- 354 -
Example 4.
4𝑥 2 +1 𝑑𝑦
Solve the equation = 𝑥𝑦 given 𝑦(1) = 0.
𝑦+1 𝑑𝑥

Solution.
4𝑥 2 +1
This can be written as 𝑑𝑥 = 𝑦(𝑦 + 1)𝑑𝑦.
𝑥

Simplify and integrate

1
∫ (4𝑥 + ) 𝑑𝑥 = ∫(𝑦 2 + 𝑦)𝑑𝑦
𝑥
2 𝑦3 𝑦2
Then the general solution is 2𝑥 + ln|𝑥 | = + + 𝐴.
3 2

When 𝑥 = 1, 𝑦 = 0, giving 𝐴 = 2. Therefore, the

particular solution is
𝑦3 𝑦2
2𝑥 2 + ln|𝑥 | = + + 2.◄
3 2

Type (III)Homogeneous differential equation

An expression of the nth degree in x and y is said to be
homogeneous of degree n, if when x and y are replaced
by𝑡𝑥 and𝑡𝑦,the result will be the original expression
multiplied by𝑡 𝑛 ; symbolically 𝑓(𝑡𝑥, 𝑡𝑦) = 𝑡 𝑛 𝑓(𝑥, 𝑦).
Example 1.
Show that 𝑥 2 + 𝑥𝑦 − 𝑦 2 is homogeneous and determine
the degree.

- 355 -
Solution.
Replace x and y by 𝑡𝑥 and𝑡𝑦 respectively, to
obtain(𝑡𝑥 )2 + (𝑡𝑥 )(𝑡𝑦) − (𝑡𝑦)2 = 𝑡 2 (𝑥 2 + 𝑥𝑦 − 𝑦 2 ).
The degree is 2.◄
♠ Consider the differential equation
𝑀(𝑥, 𝑦)𝑑𝑥 + 𝑁(𝑥, 𝑦)𝑑𝑦 = 0.
The equation is said to be homogeneous in x and y if M
and N are homogeneous functions of the same degree in x
and y. The technique for solving this equation is to make
the substitution 𝑦 = 𝑣𝑥 or 𝑥 = 𝑣𝑦 and is based upon the
following theorem.
Theorem 1.
Any homogeneous differential of the first order and first
degree can be reduced to the type of variables separable
by a substitution of 𝑦 = 𝑣𝑥 or 𝑥 = 𝑣𝑦.
Example2.
Find the general solution of the differential equation
2𝑥𝑦𝑑𝑥 − (𝑥 2 − 𝑦 2 )𝑑𝑦 = 0.
Solution.
A simple check reveals that the equation is homogeneous.

- 356 -
Let𝑦 = 𝑣𝑥and 𝑑𝑦 = 𝑣𝑑𝑥 + 𝑥𝑑𝑣, and substitute into the
differential equation to obtain
2𝑥(𝑣𝑥)𝑑𝑥 − (𝑥 2 − 𝑣 2 𝑥 2 )(𝑣𝑑𝑥 + 𝑥𝑑𝑣) = 0
𝑥 2 (𝑣 + 𝑣 3 )𝑑𝑥 + 𝑥 3 (𝑣 2 − 1)𝑑𝑣 = 0.
Dividing by 𝑥 3 (𝑣 + 𝑣 3 ) separates the variables:
𝑑𝑥 (𝑣 2 −1)𝑑𝑣
+ = 0. Integration yields
𝑥 𝑣(1+𝑣 2 )

𝑙𝑛 𝑥 − 𝑙𝑛 𝑣 + 𝑙𝑛( 𝑣 2 + 1) = 𝑙𝑛 𝑐
or 𝑥(𝑣 2 + 1) = 𝑐𝑣
𝑦
Rewriting the original variables by substituting 𝑣 = , we
𝑥

obtain 𝑦 2 + 𝑥 2 = 𝑐𝑦 as the general solution.◄

Problem.
𝑑𝑦
Solve: 2𝑥𝑦 = 𝑥 2 + 𝑦 2 , given that 𝑦(1) = 0.
𝑑𝑥

Hint: Write 𝑦 = 𝑣𝑥 implying that 𝑑𝑦 = 𝑣𝑑𝑥 + 𝑣𝑑𝑥.

Convert the resulting equations to first order separable
equation in terms of v and x and solve. Then substitute
𝑦
𝑣= to obtain the required solution in terms of x and y.
𝑥

Example3.
𝑑𝑦 𝑦(𝑥+𝑦)
Solve the equation = , given 𝑦(1) = 1.
𝑑𝑥 𝑥(𝑦−𝑥)

- 357 -
Solution.
Substitute𝑦 = 𝑣𝑥 we obtain
𝑑𝑣 𝑣𝑥(𝑥+𝑣𝑥) 𝑣(1+𝑣)𝑥 2
𝑥 +𝑣 = = .
𝑑𝑥 𝑥(𝑣𝑥−𝑥) (𝑣−1)𝑥 2

Simplifying and integrating

1 𝑑𝑥
∫ (1 − ) 𝑑𝑣 = 2 ∫
𝑣 𝑥
𝑣 − ln|𝑣| = 2 ln|𝑥 | + 𝑐
𝑣 = ln|𝑣| + 2 ln|𝑥 | + ln 𝐴 = ln(𝐴𝑥 2 𝑣)
𝑦 𝑦
= ln (𝐴𝑥 2 ) = ln(𝐴𝑥𝑦)
𝑥 𝑥
𝑦 = 𝑥 ln(𝐴𝑥𝑦).
But 𝑦(1) = 1; substituting in gives1 = ln 𝐴 ⟹ 𝐴 = 𝑒.
The solution can be written as𝑦 = 𝑥 ln(𝑒𝑥𝑦) or 𝑦 =
𝑥(1 + ln(𝑥𝑦)).◄
♠It is important to realize that there are often many
different ways of writing the final solution. It depends on
how it is to be used as to which is the most useful form.
Example 4.
Solve the equation
𝑑𝑦
𝑥 − 𝑦 = √𝑥 2 − 𝑦 2 .
𝑑𝑥

- 358 -
Solution.
𝑑𝑦 √𝑥 2 −𝑦 2 +𝑦
This can be written as = .
𝑑𝑥 𝑥

Substitute 𝑦 = 𝑣𝑥, we obtain

𝑑𝑣 √𝑥 2 −𝑣 2 𝑥 2 +𝑣𝑥 𝑥√(1−𝑣 2 )+𝑣𝑥
𝑥 +𝑣 = = .
𝑑𝑥 𝑥 𝑥
𝑑𝑣
Thus 𝑥 + 𝑣 = √1 − 𝑣 2 + 𝑣.
𝑑𝑥
𝑑𝑣 𝑑𝑥
Separate the variables = .
√1−𝑣 2 𝑥
𝑑𝑣 𝑑𝑥
Integrate ∫ =∫ . Then  𝑠𝑖𝑛−1 𝑣 = 𝑙𝑛|𝑥 | + 𝑐.
√1−𝑣 2 𝑥
𝑦 𝑦
Put 𝑣 = and 𝑐 = ln 𝐴. So  𝑠𝑖𝑛−1 = 𝑙𝑛|𝑥 | + 𝑐 =
𝑥 𝑥

𝑙𝑛|𝐴𝑥 |. Therefore  𝑦 = 𝑥 𝑠𝑖𝑛(𝑙𝑛|𝐴𝑥 |) .◄

Example5.
Solve the equation
𝑑𝑦
𝑥 = 𝑥 + 𝑦.
𝑑𝑥

Solution.
𝑑𝑦 𝑥+𝑦
Rearrange the given equation: = .
𝑑𝑥 𝑥
𝑑𝑣 𝑥+𝑣𝑥
Putting 𝑦 = 𝑣𝑥 gives:𝑥 +𝑣 = = 1 + 𝑣.
𝑑𝑥 𝑥
𝑑𝑣
Simplify: 𝑥 = 1.
𝑑𝑥
𝑑𝑥
Separate the variables: 𝑑𝑣 =
𝑥

- 359 -
 1
Integrate: 𝑣 = ∫ 𝑑𝑥 = 𝑙𝑛|𝑥 | + 𝐴.
𝑥

Put 𝐴 = ln 𝐵, we obtain 𝑣 = 𝑙𝑛|𝑥 | + 𝑙𝑛 𝐵 = 𝑙𝑛|𝐵𝑥 |.

𝑦 𝑦
Put 𝑣 = , we have = 𝑙𝑛|𝐵𝑥 |.
𝑥 𝑥

So, the solution is  𝑦 = 𝑥 𝑙𝑛|𝐵𝑥 | .◄

Type (VI)Linear Equations
𝑑𝑦
A linear equation is one of the form + 𝑃(𝑥 )𝑦 = 𝑄 (𝑥 ).
𝑑𝑥

The method used to solve such equations depends, again,

on noticing a particular pattern.
If𝑧 = 𝑦𝑒 𝑓(𝑥) then differentiating (using the product rule
and the chain rule
𝑑𝑧 𝑑𝑦 𝑓(𝑥) 𝑑𝑦
= 𝑒 + 𝑦𝑓′(𝑥 )𝑒 𝑓(𝑥) = 𝑒 𝑓(𝑥) ( + 𝑓′(𝑥 )𝑦)
𝑑𝑥 𝑑𝑥 𝑑𝑥
When compared to the linear equation above, this
suggests that if we can find an 𝑒 𝑓(𝑥) to multiply both
sides of the linear equation by then we’ll be able to
integrate the left-hand side immediately.
𝑑𝑦
Multiplying the linear equation + 𝑃(𝑥 )𝑦 = 𝑄 (𝑥 )
𝑑𝑥
𝑑𝑦
by𝑒 𝑓(𝑥) gives 𝑒 𝑓(𝑥) ( + 𝑃(𝑥 )𝑦) = 𝑒 𝑓(𝑥) 𝑄(𝑥 )
𝑑𝑥

- 360 -
 𝑑𝑦
Comparing left hand side with 𝑒 𝑓(𝑥) ( + 𝑓′(𝑥 )𝑦)
𝑑𝑥

(which can be integrated immediately) shows that we

want to find𝑓(𝑥), where𝑓′(𝑥 ) = 𝑃(𝑥 ) ⇒ 𝑓(𝑥 ) =
∫ 𝑃(𝑥 )𝑑𝑥i.e. multiply the equation by𝑒 𝑓(𝑥) = 𝑒 ∫ 𝑃(𝑥)𝑑𝑥
which is called the Integrating Factor.
Example1.
Solve the equation
𝑑𝑦 2𝑥
− 3𝑦 = 𝑒 .
𝑑𝑥
Solution.
From above the Integrating Factor is𝑒 ∫ −3𝑑𝑥 = 𝑒 −3𝑥 .
Multiplying through gives
𝑑𝑦
𝑒 −3𝑥 𝑑𝑥 − 3𝑒 −3𝑥 𝑦 = 𝑒 −3𝑥 𝑒 2𝑥 = 𝑒 −𝑥 .
The left-hand side is the differential of𝑦𝑒 −3𝑥 .
𝑑(𝑦𝑒 −3𝑥 )
Then, = 𝑒 −𝑥 .
𝑑𝑥

𝑦𝑒 −3𝑥 = ∫ 𝑒 −𝑥 𝑑𝑥 = −𝑒 −𝑥 + 𝐴

𝑦 = 𝑒 2𝑥 + 𝐴𝑒 3𝑥 .◄

- 361 -
Example 2.
Given
𝑑𝑦
2 cos 𝑥 + 4𝑦 sin 𝑥 = sin 2 𝑥
𝑑𝑥
𝜋
Find 𝑦(𝑥) (y as a function of x), given that 𝑦( ) = 0.
3

Solution.
Divide through by 2 cos 𝑥 to write it in the form
𝑑𝑦
+ 𝑃(𝑥 )𝑦 = 𝑄 (𝑥 )
𝑑𝑥
dy 2 sin x sin 2 x 2 sin x cos x
+ y= =
dx cos x 2 cos x 2 cos x
dy
+ (2 tan x ) y = sin x
dx
The equation is linear, integrating factor
2𝑥
𝑒 ∫ 2 𝑡𝑎𝑛 𝑥 𝑑𝑥 = 𝑒 ln 𝑠𝑒𝑐 = 𝑠𝑒𝑐 2 𝑥
Multiply by I. F. We obtain
𝑑𝑦
𝑠𝑒𝑐 2 𝑥 + (2 𝑠𝑒𝑐 2 𝑥 𝑡𝑎𝑛 𝑥 )𝑦 = 𝑠𝑒𝑐 2 𝑥 𝑠𝑖𝑛 𝑥
𝑑𝑥
Then
𝑑(𝑦 𝑠𝑒𝑐 2 𝑥 ) 𝑠𝑖𝑛 𝑥
   = 𝑠𝑒𝑐 2 𝑥 𝑠𝑖𝑛 𝑥 =
𝑑𝑥 𝑐𝑜𝑠 2 𝑥
𝑠𝑖𝑛 𝑥 1
Integrate 𝑦 𝑠𝑒𝑐 2 𝑥 = ∫   𝑑𝑥 = + 𝐴.
𝑐𝑜𝑠2 𝑥 𝑐𝑜𝑠 𝑥

- 362 -
Hence 𝑦 = 𝑐𝑜𝑠 𝑥 + 𝐴 𝑐𝑜𝑠 2 𝑥.
𝜋
But𝑦 = 0 when𝑥 = .
3
𝜋 𝜋
Hence 0 = 𝑐𝑜𝑠 + 𝐴 𝑐𝑜𝑠 2
3 3
1 1
0= + 𝐴 ⇒ 𝐴 = −2
2 4
Particular solution𝑦(𝑥) = 𝑐𝑜𝑠 𝑥 − 2 𝑐𝑜𝑠 2 𝑥.◄
Example 3.
𝑑𝑦
Solve the equation −𝑦 =𝑥
𝑑𝑥

Solution.
𝑑𝑦
The I. F. is 𝑒 ∫ −1𝑑𝑥 = 𝑒 −𝑥 giving 𝑒 −𝑥 − 𝑦𝑒 −𝑥 =
𝑑𝑥

𝑥𝑒 −𝑥 .
𝑑(𝑦𝑒 −𝑥 )
Hence = 𝑥𝑒 −𝑥 .
𝑑𝑥

So 𝑦𝑒 −𝑥 = ∫ 𝑥𝑒 −𝑥  𝑑𝑥.
Integrating the right side by parts we obtain
𝑦𝑒 −𝑥 = −𝑥𝑒 −𝑥 − 𝑒 −𝑥 + 𝐴.
Thus
𝑦 = 𝐴𝑒 𝑥 − 𝑥 − 1.◄

- 363 -
 Exercise Set (8.1)
1. Classify the given differential equation by order and
tell whether it is linear or non-linear.
(a) 𝑦’ + 2𝑥𝑦 = 𝑥 2 ; (b) 𝑦’ (𝑦 + 𝑥) = 5;
(c) 𝑦 sin 𝑦 = 𝑦’’; (d) 𝑦 cos 𝑦 = 𝑦’’’;
(e)cos 𝑦 𝑑𝑦 = sin 𝑥 𝑑𝑥; (f)𝑦’’ = 𝑒 𝑦 .
2. State whether the given differential equation is linear
or non-linear. Write the order of each equation.
(a) (1 − 𝑥 )𝑦’’ − 6𝑥𝑦’ + 9𝑦 = sin 𝑥;
𝑥𝑑3 𝑦 𝑑𝑦
(b) − 2( )2 + 𝑦 = 0;
𝑑𝑥 3 𝑑𝑥

(c) 𝑦𝑦’ + 2𝑦 = 2 + 𝑥 2 ;
𝑑2 𝑦
(d) + 9𝑦 = 𝑠𝑖𝑛 𝑦;
𝑑𝑥 2
𝑑𝑦 𝑑2 𝑦 2 1
(e) = (1 + ( 2 ) )2 ;
𝑑𝑥 𝑑𝑥
𝑑2 𝑟 𝑘
(f) = − 2.
𝑑𝑡 2 𝑟

3. Verify that, in problems 4 to 9 the indicated function is

a solution of the given differential equation. In some
cases, assume an appropriate interval
4. 2𝑦’ + 𝑦 = 0; 𝑦 = 𝑒 −𝑥/2 ;
5. 𝑦’ = 25 + 𝑦 2 ; 𝑦 = 5 tan 5𝑥

- 364 -
 6. 𝑥 2 𝑑𝑦 + 2𝑥𝑦 𝑑𝑥 = 0; 𝑦 = −1/𝑥 2
7. 𝑦’’’ − 3𝑦’’ + 3𝑦’ − 𝑦 = 0; 𝑦 = 𝑥 2 𝑒 𝑥
8. 𝑦’ = 𝑦 + 1; 𝑦 = 𝑒 𝑥 − 1
9. 𝑦’’ + 9𝑦 = 8 sin 𝑥 ;
𝑦 = sin 𝑥 + 𝑎 cos 3𝑥 + 𝑏 sin 3𝑥
10. Solve the following differential equations
𝑑𝑦 𝑑𝑦
(i) 𝑥 = 𝑦2 , (ii) 𝑥 (𝑥 2 + 1) = ;
𝑑𝑥 𝑑𝑥
2𝑥𝑦 𝑑𝑦 𝑑𝑦 𝑥 2 +𝑥
(iii) = ;(iv)   = ;
𝑥 2 +1 𝑑𝑥 𝑑𝑥 𝑦 2 +𝑦
𝑑𝑦 𝑑𝑦
(v) 𝑐𝑜𝑠 2 𝑥 = 𝑐𝑜𝑠 2 𝑦; (vi)  𝑒 𝑥+𝑦 = 1;
𝑑𝑥 𝑑𝑥
𝑑𝑦 𝑦−1
(vii) 𝑥 = − 𝑦;
𝑑𝑥 𝑦+1
𝑑𝑦
(viii) (𝑥 + 1)𝑦 = 𝑥 ,  where 𝑦(1) = 1;
𝑑𝑥
𝑑𝑦 𝜋
(ix) 𝑡𝑎𝑛 𝑥 = 𝑐𝑜𝑡 𝑦 where 𝑦(0) = ;
𝑑𝑥 2
𝑑𝑦
(x)   (𝑠𝑖𝑛 𝑥 + 𝑐𝑜𝑠 𝑥 ) = 𝑐𝑜𝑠 𝑥 − 𝑠𝑖𝑛 𝑥.
𝑑𝑥

11. Find the equation of the curve which has gradient

𝑑𝑦 𝑦
function  =   and passes through the point (2, −1).
𝑑𝑥 𝑥 2 −1

12. Solve the following differential equations

𝑑𝑦 𝑥−2𝑦 𝑑𝑦 𝑥 2 +𝑦 2 𝑑𝑦
1. = 2. = 3.2𝑥 2 = 𝑥 2 + 𝑦2
𝑑𝑥 𝑥 𝑑𝑥 𝑥𝑦 𝑑𝑥

- 365 -
 𝑑𝑦 𝑑𝑦
4. 𝑥𝑦 2 = 𝑥 3 − 𝑦 3 5.𝑥 = 𝑦 + √𝑥 2 − 𝑦 2
𝑑𝑥 𝑑𝑥
𝑑𝑦
6. 𝑦 = −𝑥 + 2𝑦
𝑑𝑥
𝑑𝑦
7. 𝑥 (𝑥 2 + 𝑦 2 ) = 𝑦 (𝑦 2 − 𝑥 2 )
𝑑𝑥
𝑑𝑦
8. (𝑥 3 + 𝑥𝑦 2 ) = (𝑥 2 𝑦 − 𝑦 3 )
𝑑𝑥

13. Solve the following differential equations

𝑑𝑦 𝑑𝑦 𝑑𝑦 𝑦 1
1. + 𝑦 = 𝑒 −𝑥 2. − 𝑥𝑦 = 0 3. + =
𝑑𝑥 𝑑𝑥 𝑑𝑥 𝑥 𝑥
𝑑𝑦 𝑑𝑦 𝑑𝑦
4. 𝑥 − 𝑦 = 𝑥 2 5. + 𝑦 =   𝑥 6. − 2𝑦 = 𝑥
𝑑𝑥 𝑑𝑥 𝑑𝑥
𝑑𝑦 𝑑𝑦
7. + 𝑦 𝑐𝑜𝑡 𝑥 = 𝑐𝑜𝑠 𝑥8. 𝑥 − 2𝑦 = 𝑥 3
𝑑𝑥 𝑑𝑥
𝑑𝑦
9.   − 𝑦 𝑡𝑎𝑛 𝑥 = 𝑐𝑜𝑠 2 𝑥
𝑑𝑥
𝑑𝑦
10. (𝑥 2 − 1) + 2𝑥𝑦 = 𝑥.
𝑑𝑥

- 366 -
 MISCELLANEOUS PROBLEMS
1- Suppose𝐴 is an 𝑚 × 𝑛 matrix such that 𝐴𝐴𝑡 = 𝟎.
Prove that 𝐴 = 𝟎.
1 2 1 2 1 0 7
1 2 4 −1 4 3 1
2- Put the matrix 𝐴 = 2 4 5 1 5 −2 3 in row-
2 4 8 −2 8 6 2
[1 2 1 2 6 0 7]
echelon form.
3- Let 𝐴 and 𝐵 be an 𝑛 × 𝑛 non-singular matrices prove
that 𝐴𝐵 = 𝐵𝐴 if and only if (𝐴𝐵 )2 = 𝐴2 𝐵2
4- Determine whether the following matrix is unitary and
find the inverse of it if it is unitary.
𝑖 1 1
1+ −
2 2 2
𝑖 1 𝑖
−
√3 √3 √3
3+𝑖 4 + 3𝑖 5𝑖
[2√15 2√15 2√15]

−1 3 5
5- Prove that [ 1 −3 −5] is idempotent.
−1 3 5

- 367 -
 6- Prove that the following system is consistent if and
only if 𝛿𝛼 − 𝛽 − 11𝛾 + 5𝛿 = 0.
Solve the system when 𝛼 = 𝛽 = −1, 𝛾 =3 and 𝛿 = 8.
𝑥1 − 𝑥2 − 𝑥4 − 5𝑥5 = 𝛼
2𝑥1 + 𝑥2 − 𝑥3 − 4𝑥4 + 𝑥5 = 𝛽
𝑥1 + 𝑥2 + 𝑥3 − 4𝑥4 − 6𝑥5 = 𝛾
𝑥1 + 4𝑥2 + 2𝑥3 − 8𝑥4 − 5𝑥4 = 𝛿
7- Write the general solution of the following systems as
a linear combination of particular solutions.
𝑥1 − 𝑥2 + 𝑥3 = 0
−2𝑥1 − 3𝑥2 = 0
𝑥1 + 2𝑥2 − 𝑥3 = 0
8. Find the minors and cofactors of the matrix
2 −3 1
𝐴 = [4 −2 5].
1 −1 3
1 𝑥 𝑥2
9. Prove that |1 𝑦 𝑦 2 | = (𝑥 − 𝑦)(𝑦 − 𝑧)(𝑧 − 𝑥 ).
1 𝑧 𝑧2
10. Use Cramer's Rule to find 𝑥1 and 𝑥2 for the system
𝑥 = 𝑥1 cos 𝜃 − 𝑦1 sin 𝜃
𝑥 = 𝑥1 sin 𝜃 + 𝑦1 cos 𝜃

- 368 -
11. Prove that 𝑃2 is a linear space.
12. Prove that 𝑉 = ℝ2 with addition and scalar
multiplication defined by: (𝑎, 𝑏) + (𝑐, 𝑏) =
(𝑎 + 𝑐, 𝑏 + 𝑑) and 𝛼 (𝑎, 𝑏) = (|𝛼 |𝑎, |𝛼 |𝑏) is not a linear
space.
13. Let 𝑉 be a linear space and 𝒖, 𝒗 ∈ 𝑉, 𝛼, 𝛽 ∈ ℝ.
(a) If 𝛼𝒗 = 𝛽𝒗, 𝒗 ≠ 𝟎, prove that 𝛼 = 𝛽.
(b) If 𝛼𝒗 = 𝛼𝒖 and ≠ 0 , prove that 𝒗 = 𝒖.
14. Let 𝑉 = {𝐴𝑋: 𝐴 ∈ 𝑀𝑚×𝑛 and X ∈ Mn×1 }. Prove that
𝑉 is a linear space with addition and scalar
multiplication of matrices
15. Let 𝑈, 𝑊 be two subspaces of V. Prove that
(a) 𝑈 ∩ 𝑊 is a subspace of 𝑉.
(b) 𝑈 ∪ 𝑊 may not be a subspace of 𝑉.
(c) 𝑈 ∪ 𝑊 is a subspace of 𝑉 if and only if
𝑈 ⊆ 𝑉 or W ⊆ V.
(d) 𝑈 + 𝑊 = {𝒖 + 𝒘: 𝒖 ∈ 𝑈, 𝒘 ∈ 𝑊 } is a subspace
of 𝑉.
16. Let 𝒖1 = (1,0,1), 𝒖2 = (−1, −1,0),
𝒗1 = (1,2,0), 𝒗2 = (0,0,1)
(a) Prove that {𝒖1 , 𝒖2 } ≠ {𝒗1 , 𝒗2 }.
- 369 -
(b) Find 𝛼1 , 𝛼2 , 𝛽1 and 𝛽2 such that
𝛼1 𝑢1 + 𝛼2 𝑢2 = 𝛽1 𝑣1 + 𝛽2 𝑣2
17. Let {𝒖, 𝒗} be linearly independent in a vector space V.
Prove that
(a) {𝒖, 𝒖 + 𝒗} is linearly independent.
(b){𝒖, 𝒗, 𝒖 + 𝒗} is linearly dependent.
18. Determine the dimension of and a basis for the
solution space of the systems:
𝑥 + 𝑦 + 𝑧 = 0, 2𝑥 + 2𝑦 − 2𝑧 = 0
4𝑥 + 3𝑦 − 𝑧 = 0, 6𝑥 + 5𝑦 + 𝑧 = 0.
19. Find the coordinate vector of v relative to the basis
𝑆 = {𝒗1 , 𝒗2 , 𝒗3 } , where 𝑣 = (2, −1, 3); 𝒗1 =
(1, 0, 0) , 𝒗2 = (2, 2, 0) , 𝒗3 = (3, 3, 3)

20. Let 𝐴 = [𝑎𝑎1 𝑎𝑎2 ] , 𝐵 = [𝑏𝑏1 𝑏𝑏2 ] ∈ 𝑀2×2 and

3 4 3 4

〈𝐴, 𝐵〉 = 𝑎1 𝑏1 + 𝑎2 𝑏2 + 𝑎3 𝑏3 + 𝑎4 𝑏4 .

Prove that 𝑀2×2 is an inner product space.

21. Let V be an inner product space and 𝟎 ≠ 𝒗∘ ∈ 𝑉.

Prove that𝑊 = {𝒖 ∈ 𝑉: 〈𝒖, 𝒗∘ 〉 = 𝟎} is subspace of V.

- 370 -
22. Let 𝒖, 𝒗 be a linearly dependent vectors in an inner
product space V. Compute the determinate
〈𝒗, 𝒗〉 〈𝒖, 𝒗〉
| |
〈𝒗, 𝒖〉 〈𝒖, 𝒖〉

23. Let V be an inner product space and 𝒗𝑖 , 𝒘 ∈ 𝑉, 𝛼𝑖 ∈

ℝ. Prove that

〈𝛼1 𝒗1 + ⋯ 𝛼𝑛 𝒗𝑛 , 𝒘〉 = 𝛼1 〈𝒗1 , 𝒘〉 + ⋯ +𝛼𝑛 〈𝒗𝑛 , 𝒘〉

24. Let 𝑥0 , 𝑥1 , … , 𝑥𝑛 ∈ ℝ, 𝑝, 𝑞 ∈ 𝑃2 and

〈𝑝, 𝑞 〉 = 𝑝(𝑥0 )𝑞(𝑥0 ) + 𝑝(𝑥1 )𝑞(𝑥1 ) + ⋯ + 𝑝(𝑥𝑛 )𝑞 (𝑥𝑛 ).

Prove that 𝑃𝑛 is an inner product space.

25. Let V be an inner product space and 𝒗1 , … , 𝒗𝑛 ∈ 𝑉

such that 𝒗𝑖 and 𝒗𝑗 are orthogonal for every 𝑖 ≠ 𝑖
(𝑖. 𝑒, 〈𝒗𝑖 , 𝒗𝑗 〉 = 0, 𝑖 ≠ 𝑗).
Prove that
‖𝒗1 + ⋯ + 𝒗𝑛 ‖2 = ‖𝒗1 ‖2 + ⋯ + ‖𝒗𝑛 ‖2 .
26. Apply Gram – Schmidt process to obtain an
orthonormal basis for the solution space for the matrix A,
where the inner product is the Euclidean.

- 371 -
 1 0
03 −4
1 0 5 −2 −1
𝐴 = [0 04 2 ] , 𝐴 = [
1 ]
0 1 −2 4 2
0 0
15 0
1 1
27. Let 𝐴 = [ ]. Find the eigenvectors of A.
−2 4
5 −2 6 −1
0 3 −8 0
28. Let 𝐴 = ( ).
0 0 5 4
0 0 0 1
Find the eigenvalues and eigenvectors of A.
100𝑛 +∞
29. Is{ } a monotone sequence? Is it a convergent
𝑛! 𝑛=1

sequence? Justify your answer.

3𝑛 +∞
30. Show that{ } is a decreasing sequence.
1+32 𝑛 𝑛=1

1 .3 .5…(2 𝑛−1) +∞
31. Show that{ } is an increasing
𝑛! 𝑛=1

sequence.
32. Show that if {𝑎𝑛 }+∞
𝑛=1 a nonincreasing sequence,

then{−𝑎𝑛 }+∞
𝑛=1 is a nondecreasing sequence.

In Exercises33 to 38, find𝐿 = lim 𝑎𝑛 if it exists.

𝑛→+∞

33. 𝑎𝑛 = (−1)𝑛 /𝑛; 34. 𝑎𝑛 = 𝑒 1/𝑛 ;

1 1
35. 𝑎𝑛 = − ; 36. 𝑎𝑛 = sin(𝜋𝑛);
√𝑛 𝑛+1

- 372 -
 (2 𝑛−1)𝜋 𝑛+1
37. 𝑎𝑛 = sin ( ); 38. 𝑎𝑛 = 𝑛(𝑛+2).
2

39.Which of the sequence{𝑎𝑛 }+∞

𝑛=1 in Exercises33 to 38,

are(a) Decreasing? (b) Nondecreasing?(c) Alternating?

40. Suppose 𝑓(𝑥) satisfies𝑓′(𝑥) > 0 and𝑓(𝑥 ) ≤ 1 −
𝑒 −𝑥 for all𝑥 ≥ 1. What can you conclude about the
convergence of{𝑎𝑛 } if𝑎𝑛 = 𝑓 (𝑛), 𝑛 = 1,2, … ?
41. Use your knowledge of geometric series and p-series
to determine all values of q for which the following
series converge.
∑∞ 𝑘 2𝑘 ∑∞ 𝑞 3
(a) 𝑘=0 𝜋 /𝑞 ; (b) 𝑘=1(1/𝑘 ) ;

∑∞ 𝑘 ∑∞ 𝑘
(c) 𝑘=2 1/(ln𝑞 ); (d) 𝑘=2 1/(ln𝑞 ) ;

42.(a) Use a suitable test to find all values of q for

𝑞
which∑∞
𝑘=2 1/[𝑘 (ln 𝑘 ) ] converges.

(b) Why can't you use the integral test for the series
∑∞ 2
𝑘=1(2 + cos 𝑘 )/𝑘 ?Test for convergence using a

test that does apply?

43.Express 1.3636… as:
(a) an infinite series in sigma-notation.
(b) a ratio of integers.

- 373 -
44.In parts (a)-(d), use the comparison test to
determine whether the series converges.
2 𝑘−1 2 𝑘+1
(a) ∑∞
𝑘=1 ; (b) ∑∞
𝑘=1 ;
3 𝑘 2 −𝑘 3 𝑘 2 +𝑘
2 𝑘−1 2 𝑘+1
(c) ∑∞
𝑘=1 2; (d) ∑∞
𝑘=1 .
3 𝑘 3 −𝑘 3 𝑘 3 +𝑘 2

45.Find the sum of the series (if it converges).

2𝑘 +3𝑘 1
(a) ∑∞
𝑘=1 ; (b) ∑∞
𝑘=2 ln (1 + 𝑘 );
6𝑘+1
−1/2
(c) ∑∞
𝑘=1[𝑘 − (𝑘 + 1)−1/2 ].
In Exercises 46-55, determine whether the series
converges or diverges. You may use the following limits
without proof:
𝑘 𝑘
lim (1 + 1/𝑘)𝑘 = 𝑒, lim √𝑘 = 1 , lim √𝑎 = 1
𝑘→+∞ 𝑘→+∞ 𝑘→+∞
−𝑘 −𝑘 2 𝑘
46. ∑∞ ∞
𝑘=0 𝑒 ; 47.∑𝑘=1 𝑘𝑒 ;48. ∑∞
𝑘=1 ;
𝑘 2 +2𝑘+7

√𝑘 𝑘 𝑘 3𝑘 𝑘!
49. ∑∞
𝑘=1 𝑘 2 +7; 50. ∞ ∞
∑𝑘=1 ( ) ;51.∑𝑘=1 ;
𝑘+1 (2 𝑘)!

∞ 𝑘 6 3𝑘 ∞ 5 𝑘 3𝑘
52 ∑𝑘=1 ; 53.∑𝑘=1 ( ) .
(𝑘+1) 2 𝑘+1

In Exercise 54-57, determine if the given series is

absolutely convergent, conditionally convergent, or
divergent.
𝑘 1/𝑘 𝑘 𝑘
57. ∑∞
𝑘=1(−1) /𝑒 ;58.∑∞
𝑘=0(−2) /(3 + 1);

- 374 -
 𝑘 𝑘 𝑘 𝑘+1
59. ∑∞ ∞
𝑘=0(−1) /(2 𝑘 + 1); 60.∑𝑘=0(−1) 3 /2 .
In Exercise (27)-(32), determine the radius of
convergence and the interval of convergence of the given
power series.
(𝑘−1)𝑘 (2 𝑥)𝑘 (1−𝑥)2𝑘
61. ∑∞
𝑘=1 ; 62. ∑∞
𝑘=1 ; 63.∑∞
𝑘=1 ;
𝑘 √𝑘 3𝑘 4𝑘𝑘

𝑘 2 (𝑥+2)𝑘 𝑘!(𝑥−1)𝑘 (2𝑘) !𝑥 𝑘

64.∑∞
𝑘=1 ; 65. ∑∞
𝑘=1 ; 66. ∑∞
𝑘=1 (2𝑘+1)! .
(𝑘+1)! 5𝑘

In Exercise 67-69, find the nth Taylor polynomial for f

about 𝑥 = 𝑎 (for the stated value of n and a );
3
67. 𝑓(𝑥 ) = 𝐼𝑛(𝑥 − 1) ; 𝑎 = 2 ; 𝑛 = 3 ; [ ,2]
2

68. 𝑓(𝑥 ) = 𝑒 𝑥/2 ; 𝑎 = 0 ; 𝑛 = 4 ; [−1,0]

4
69. 𝑓(𝑥 ) = √𝑥 ; 𝑎 = 1; 𝑛 = 2 ; [ ,1]
9

70. (a) Use the identity𝑎 − 𝑥 = 𝑎(1 − 𝑥/𝑎) to find the

Maclaurin series for 1/(𝑎 − 𝑥)from the geometric
series. What is its radius of convergence?
(b) Find the Maclaurin series and radius of
convergence of 1/(3 + 𝑥 ).
71. Use the known Maclaurin series in (1 + 𝑥) to find the
Maclaurin series and radius of convergence of ln
(𝑎 + 𝑥) (𝑎 > 0).

- 375 -
72 Use the identity 𝑥 = 𝑎 + (𝑥 − 𝑎) and the known
Maclaurin series for𝑒 𝑥 , sin 𝑥 , cos 𝑥 , and 1 / (1 − 𝑥) to
find the Taylor series about 𝑥 = 𝑎 for
(a)𝑒 𝑥 (b) sin x and (c) 1 / x.
73. Find the partial fraction decomposition of
𝑥 2 +2𝑥+5 𝑥 2 +1
; ;
𝑥 3 +3𝑥 2 −𝑥−3 𝑥(𝑥−1)3

2𝑥 3 +7𝑥+5 𝑥 4 +𝑥 3 +𝑥 2 −𝑥+1
(𝑥 2 +𝑥+2)(𝑥 2 +1)
; .
𝑥(𝑥 2 +1)2

74. Find the solution of the differential equation

𝑑𝑦
= 𝑥𝑦 ln 𝑥
𝑑𝑥

which satisfies the conditions 𝑥 = 1, 𝑦 = 1. Give ln 𝑦in

terms of x.
𝑑𝑦
75. Solve the differential equation 𝑥 + 2𝑦 =
𝑑𝑥

𝑒 𝑥 (𝑥 > 0) given that 𝑦 = 1when𝑥 = 1.

𝑑𝑦
77. Find the solution of the differential equation +
𝑑𝑥

𝑦 𝑐𝑜𝑡 𝑥 = 𝑠𝑖𝑛 𝑥 for which 𝑦 = 1 when𝑥 = 𝜋.

78. Find the solution of the differential equation
𝑑𝑦
𝑦𝑒 2𝑥 = 𝑥 which satisfies the boundary condition 𝑦 =
𝑑𝑥

2 when𝑥 = 0.
Give y in terms of x.
- 376 -
 REFERENCES

[1] Axler, Sheldon (February 26, 2004), Linear Algebra Done

Right (2nd ed.), Springer, ISBN 978-0-387-98258-8

[2] Bretscher, Otto (June 28, 2004), Linear Algebra with

Applications (3rd ed.), Prentice Hall,

[3] Garner, Lynn E. (1991), Calculus and analytic geometry,

Dellen Publishing Company,.

[4] Hazewinkel, Michiel, ed. (2001), Differential equation,

ordinary, Encyclopedia of Mathematics, Springer, ISBN 978-1-

55608-010-4

[5] Kolman, Bernard; Hill, David R. (May 3, 2007), Elementary

Linear Algebra with Applications (9th ed.), Prentice Hall,

ISBN 978-0-13-229654-0

[6] Shores, Thomas S. (December 6, 2006), Applied Linear

Algebra and Matrix Analysis, Undergraduate Texts in

Mathematics, Springer, ISBN 978-0-387-33194-2

[7] Smith, Larry (May 28, 1998), Linear Algebra, Undergraduate

Texts in Mathematics, Springer, ISBN 978-0-387-98455-1

- 377 -