III SEM ME - CH 31-301 - Unit - II

Unit - II: LINEAR ALGEBRA
Linear Algebra:
Introduction: We have studied the solution of a system of linear equation in one
variable x, the solution of a system of linear equations in two variables in x and y.
The method of solving a system of linear equations using different methods like
Graphical method, Substitution method, Elimination method and cross
multiplication method and the nature of solution of the system of linear equations.
Linear algebra comprises of matrices and applications of the system of linear
equations. In the present chapter, we discuss the rank of the matrix by row Echelon
form, the consistency and inconsistency of the system of linear equations using
matrix method. The solution of the system of linear equations using Gauss -
Elimination method. The solution of the system of linear equations using Gauss -
Seidel method. Eigen value and Eigen vector of a matrix. The largest Eigen vector
and corresponding Eigen value of a matrix using Rayleigh power method.
Diagonalization of a square matrix. The solution of the system of first order and first
degree ordinary differential equations using matrix method.
Dr. Vijaya kumar, MSRIT, Bengalur Engineering Mathematics - III [ME/CH 31] August 17, 2021 1 / 112
Continued Unit - II on Rank of the Matrix
Rank of a Matrix: If A is the given matrix, then the number of non zero rows in
the row echelon form of a given matrix A is the rank of the given matrix. The rank
of the given matrix A is denoted by ρ(A).
Row echelon form: A non zero matrix A is said to be in row echelon form if it
satisfies the following conditions.
1. All the the zero rows are below the non zero rows.
2. The first non zero entry in any non zero row is 1
Note :The row echelon form is nothing but to reducing the given matrix to an upper
triangular matrix. The row echelon form generally is of the form
 
∗∗

 ∗∗ 


 ∗∗ 

 ∗∗
below the diagonal elements are zero and leading and above the diagonal elements
are non zero or zero.
Continued Unit - II on Rank of the Matrix
Examples :
1.  
1 2 3 2

 0 1 3 2
A=
 0 0 1 2
 0 0 0 1
The given matrix A is in row echelon form and it has 4 non zero rows. The rank of
the matrix A is 4 ∴ ρ(A) = 4.
2.  
1 4 6 3
 0 1 3 7
 
B=  0 0 1 3
 0 0 0 0
The given matrix B is in row echelon form and it has 3 non zero rows. The rank of
the matrix B is 3 ∴ ρ(B) = 3.
Continued Unit - II on Rank of the Matrix ...
3.  
1 −5 6
 0 1 0
D= 
 0 0 0
The given matrix D is in row echelon form and it has 2 non zero rows. The rank of
the matrix D is 2 ∴ ρ(D) = 2.
Elementary row Transformations (OR) Elementary row operations:
We are using the following row transformations to reduce the given matrix to roe
echelon form
1 Interchanging any two rows of the matrix and it is denoted by
R1 ←→ R2 , R2 ←→ R3
2 Any row is multiple by a non zero constant k anf it is denoted by
R2 → kR1 , R3 → −kR2
3 Adding or Subtracting any row a non zero constant multiple of any other row
and it is denoted by
R2 → R2 + kR1 , R3 → R3 − kR1
Continued Unit - II Problems on Rank of the Matrix ...
Problems:
Problem - 01: Find the rank of the matrix by reducing to row echelon form.
 
1 3 −2
 2 −1 4
 
 1 −11 14 
Solution: Take  
1 3 −2
 2 −1 4
A= 
 1 −11 14 
Applying elementary row transformations reducing the given matrix to row echelon
form
R2 −→ R2 − 2R1 and R3 −→ R3 − R1
 
1 3 −2
 0 −7 8
A≈  0 −14 16 

R3 −→ R3 − 2R2  
1 3 −2
 0 −7 8
A≈ 
 0 0 0
Therefore, the given matrix A is in row echelon form and it has 2 non zero rows. The
rank of the matrix A is 2. ∴ ρ(A) = 2.
Problem - 02: Find the rank of the matrix by reducing to row echelon form.
 
2 −1 1

 3 −1 1

 4 −1 2
 −1 1 −1
2 −1 1

 3 −1 1
A=
 4 −1 2
 −1 1 −1
Applying elementary row transformations to the given matrix and reducing to row
echelon form
R2 −→ 2R2 − 3R1 ,
R3 −→ R3 − 2R1
R4 −→ 2R4 + R1  
2 −1 1
 0
 1 −1
A≈  0 1 0 
 0 1 −1
R3 −→ R3 − R2
R4 −→ R4 − R2  
2 −1 1

 0 1 −1
A≈
 0 0 1 
 0 0 0
Therefore, the given matrix A is reduced to row echelon form. The number of non
zero rows are 3 and hence the rank of the given matrix A is 3. ∴ ρ(A) = 3.
Continued Unit - II Problems on Rank of a Matrix ...
Problem - 03: Find the rank of the matrix by reducing to row echelon form
 
3 −1 2
 −6 2 4
 
 −3 1 2
3 −1 2
 −6 2 4
A= 
 −3 1 2
Reducing the given matrix to row echelon form by applying elementary row
transformations
R2 −→ R2 + 2R1
R3 −→ R3 + R1  
3 −1 2
 0 0 8
A≈  0

0 4
R3 −→ 2R3 − R2  
3 −1 2
 0 0 8
A≈ 
 0 0 0
Now, it is in row echelon form and has two non zero rows and hence the rank of the
given matrix A is 2.
∴ ρ(A) = 2.
 
0 2 3 4
 2 3 5 4
 
 4 8 13 12
0 2 3 4
 2 3 5 4
B= 
 4 8 13 12
transformations
R1 ←→ R2  
2 3 5 4
 0 2 3 4
B≈  4

8 13 12
R3 −→ R3 − 2R1  
2 3 5 4
 0 2 3 4
B≈ 
 0 2 3 4
R3 −→ R3 − R2  
2 3 5 4
 0 2 3 4
B≈ 
 0 0 0 0
given matrix B is 2.
∴ ρ(B) = 2.
 
0 1 −3 −1
 1 0 1 1 
 
 3 1 0 2 
 
 1 1 −2 0
0 1 −3 −1

 1 0 1 1 
A=
 3 1 0 2 
 1 1 −2 0
transformations
R1 ←→ R2  
1 0 1 1
 0 1 −3 −1
 
A≈  3 1 0 2 

 1 1 −2 0
R3 −→ R3 − 3R1
R4 −→ R4 − R1  
1 0 1 1

 0 1 −3 −1

A≈
 0 1 −3 −1

 0 1 −3 −1
R3 −→ R3 − R2
R4 −→ R4 − R2  
1 0 1 1

 0 1 −3 −1

A≈
 0 0 0 0 

 0 0 0 0 
given matrix A is 2.
∴ ρ(A) = 2.
Problems: Find the rank of the following matrices, by elementary row
transformations.
 
  1 2 3 0    
1 4 5 1 2 3 2 1 1 1 −3
 2 4 3 2
 2 6 8    2 3 5 1  3 1 −2 −2
(1).   (2).  3 2 1 3 (3). 
 (4).  
 3 7 22   1 3 4 5   2 4 7 7 
 6 8 7 5
   
1 3 2 3 4 −1 −6  
1 2 3 4
 2 −1 3  2 3 2 −3
     3 4 1 2
(5).  3 −5 4 (6).  2 1 −14 −9
 (7). 
 
   4 3 1 2
 1 17 4  1 3 13 3
 
2 −2 0 6

 4 2 0 2

(8). 
 1 −1 0 3

 1 −2 −1 2
Continued Unit - II on Consistent of a system of Linear Equations
Consistency of a system of linear equations:

A system of equations are said to be linear if in which all unknown quantities (or
variables) x1 , x2 , x3 , ..., xn are in first degree only.
Consider a set of m system of linear equations in n unknowns quantities (variables)
x1 , x2 , x3 , ..., xn as follows.
a11 x1 + a12 x2 + ... + a1n xn = b1
a21 x1 + a22 x2 + ... + a2n xn = b2

...
...
am1 x1 + am2 x2 + ..... + amn xn = bm
where amn ’s and b1 , b2 , ...bm are constants. If the RHS quantities b1 , b2 , ...bm all are
equal to zero then the above system of linear equations are said to be a
Homogeneous system of a linear equations. Suppose the RHS quantities
b1 , b2 , ...bm all are not equal to zero then the above system of linear equation are said
to be Non homogeneous system of a linear equations.
Continued Unit - II on Consistent of a system of Linear Equations ...
Solution : The set unknown quantities x1 , x2 , x3 , ..., xn are satisfies the system of
linear equations is called as a solution of the system of linear equations.
Consistency: A system of a linear equations are said to be consistency if it has a
solution. A system of linear equations is consistent and has an unique (OR) an
infinitely many solutions.
Inconsistency : A system of a linear equations are said to be Inconsistency if it
has a no solution.
Trivial Solution : If all the unknown quantities x1 , x2 , x3 , ...., xn are all zero then
the solution is called as trivial solution and it is a solutions for a homogeneous
system of linear equations.
Non trivial solution : If at least one unknown quantity is not equal to zero is
called as non trivial solution.
Therefore, the above system of linear equations are representing in matrix equation
of the form
AX = B
Where A is called the coefficient matrix, X is a column matrix of all the unknown
quantities x1 , x2 , x3 , ..., xn and B is a column matrix of the RHS values of the above
system of linear equations
    
a11 a12 a13 ... a1n x1 b1

 a21 a22 a23 ... a2n 

 x2  
  b2 

 a31 a32 a33 ... a3n  

 x3  

 b3 

 .. .. .. ... .. 

 .. 
=
 .. 


 .. .. .. ... .. 

 ..  
  .. 

 am1 am2 am3 ... amn 
 xn   bm 
and the augmented matrix of the above system of linear equations is of the form
[A : B] and it is given by
 
a11 a12 a13 ... a1n : b1
 a21 a22 a23 ... a2n : b2 
 
 a31 a32 a33 ... a3n : b3 
 
[A : B] = 
 .. .. .. ... .. : .. 

 .. .. .. ... .. : .. 
 am1 am2 am3 ... amn : bm 
Note : The system of linear equations AX = B may or may not possess a solution.
Suppose the system of linear equations possesses a solution it may or may not be the
only solution. The concept of the rank of the matrix is useful us to conclude the
system of linear equation is consistent or not and the system of linear equations has
an unique solution or an infinitely many solutions.
Conditions for Consistent and Inconsistent :
Consistent: A system of linear equation are said to be consistent if ρ(A) = ρ(A : B)
and it has two types of solutions, they are,
(i) Consistent with an unique solution: A system of linear equation are said to
be consistent and has an unique solution if ρ(A) = ρ(A : B) = r = n.
where r is the rank and n is the number of unknowns in the given system of linear
equations.
(ii) Consistent with an infinitely many solutions: A system of linear equation
are said to be consistent and has an infinitely many solutions if
ρ(A) = ρ(A : B) = r < n. here we have assume that (n − r) unknowns are arbitrary
constants.
Inconsistent or No solution: A system of linear equation are said to be
inconsistent or no solution if ρ(A) 6= ρ(A : B)
Geometrical Representation of a system of the linear equations for

consistent and Inconsistent:
Consistent and has an unique solution: A system of the linear equations in two
variables x and y are consistent and has an unique solution then geometrical
represents that the straight lines are intersects at one point in the xy plane.
Example: Consider a system of the linear equations in two variables x and y are
x + 2y = 3, 4x + 5y = 6, the point of intersection is (−1, 2).
Consistent and has an infinitely many solutions: A system of the linear
equations in two variables x and y are consistent and has an infinitely many solutions
then geometrical represents that the straight line are intersects at many points or
coincide each other in the xy plane.
x + 2y = 3, 4x + 8y = 12, the points of intersection are more than one point.
Inconsistent or No solution: A system of the linear equations in two variables x
and y are inconsistent or no solution then geometrical represents that the straight
lines are not intersects at any point or parallel to each other in the xy plane.
x + 2y = 3, 4x + 8y = 6, there is no point of intersection.
Note - 01: Homogeneous system of linear equations is always consistent because it

has at least one trivial solution.
Note - 02: A homogeneous system of linear equations has the non - trivial solution
if and only if the determinant of the coefficient matrix of the homogeneous system of
linear equations is zero. that is |A| = 0
Note - 03: Non homogeneous system of a linear equations may or may not be
consistent.
Working procedure:
Step - 01: Write the given system of linear equations in matrix equation AX = B.
Step - 02: Write the augmented matrix of the form [A : B] .
Step - 03: Applying elementary row transformations and reduce the augmented
matrix [A : B] to row echelon form.
Step - 04: Find the rank of the coefficient matrix A and augmented matrix [A : B].
Verify the conditions for consistent and inconsistent of the system of equations.
Step - 05: If it is consistent solve the unknown quantities of the given system of
linear equations.
Note: If the given system of linear equations are consistent and has infinitely many
solutions, we have to assume that (n − r) unknowns are arbitrary constants.
Continued Unit - II Problems on Consistent of a system of Linear
Equations ...
Problems:
Problem - 01: Find the value of k for which the system of of linear equation
(3k − 8)x + 3y + 3z = 0, 3x + (3k − 8)y + 3z = 0, 3x + 3y + (3k − 8)z = 0 has a non -
trivial solution.
Solution: Given the homogeneous system of linear equations are
(3k − 8)x + 3y + 3z = 0, 3x + (3k − 8)y + 3z = 0, 3x + 3y + (3k − 8)z = 0
We know that, the homogeneous system of linear equations has non trivial solution if
and if the determinant of the coefficient matrix is equal to zero, i.e. |A| = 0
|A| = 0

3k − 8 3 3
3 3k − 8 3
=0

3 3 3k − 8

Now, expanding the determinant, we get
[(3k − 8)(3k − 8)2 − 9] − 3[3(3k − 8) − 9] + 3[9 − 3(3k − 8)] = 0
(3k − 8)(3k − 11)(3k − 5) − 18(3k − 11) = 0
(3k − 11)2 (3k − 2) = 0
Equations ...
From this we get, k = 11

3
or k = 23
Hence the given homogeneous system of linear equations has a non - trivial solution
for k = 11
3
or k = 23 .
Problem - 02: Test for consistency and solve the system of linear equation
x + y + z = 4, 2x + y − z = 1, x − y + 2z = 2 .
Solution: Given the system of linear equations
x + y + z = 4, 2x + y − z = 1, x − y + 2z = 2
The matrix equation of the given system of linear equations is
AX = B
    
1 1 1 x 4

 2 1 −1 
 y 
= 1

 1 −1 2  z  2
Equations ...
The augmented matrix of the given system of linear equations is
 
1 1 1 :4
 2 1 −1 : 1
[A : B] =  
 1 −1 2 : 2
Reducing the augmented matrix [A : B] to row echelon form by applying the

elementary row transformations.
R2 −→ R2 − 2R1
R3 −→ R3 − R1  
1 1 1 :4
 0 −1 −3 : −7
[A : B] ≈ 
 0 −2

1 : −2
R3 −→ R3 − 2R2  
1 1 1 :4
 0 −1 −3 : −7 
[A : B] ≈  
 0 0 7 : −12
Therefore, the augmented matrix [A; B] is in row echelon and ρ(A) = 3, ρ(A : B) = 3
∴ ρ(A : B) = ρ(A) = 3 = r = (n = 3). Hence the given system of a linear equations
are consistent and has an unique solution.
x+y+z =4 (1)
−y − 3z = −7 (2)
7z = −12 (3)
Now, solve by back substitution method, from the Eqn. (3) z = −127
, from Eqn. (2)
−y = −7 + 3z, y = −857
and from the Eqn. x = 4 − y − z, x = 125
7
Hence, x = 125
7
, −85
7
and z = −12
7
are the unique solution for the given system of a
linear equations.
Continued Unit - II Problems on Consistent of a system of a linear
equations ...
Problem - 02: Test consistency and solve 5x + 3y + 7z = 4, 3x + 26y + 2z = 9,

7x + 2y + 10z = 5
5x + 3y + 7z = 4, 3x + 26y + 2z = 9, 7x + 2y + 10z = 5
The given system of a linear equations are in matrix equation
AX = B
    
5 3 7 x 4

 3 26 2   y = 
   9

 7 2 10  z   5
The augmented matrix of the given system of linear equations [A : B] is
Continued Unit - II Problems on Consistent of a system of a linear
equations ...
 
5 3 7 :4
 3 26 2 : 9
[A : B] =  
 7 2 10 : 5

R2 −→ 5R2 − 3R1
R3 −→ 5R3 − 7R1  
5 3 7 :4
 0 121 −11 : 33 
[A : B] ≈ 
 0 −11

1 : −3
R3 −→ 11R3 + R2  
5 3 7 :4
 0 121 −11 : 33
[A : B] ≈  
 0 0 0 :0
The augmented matrix [A : B] is in row echelon form, ρ(A) = 2 and ρ(A : B) = 2.

Equations ...
∴ ρ(A : B) = ρ(A) = 2 and number of unknowns in the given system of a linear
equations are n = 3.
Therefore, the given system of a linear equations are consistent and has an infinitely
many solutions. Let us assume that n − r = 3 − 2 = 1 unknown quantities are an
arbitrary constant k. Let z = k
5x + 3y + 7z = 4 (4)
121y − 11z = 33 (5)
Now, solving equations (4) and (5)
From Eqn. (5)
121y = 33 + 11k
dividing by 11
11y = 3 + k
3 k
→ y = 11 + 11
3 K
From Eqn, (4) 5x + 3( 11 + 11 ) + 7k = 4
35−80k
→ x = 55
∴ x = 35−80k
55
, y = 3+k
11
and z = k are the infinitely many solutions for the given
system of a linear equations.
Equations ...
Problem - 03: Show that the system of a linear equations x + y + z = −3,
3x + y − 2z = −2, 2x + 4y + 7z = 7 is not consistent.
x + y + z = −3, 3x + y − 2z = −2, 2x + 4y + 7z = 7
AX = B
    
1 1 1 x −3

 3 1 −2  y = 
   −2

 2 4 7   z  7

 
1 1 1 : −3
 3 1 −2 : −2
[A : B] = 
 
2 4 7 :7 
Equations ...

R2 −→ R2 − 3R1
R3 −→ R3 − 2R1  
1 1 1 : −3
 0 −2 −5 :7 
[A : B] ≈ 
 0

2 5 : 13 
R3 −→ R3 + R2  
1 1 1 : −3
 0 −2 −5 :7 
[A : B] ≈  
 0 0 0 : 20 
The augmented matrix [A : B] is in row echelon form, ρ(A) = 2 and ρ(A : B) = 3.

∴ ρ(A) 6= ρ(A : B), the given system of a linear equations has no solution and hence
it not consistent.
Equations ...
Problem - 04: Find the value of λ for which the system of a linear equations
x + y + z = 1, x + 2y + 4z = λ, x + 4y + 10z = λ2 has a solution and solve it in each
case.
Solution: Given the system of a linear equations
x + y + z = 1, x + 2y + 4z = λ, x + 4y + 10z = λ2
AX = B
    
1 1 1 x 4

 1 2 4  y = 
   λ 
 1 4 10  z   λ2 

 
1 1 1 :1
 1 2 4 :λ
[A : B] =  
 1 4 10 : λ2 
Equations ...

R2 −→ R2 − R1
R3 −→ R3 − R1  
1 1 1 :1
 0 1 3 :λ−1
[A : B] ≈ 
 0 3 9 : λ 2 − 1

R3 −→ R3 − 3R2  
1 1 1 :1
 0 1 3 :λ−1 
[A : B] ≈  2

 0 0 0 : λ − 3λ + 2
The augmented matrix [A : B] is in row echelon form, ρ(A) = 2 and ρ(A : B) = 3,

but the given system of a linear equations has a solution if ρ(A : B) = 2 only when
λ2 − 3λ + 2 = 0. Thus, the given system of a linear equations is consistent for λ = 1
or λ = 2
Equations ...
Case - 01: When λ = 1 and the given system of a linear equations reduces to
x+y+z =1
y + 3z = 0
From these equations we get, z = k1 , y = −3k1 and x = 1 − 2k1 are the infinitely
many solutions for the given system of a linear equations. Where k1 be a arbitrary
constant.
Case - 02: When λ = 2 and the given system of a linear equations reduces to
x+y+z =1
y + 3z = 1
From these equations we get, z = k2 , y = 1 − 3k2 and x = 2k2 are the infinitely many
solutions for the given system of a linear equations. Where k2 be a arbitrary
constant.
Equations ...
Problem - 05: Find the values of λ and µ for which the system of a linear equations
x + y + z = 6, x + 2y + 3z = 10, x + 2y + λz = µ has (i) a unique solution, (ii) an
infinitely solution and (iii) no solution.
Solution: Given the system of a linear equations
x + y + z = 6, x + 2y + 3z = 10, x + 2y + λz = µ
AX = B
    
1 1 1 x 6

 1 2 3  y = 
   10
 1 2 λ  z   µ

 
1 1 1 :6
 1 2 3 : 10
[A : B] = 
 1

2 λ : µ
Equations ...

R2 −→ R2 − R1
R3 −→ R3 − R1  
1 1 1 :6
 0 1 2 :4 
[A : B] ≈ 
 
0 1 λ−1 : µ − 6
R3 −→ R3 − R2  
1 1 1 :6
 0 1 2 :4 
[A : B] ≈  
 0 0 λ−3 : µ − 10
From this, we observed that the rank of the coefficient matrix A depends on the
value of λ and that the rank of augmented matrix [A : B] depends on the values of
both λ and µ.
Equations ...
(i) An unique solution: The given system of a linear equations is consistent and
has an unique solution if ρ(A) = ρ(A : B) = r = n. This is possible only when λ 6= 3,
any value of µ.
(ii) An infinitely many solutions: The given system of a linear equations is
consistent and has an infinitely many solutions if ρ(A) = ρ(A : B) = r < n. This is
possible only when λ = 3 and µ = 10.
(iii) No solution: The given system of a linear equations is inconsistent or no
solution if ρ(A) 6= ρ(A : B). This is possible only when λ = 3 and µ 6= 10.
Continued Unit - II Problems on Gauss - Elimination Method to solve a
system of a linear equations ...
Gauss - Elimination Method:
By using Gauss - elimination method, we reducing the augmented matrix [A : B] to
an upper triangular matrix by applying elementary row transformations and solve
the unknown quantities using back substitution method.
Problem - 01: Solve the system of a linear equations by Gauss - elimination method
2x1 + x2 + 4x3 = 12
4x1 + 11x2 − x3 = 33
8x1 − 3x2 + 20x3 = 20
Solution: The augmented matrix for the given system of linear equations is
 
2 1 4 : 12
 4 11 −1 : 33
[A : B] = 
 8 −3 20 : 20


R2 −→ R2 − 2R1
R3 −→ R3 − 4R1
Continued Unit - II Problems on Gauss - Elimination Method ...
 
2 1 4 : 12
 0 9 −9 :9 
[A : B] ≈  
 0 −7 4 : −28
R2 −→ 91 R2
 
2 1 4 : 12
 0 1 −1 :1 
[A : B] ≈  
 0 −7 4 : −28
R3 −→ R3 + 7R2  
2 1 4 : 12
 0 1 −1 :1 
[A : B] ≈  
 0 0 −3 : −21
Now, converted to a system of linear equations

2x1 + x2 + 4x3 = 12 (6)
x2 − x3 = 1 (7)
−3x3 = −21 (8)
solve by back substitution method
From Eqn. (8), x3 = 7, Eqn. (7) x2 − 7 = 1, x2 = 8 and Eqn. (6)
2x1 + 8 + 4(7) = 12, x1 = −12
∴ x1 = −12, x2 = 8 and x3 = 7 are the solutions for the given system of a linear
equations.
Problem - 02: Solve the system of a linear equations by Gauss - elimination method
2x + 5y + 7z = 52
2x + y − z = 0
x+y+z =9
Solution: The augmented matrix for the given system of linear equations is
 
2 5 7 : 52
 2 1 −1 : 0 
[A : B] = 
 1 1

1 :9

R2 −→ R2 − R1
R3 −→ 2R3 − R1
 
2 5 7 : 52
 0 −4 −8 : −52
[A : B] ≈  
 0 −3 −5 : −43
−1
R2 −→ 4
R2
 
2 5 7 : 52
 0 1 2 : 13 
[A : B] ≈  
 0 −3 −5 : −43
R3 −→ R3 + 3R2  
2 5 7 : 52
 0 1 2 : 13 
[A : B] ≈  
 0 0 1 : −4
Now, converted to a system of linear equations

2x + 5y + 7z = 52 (9)
y + z = 13 (10)
z = −4 (11)
solve by back substitution method

From Eqn. (11), z = −4, Eqn. (10) y + 4 = 13, y = 9 and Eqn. (9) x + 9 − 4 = 52,
x = 47
∴ x = 47, y = 9 and z = −4 are the solutions for the given system of a linear
equations.
Problems: Solve the following system of linear equations using Gauss - elimination
method.
(1)
2x + 5y + 7z = 52, 2x + y − z = 0, x + y + z = 9
(2)
2x1 − x2 + 3x3 = 1, −3x1 + 4x2 − 5x3 = 0, x1 + 3x2 − 6x3 = 0
(3)
x + 2y − z = 3, 3x − y + 2z = 1, 2x − 2y + 3z = 2
Continued Unit - II on Gauss - Seidel Iterative Method to solve a
system of a linear equations ...
Gauss - Seidel Iterative Method:
This is a numerically an iterative method to find an approximate solution of the
system of a linear equations. This method is applicable when the system of a linear
equations are in diagonally dominant system.
Diagonally dominant system of a linear equations: A system of a linear
equations are said to be in diagonally dominant system if the numerically large
coefficients are along the leading diagonal of the coefficient matrix.
Consider the system of equations as follows
a11 x + a12 y + a13 z = b1 (12)

a21 x + a22 y + a23 z = b2 (13)
a31 x + a32 y + a33 z = b3 (14)
The above system of a linear equations is solvable if it is in diagonally dominant
form, that is
|a11 | > |a12 | + |a13 |
|a22 | > |a21 | + |a23 |
|a33 | > |a31 | + |a32 |
Continued Unit - II Problems on Gauss - Seidel Iterative Method ...
Problems:
Problem - 01: Use Gauss - Seidel iteration method to solve the system of a linear
equations.
10x + y + z = 12, 2x + 10y + z = 13, 2x + 2y + 10z = 14
Solution: By data the system of a linear equations are
10x + y + z = 12
2x + 10y + z = 13
2x + 2y + 10z = 14
We can observe that the given system of a linear equations, it is in diagonally
dominant form and we can solve x, y and z as follows
1
x= (12 − y − z) (15)
10
1
y= (13 − 2x − z) (16)
10
1
z= (14 − 2x − 2y) (17)
10
Let us choose an initial approximations are x0 = 0, y 0 = 0 and z 0 = 0.

I - iteration:
Substitute y = y (0) = 0 and z = z (0) = 0 in Eqn. (15)
1 12
x(1) = (12 − 0 − 0) = = 1.2
10 10
Substitute x = x(1) 1.2 and z = z (0) = 0 in Eqn. (16)
1 10.6
y (1) = (13 − 2(1.2) − 0) = = 1.06
10 10
Substitute x = x(1) = 1.2 and y = y (1) = 1.06 in Eqn. (17)
1
z (1) = (14 − 2(1.2) − 2(1.06)) = 0.948
10
II - iteration:
Substitute y = y (1) = 1.06 and z = z (1) = 0.948 in Eqn. (15)
1 9.992
x(2) = (12 − 1.06 − 0.948) = = 0.9992
10 10
Substitute x = x(2) = 0.9992 and z = z (1) = 0.948 in Eqn. (16)
1 10.0536
y (2 = (13 − 2(0.9992) − 0.948) = = 1.00536
10 10
1
z (2) = (14 − 2(0.9992) − 2(1.0536)) = 0.98944
10
III - iteration:
1
x(3) = (12 − 1.00536 − 0.98944) = 1.00052
10
1
y (3) = (13 − 2(1.00052) − 0.98944) = 1.000952
10
1
z (3) = (14 − 2(1.00052) − 2(1.000952)) = 0.9997
10
Continued Unit - II Problems on Gauss - Seidel Method ...
IV - iteration:
1
x(4) = (12 − 1.000952 − 0.9997) = 0.999934
10
1
y (4) = (13 − 2(0.999934) − 0.9997) = 1.0000432
10
1
z (4) = (14 − 2(0.999934) − 2(1.000043) = 1.000000456
10
Hence, x = 0.999934 ≈ 1, y = 1.0000432 ≈ 1 and z = 1.00000456 ≈ 1 are the
solutions for the given system of a linear equations.
Problem - 02: Apply Gauss - Seidel iteration method to solve the system of a
linear equations.
−x−y−2z+10u = −9, −2x+10y−z−u = 15, −x−y+10z−2u = 27, 10x−2y−z−u = 3
−x − y − 2z + 10u = −9
−2x + 10y − z − u = 15
−x − y + 10z − 2u = 27
10x − 2y − z − u = 3
We can observe that the given system of a linear equations is not in diagonally
dominant form and we can rearrange it to reach the diagonally dominant form, that is
10x − 2y − z − u = 3
−2x + 10y − z − u = 15
−x − y + 10z − 2u = 27
−x − y − 2z + 10u = −9
Now, from these system of a linear equations, we can solve x, y, z and u, we get
1
x= (3 + 2y + z + u) (18)
10
1
y= (15 + 2x + z + u) (19)
10
1
z= (27 + x + y + 2u) (20)
10
1
u= (−9 + x + y + 2z) (21)
10
Taking the initial approximations are x(0) = 0, y (0) = 0, z (0) = 0 and u(0) = 0.
Iteration - I:
Substitute y = y (0) = 0, z = z (0) = 0 and u = u(0) = 0 in Eqn. (18)
1
x(1) = (3 + 2(0) + (0) + (0)) = 0.3
10
Substitute x = x(1) = 0.3, z = z (0) = 0 and u = u(0) = 0 in Eqn. (19)
1
y (1) = (15 + 2(0.3) + (0) + (0)) = 1.56
10
Substitute x = x(1) = 0.3, y = y (1) = 1.56 and u = u(0) = 0 in Eqn. (20)

1
z (1) = (27 + (0.3) + (1.56) + 2(0)) = 2.886
10
Substitute x = x(1) = 0.3, y = y (1) = 1.56 and z = z (1) = 2.886 in Eqn. (21)
1
u(1) = (−9 + (0.3) + (1.56) + 2(2.884)) = −0.1368
10
Iteration - II:
Substitute y = y (1) = 1.56, z = z (1) = 2.886 and u = u(1) = −0.1386 in Eqn. (18)
1
x(2) = (3 + 2(1.56) + (2.886) + (−0.1368)) = 0.88692
10
Substitute x = x(2) = 0.88692, z = z (1) = 2.886 and u = u(1) = −0.1368 in Eqn. (19)
1
y (2) = (15 + 2(0.88692) + (2.886) + (−0.1368)) = 1.952304
10
Substitute x = x(2) = 0.88692, y = y (2) = 1.952304 and u = u(1) = −0.1368 in Eqn.

(20)
1
z (2) = (27 + (0.88692) + (1.952304) + 2(−0.1368)) = 2.95656
10
Substitute x = x(2) = 0.88692, y = y (2) = 1.952304 and z = z (2) = 2.95656 in Eqn.
(21)
1
u(2) = (−9 + (0.88692) + (1.952304) + 2(2.95656)) = −0.02476
10
Iteration - III:
Substitute y = y (2) = 1.952304, z = z (2) = 2.95656 and u = u(2) = −0.02476 in Eqn.
(18)
1
x(3) = (3 + 2(1.952304) + (2.95656) + (−0.02476)) = 0.9836
10
Substitute x = x(3) = 0.9836, z = z (2) = 2.95656 and u = u(2) = −0.02476 in Eqn.

(19)
1
y (3) = (15 + 2(0.9836) + (2.95656) + (−0.02476)) = 1.9899
10
Substitute x = x(3) = 0.9836, y = y (3) = 1.9899 and u = u(2) = −0.02476 in Eqn.
(20)
1
z (2) = (27 + (0.9836) + (1.9899) + 2(−0.02476)) = 2.99239
10
Substitute x = x(3) = 0.9836, y = y (3) = 1.9899 and z = z (3) = 2.99239 in Eqn. (21)
1
u(2) = (−9 + (0.9836) + (1.9899) + 2(2.99239)) = −0.004172
10
Proceeding in this way, we get
Iteration - IV:
x(4) = 0.9968, y (4) = 1.9982, z (4) = 2.9987, u(4) = −0.00038
Iteration - V:
x(5) = 0.9994, y (5) = 1.9997, z (5) = 2.9998, u(5) = −0.0001

Iteration - VI:
x(6) = 0.9999, y (6) = 1.9999, z (6) = 3.0000, u(6) = 0.0000

Hence, x = 1, y = 2, z = 3 and u = 0 are the solutions of the given system of a linear
equations.
Problem - 03: Apply Gauss - Seidel iteration method to solve the system of a
linear equations
x + y + 54z = 110, 27x + 6y − z = 85, 6x + 15y + 2z = 72
Perform three iterations.

x + y + 54z = 110
27x + 6y − z = 85
6x + 15y + 2z = 72
We can observe that the given system of a linear equations are not in diagonally
dominant. Therefore, rearrange the given system of a linear equations as
27x + 6y − z = 85
6x + 15y + 2z = 72
x + y + 54z = 110
Now, solve x, y and z we get
1
x= (85 − 6y + z) (22)
27
1
y= (72 − 6x − 2z) (23)
15
1
z= (110 − x − y) (24)
54
By taking an initial approximations are x0 = 0, y 0 = 0 and z 0 = 0.
I - iteration:
Substitute y = y (0) = 0 and z = z (0) = 0 in Eqn. (22)
1 85
x(1) = (85 − 0 − 0) = = 3.1481
27 27
Substitute x = x(1) = 3.1481 and z = z (0) = 0 in Eqn. (23)
1
y (1) = (72 − 6(3.1481) − 0) = 3.54076
15
1
z (1) = (110 − 3.1481 − 3.54076) = 1.9131
54
II - iteration:
1
x(2) = (85 − (6X3.54076) + 1.9131) = 2.4321
27
1
y (2) = (72 − 6(2.4321) − 1.9131) = 3.69962
15
1
z (1) = (110 − 2.4321 − 3.69962) = 1.9234
54
III - iteration:
1
x(3) = (85 − (6X3.69962) + 1.9234) = 2.3972
27
1
y (3) = (72 − 6(2.3972) − 1.9234) = 3.7128
15
1
z (3) = (110 − 2.3972 − 3.7128) = 1.9238
54
Hence, x = 2.3972, y = 3.7128, z = 1.9238 are the solutions of the given system of a
linear equations.
Problems: Apply Gauss - Seidel iteration method to solve the following systems of
a linear equations
(1)
20x + y − 2z = 17, 3x + 20y − z = −18, 2x − 3y + 20z = 25
(2)
10x + 2y + z = 9, 2x + 20y − 2z = −44, −2x + 3y + 10z = 22
(3)
5x + 2y + z = 12, x + 4y + 2z = 15, x + 2y + 5z = 20
(4)
13x+5y−3z+u = 18, 2x+12y+z−4u = 13, 3x−4y+10z+u = 29, 2x+y−3z+9u = 31
Continued Uni - II on Eigen values and Eigen vector of a Matrix ...
Definition of a Eigen value and Eigen vector of the matrix:
Given a square matrix A, if there exists a scalar λ and a non zero column matrix X
such that AX = λ, then λ is called Eigen value of a matrix A and X is called an
Eigen vector of a matrix A corresponding to an Eigen value λ.
Example: Consider a square matrix
 
1 0 −1
 1 2 1
A=  2 2

3
and  
1
 −1
X= 
 −2
then       
1 0 −1 1 3 1
 1 2 1  −1  −3  −1
AX =   =  = 3  = λX
 2 2 3  −2  −6  −2
Continued Unit - II on Eigen values and Eigen vector of a Matrix ...
Therefore, λ = 3 is called an eigen value of a matrix A and X = [1, −1, −2]T is called
an eigen vector of a matrix A corresponding to an eigen value λ = 3.
Let I be the unit matrix and X = IX, then we can write AX = λX as,
AX = λIX
→ AX − λIX = 0
→ (A − λI)X = 0
Where,  
a1 b1 c1
 a2 b2 c2 
A=  a3 b3 c3 

and  
x1
 x2 
X= 
 x3 
and  
λ 0 0
 0 λ 0
λI =  
 0 0 λ
continued Unit - II on Eigen values and Eigen vector of a Matrix ...
Therefore, (A − λI)X = 0 becomes in the matrix form

    
a1 − λ b1 c1 x1 0
 a2 b2 − λ c2   x2   0
[A − λI]X =   = 
 a3 b3 c 3 − λ  x3   0
(a1 − λ)x1 + b1 x2 + c1 x3 = 0
a2 x1 + (b2 − λ)x2 + c2 x3 = 0
a3 x1 + b3 x2 + (c3 − λ)x3 = 0
Therefore, this homogeneous system of linear equations has a non trivial solution
exists if the determinant of the co efficient matrix is equal to zero.

a1 − λ b1 c1

a2 b2 − λ c2
=0

a3 b 3 c 3 − λ

∴ |A − λI| = 0 is called as a characteristics equation.
On expanding this determinant, we get a polynomial equation in λ of degree 3 is

called characteristics equation. The roots of this characteristics equation are called as
characteristics roots or Eigen roots or Eigen values of the matrix A. For each eigen
value λ, there will be an eigen vector X 6= 0, is called characteristic vector or eigen
vector of the matrix A.
Continued Unit - II on Properties of an Eigen values and Eigen vector
of a Matrix ...
Property - 01: If λ be an eigen value of a square matrix A with X as a
corresponding eigen vector, then λ1 is an eigen value of A−1 .
Proof: We know that, by definition of an eigen value and eigen vector of a square
matrix A is
AX = λX (25)
Pre - multiplying by A−1 on both sides, we get
A−1 (AX) = A−1 (λX)
(A−1 A)X = (λ)(A−1 X)

But, A−1 A = I and IX = X
(I)X = (λ)(A−1 X)
IX = (λ)(A−1 X)
X = (λ)(A−1 X)
1
A−1 X = ( )X (26)
λ
of a Matrix ...
Comparing the Eqns. (14) and (15) and hence we can conclude that, λ1 is a eigen
value of a A−1 .
Property - 02: If λ be an eigen value of a square matrix A with X as a
corresponding eigen vector, then λ2 is an eigen value of A2 and hence λm is an eigen
value of Am , where m is a positive integer.
Proof: We know that, by definition of an eigen value and eigen vector of a square
matrix A is
AX = λX (27)
Multiplying by A on both sides, we get
A(AX) = A(λX)
(A2 )X = (λ)(AX)
But AX = λX
(A2 )X = (λ)(λX)
A2 X = λ2 X (28)
2
Comparing the Eqns. (16) and (17) and hence we can conclude that, λ is a eigen
value of a A2 .
of a Matrix ...
Again, multiplying by A on both sides, we get
A(A2 X) = (A)λ2 X
A3 X = λ2 (AX)
But AX = λX
A3 X = λ2 (λX)
A3 X = λ3 X
Therefore, λ3 is a eigen value of a A3 . Similarly, continue like this m times, we get
Am X = λm X
Hence, λm is a eigen value of a Am .
of a Matrix ...
Property - 03: If A is an upper triangular matrix with X as a corresponding eigen
vector then the eigen valuee of A are the just the diagonal elements of the matrix A.
Proof: Consider an upper triangular matrix A of the form
 
a1 b1 c1
 0 b2 c 2 
A=  0

0 c3 
The characteristics equation of a matrix A is

|A − λX| = 0

(a1 − λ) b1 c1
0 (b2 − λ) c2
=0

0 0 (c3 − λ)

Now, expanding the above determinant we get
(a1 − λ)[(b2 − λ)(c3 − λ) − 0] − b1 [0 − 0] + c1 [0 − 0] = 0
(a1 − λ)(b2 − λ)(c3 − λ) = 0
of a Matrix ...
Therefore,
λ = a1 , λ = b2 , λ = c3
a1 , b2 and c3 are the diagonal elements of the matrix A and it is a triangular matrix.
Hence the eigen values of the triangular matrix are just the diagonal elements of the
matrix.
Property - 04: The sum of the eigen values of the matrix A is equal to the trace of
the matrix A.
Note: The trace of the matrix A is equal to the sum of the diagonal elements of the
matrix A. ∴ Trace(A) = Sum of the diagonal elements of the matrix A.
Property - 05: The product of the eigen values of the matrix A is equal to the
determinant of the matrix A.
∴ Product of the eigen values of the matrix A = |A|.
Working procedure to find an eigen values and eigen vectors of the given
matrix:
Step - 01: Write the characteristic equation of the given matrix A
|A − λI| = 0
Step - 02: Expanding the determinant, find the characteristics equation and roots
of the characteristics equation are called an eigen values of the given matrix A.
Step - 03: Find the eigen vectors X 6= 0 for each eigen value λ of the given matrix.
Continued Unit - II Problems on Eigen values and Eigen vector of a
Matrix ...
Problem - 01: Find all eigen values and corresponding eigen vectors of the matrix
 
−3 8
 −2 7
−3 8
A= −2 7
We know that, the characteristics equation of the given matrix is
|A − λI| = 0

−3 − λ 8

−2 7 − λ = 0

Now, expanding this determinant, we get
Matrix ...
(−3 − λ)(7 − λ) + 16 = 0
λ2 − 4λ − 5 = 0
(λ − 5)(λ + 1) = 0
∴ λ1 = 5 and λ2 = −1 are the eigen values of the given matrix A.
Let X = [x, y]T be the eigen vector of the matrix A. The matrix equation is
[A − λI][X] = [0]
    
−3 − λ 8 x 0
 −2 7 − λ  y  =  0 (29)
Case - 01: Put λ = λ1 = 5 in Eqn. (18), we get

    
−8 8 x 0
 −2 2  y  =  0
Matrix ...
From this we get the system of a linear equations
−8x + 8y = 0, −2x + 2y = 0,
The above equations are reduces to x − y = 0
x=y
x y
= = k1 (k1 6= 0)
1 1
where k1 is a arbitrary constant, x = k1 and y = k1 . Substitute x and y in the eigen
vector X, we get the eigen vector X1 = [k1 , k1 ]T or X1 = [1, 1] if k1 = 1
corresponding to the eigen value λ = λ1 = 5.
Matrix ...
Case - 02: Put λ = λ2 = −1 in Eqn. (18), we get

    
−2 8 x 0
 −2 8  y  =  0
From this we get the system of a linear equations
−2x + 8y = 0, −2x + 8y = 0,
The above equations are reduces to x − 4y = 0
x = 4y
x y
= = k2 (k2 6= 0)
4 1
where k2 is a arbitrary constant, x = 4k2 and y = k2 . Substitute x and y in the eigen
vector X, we get the eigen vector X2 = [4k2 , k2 ]T or X1 = [4, 1] if k2 = 1
corresponding to the eigen value λ = λ2 = −1.
Matrix ...
 
2 0 1
 0 2 0
 
 1 0 2
2 0 1
 0 2 0
A= 
 1 0 2

|A − λI| = 0

2−λ 0 1
0 2−λ 0
=0

1 0 2 − λ

Matrix ...
(2 − λ)[(2 − λ)2 − 0] − 0[0 − 0] + 1[0 − (2 − λ)] = 0

(2 − λ)[(2 − λ)2 ] − (2 − λ) = 0
(2 − λ)[(2 − λ)2 − 1] = 0
(2 − λ)(4 + λ2 − 4λ − 1) = 0
(2 − λ)(λ2 − 4λ + 3) = 0
→ λ = 2, λ = 1 and λ = 3
∴ λ = 1, λ = 2 and λ = 3 are the eigen value of the given matrix A
Consider the system of matrix equation
[A − λI][X] = [0]
where  
x1
 x2 
X= 
 x3 
Matrix ...
    
2−λ 0 1 x1 0

 0 2−λ 0 
 x2  
= 0
 1 0 2 − λ  x3   0
(2 − λ)x1 + 0x2 + x3 = 0
0x1 + (2 − λ)x2 + 0x3 = 0
1x1 + 0x2 + (2 − λ)x3 = 0
Case - 01: Substituting λ = 1 in the above system of a linear equations, we get
1x1 + 0x2 + x3 = 0
0x1 + 1x2 + 0x3 = 0

1x1 + 0x2 + 1x3 = 0
solve by using cross multiplication method
x1 x2 x3
=− =
(0 − 1) (0 − 0) (1 − 0)
Matrix ...
Let k1 be non zero arbitrary constant
x1 x2 x3
=− = = k1 , (k1 6= 0)
−1 0 1
∴ x1 = −k1 , x2 = 0 and x3 = k1
Hence the eigen vector corresponding to eigen value λ = 1 is
X1 = [−k1 , 0, k1 ]T = [−1, 0, 1]T .
0x1 + 0x2 + x3 = 0
0x1 + 0x2 + 0x3 = 0
1x1 + 0x2 + 0x3 = 0
From these, we get x1 = 0, x3 = 0 and let x2 = k2 , where k2 be a non zero arbitrary
constant. Hence the eigen vector corresponding to the eigen value λ = 2 is
X2 = [0, k2 , 0]T = [0, 1, 0]T .
(−1)x1 + 0x2 + x3 = 0
0x1 + (−1)x2 + 0x3 = 0
1x1 + 0x2 + (−1)x3 = 0
Matrix ...
From these, we get −x1 + x3 = 0, −x2 = 0 and x1 − x3 = 0 let x1 = k3 and x3 = k3
where k3 be a non zero arbitrary constant. Hence the eigen vector corresponding to
the eigen value λ = 3 is X3 = [k3 , 0, k3 ]T = [1, 0, 1]T .
Problem - 02: Find all the Eigen values and Eigen vectors for the matrix
 
8 −6 2
 −6 7 −4
 
 2 −4 3
8 −6 2
 −6 7 −4
A= 
 2 −4 3

|A − λI| = 0

8−λ −6 2
−6 7−λ −4
=0

2 −4 3 − λ

Matrix ...
 
2 0 1
 0 2 0
 
 1 0 2
2 0 1
 0 2 0
A= 
 1 0 2

|A − λI| = 0

2−λ 0 1
0 2−λ 0
=0

1 0 2 − λ

Matrix ...
(2 − λ)[(2 − λ)2 − 0] − 0[0 − 0] + 1[0 − (2 − λ)] = 0

(2 − λ)[(2 − λ)2 ] − (2 − λ) = 0
(2 − λ)[(2 − λ)2 − 1] = 0
(2 − λ)(4 + λ2 − 4λ − 1) = 0
(2 − λ)(λ2 − 4λ + 3) = 0
→ λ = 2, λ = 1 and λ = 3
[A − λI][X] = [0]
where  
x1
 x2 
X= 
 x3 
Matrix ...
    
2−λ 0 1 x1 0

 0 2−λ 0 
 x2  
= 0
 1 0 2 − λ  x3   0
(2 − λ)x1 + 0x2 + x3 = 0
0x1 + (2 − λ)x2 + 0x3 = 0
1x1 + 0x2 + (2 − λ)x3 = 0
1x1 + 0x2 + x3 = 0
0x1 + 1x2 + 0x3 = 0

1x1 + 0x2 + 1x3 = 0
x1 x2 x3
=− =
(0 − 1) (0 − 0) (1 − 0)
Matrix ...
Let k1 be non zero arbitrary constant
x1 x2 x3
=− = = k1 , (k1 6= 0)
−1 0 1
∴ x1 = −k1 , x2 = 0 and x3 = k1
Hence the eigen vector corresponding to eigen value λ = 1 is X1 = [−k1 , 0, k1 ]T , if
k1 = 1, then X1 = [−1, 0, 1]T .
0x1 + 0x2 + x3 = 0
0x1 + 0x2 + 0x3 = 0
1x1 + 0x2 + 0x3 = 0
From these, we get x1 = 0, x3 = 0 and let x2 = k2 , where k2 be a non zero arbitrary
constant. Hence the eigen vector corresponding to the eigen value λ = 2 is
X2 = [0, k2 , 0]T , if k2 = 1, then X2 = [0, 1, 0]T .
(−1)x1 + 0x2 + x3 = 0
0x1 + (−1)x2 + 0x3 = 0
1x1 + 0x2 + (−1)x3 = 0
Matrix ...
From these, we get −x1 + x3 = 0, −x2 = 0 and x1 − x3 = 0 let x1 = k3 and x3 = k3
where k3 be a non zero arbitrary constant. Hence the eigen vector corresponding to
the eigen value λ = 3 is X3 = [k3 , 0, k3 ]T , if k3 = 1, then X3 = [1, 0, 1]T .
Problem - 03: Find all the Eigen values and Eigen vectors for the matrix
 
8 −6 2
 −6 7 −4
 
 2 −4 3
8 −6 2
 −6 7 −4
A= 
 2 −4 3

|A − λI| = 0

8−λ −6 2
−6 7−λ −4
=0

2 −4 3 − λ

Matrix ...
(8 − λ)[(7 − λ)(3 − λ) − 16] + 6[−6(3 − λ) + 8] + 2[24 − 2(7 − λ)] = 0
(8 − λ)[λ2 − 10λ + 5] + 6[6λ) − 10] + 2[10 + 2λ] = 0

−λ3 + 18λ2 − 85λ + 40 + 40λ) − 40 = 0
−λ(λ2 − 18λ + 45) = 0
→ λ = 0, λ = 15 and λ = 3
[A − λI][X] = [0]
where  
x1
 x2 
X= 
 x3 
Matrix ...
    
8−λ −6 2 x1 0
 −6 7−λ −4   x2  
= 0
   (30)
 2 −4 3 − λ  x3   0
Case - 01: Substituting λ = 0 in the above matrix equation (1), we get

8x1 − 6x2 + 2x3 = 0
−6x1 + 7x2 − 4x3 = 0
2x1 − 4x2 + 3x3 = 0
x1 x2 x3
=− =
(24 − 14) (−32 + 12) (56 − 36)
x1 x2 x3
= =
10 20 20
x1 x2 x3
=− = = k1 , (k1 6= 0)
1 2 2
∴ x1 = k1 , x2 = 2k1 and x3 = 2k1 . Substitute x1 , x2 and x3 in the eigen vector X,
X1 = [k1 , 2k1 , 2k1 ]T , if k1 = 1, thenX1 = [1, 2, 2]T is eigen vector corresponding to
the eigen value λ = 0 of the given matrix A.
Matrix ...
5x1 − 6x2 + 2x3 = 0
−6x1 + 4x2 − 4x3 = 0

2x1 − 4x2 + 0x3 = 0
x1 x2 x3
=− =
(24 − 8) (−20 + 12) (20 − 36)
x1 x2 x3
=− =
16 −8 −16
x1 x2 x3
= = = k2 , (k2 6= 0)
2 1 −2
∴ x1 = 2k2 , x2 = k2 and x3 = −2k2 . Substitute x1 , x2 and x3 in the eigen vector X.
X2 = [2k2 , k2 , −2k2 ]T , if k2 = 1, then X2 = [2, 1, −2]T is eigen vector corresponding
to the eigen value λ = 3 of the given matrix A.
Matrix ...
−7x1 − 6x2 + 2x3 = 0
−6x1 − 8x2 − 4x3 = 0

2x1 − 4x2 − 12x3 = 0
x1 x2 x3
=− =
(24 + 16) (28 + 12) (56 − 36)
x1 x2 x3
=− =
40 40 20
x1 x2 x3
= = = k3 , (k3 6= 0)
2 −2 1
∴ x1 = 2k3 , x2 = −2k3 and x3 = k3 . Substitute x1 , x2 and x3 in the eigen vector X.
X3 = [2k3 , −2k3 , k3 ]T , if k3 = 1, then X3 = [2, −2, 1]T is eigen vector corresponding
to the eigen value λ = 15 of the given matrix A.
Matrix ...
Problems: Find all the eigen values and corresponding eigen vectors of the
following matrices.
     
7 −2 0 −2 2 −3 1 1 3
 −2 6 −2 2 1 −6 (3)  1 5 1
  
(1)   (2) 
 0 −2 5  −1 −2 0  3 1 1
Continued Unit - II on Rayleigh’s Power Method (OR) Power Method ...
Rayleigh’s Power Method (OR) Power Method:
This is a numerically an iteration method to obtain an approximately the largest
eigen value and corresponding eigen vector of the given matrix A. This method is
useful to write the programming for find the eigen values and corresponding eigen
vectors of the matrix A.
If X (0) is the initial eigen vector and A be the given matrix, then
Iteration - 1:
AX (0) = λ(1) X (1)
Iteration - 2:
AX (1) = λ(2) X (2)
Iteration - 3:
AX (2) = λ(3) X (3)
Iteration - 4:
AX (3) = λ(4) X (4)
Iteration - 5:
AX (4) = λ(5) X (5)
Iteration - 6:
AX (5) = λ(6) X (6)
Continued Unit - II on Rayleigh’s Power Method (OR) Power Method ...
Continue this iteration process till two consecutive iterative values of λ and X are
same atleast two decimal places.The value so obtained at last iteration values are
respectively the largest eigen value and the corresponding eigen vector of the given
matrix A.
Note: Suppose an initial eigen vector X (0) is not given in the problem, choose an
initial eigen vector like X (0) = [1, 0, 0]T (or) X (0) = [0, 1, 0]T (or) X (0) = [0, 0, 1]T
(or) X (0) = [1, 1, 1]T for the matrix of order 3.
Continued Unit - II Problems on Rayleigh’s Power Method (OR) Power
Method ...
Problems:
Problem - 01: Find the largest eigen value and the corresponding eigen vector of
the matrix A by using Power method.
 
2 0 1
 0 2 0
A=  
1 0 2
Solution: Given  
2 0 1
 0 2 0
A= 
 1 0 2
Now, taking X (0) = [1, 0, 0]T be the initial eigen vector

Iteration - 1:
      
2 0 1 1 2 1
 0 2 0   0  0 0
AX (0) =   = λ(1) X (1)

 1 0 2   0 =  1 = 2 
    
0.5
Continued Unit - II Problems on Rayleigh Power Method ...
Iteration - 2:
      
2 0 1 1 2.5 1
0 2 0 0 0 0
AX (1) =   = λ(2) X (2)
   
 =  = 2.5 
 1 0 2 
 0.5  2   0.8
Iteration - 3:
      
2 0 1 1 2.8 1
0 2 0 0 0 0 
AX (2) =   = λ(3) X (3)
   
 =  = 2.8 
 1 0 2 
 0.8  2.6   0.93
Iteration - 4:
      
2 0 1 1 2.93 1
0 2 0 0  0  0 
AX (3) =   = λ(4) X (4)
   
 =  = 2.93 
 1 0 2  0.93  2.86  0.98
Iteration - 5:
      
2 0 1 1 2.98 1
0 2 0 0  0  0 
AX (4) =   = λ(5) X (5)
   
 =  = 2.98 
 1 0 2  0.98  2.96  0.99
Iteration - 6:
      
2 0 1 1 2.99 1
0 2 0 0  0  0 
AX (5) =   = λ(6) X (6)
   
 =  = 2.99 
 1 0 2  0.99  2.98  0.997
Iteration - 7:
      
2 0 1 1 2.997 1
0 2 0 0  0  0 
AX (6) =   = λ(7) X (7)
   
 =  = 2.997 
 1 0 2 
 0.997  2.994  0.999
Therefore, we can observe that the values of λ(6) ≈ λ(7) and x(6) ≈ x(7) . Now, stop
the iteration process and hence λ = 2.997 ≈ 3 is the largest eigen value and the
corresponding eigen vector X = [1, 0, 0.999]T ≈ [1, 0, 1]T of the given matrix A.
Problem - 02: Apply Rayleigh’s power method to find the dominant eigen value
and the corresponding eigen vector of the matrix A, by taking the initial eigen vector
as [1, 1, 1]T .
 
6 −2 2
 −2 3 −1
A= 
 2 −1 3
Solution: By data  
6 −2 2
 −2 3 −1
A= 
 2 −1 3
and initial eigen vector X (0) = [1, 1, 1]T

Iteration - 1:
      
6 −2 2 1 6 1
 −2 3 −1 1 0 0 
AX (0) =   = λ(1) X (1)
  
 =  = 6
 2 −1 3  1  4  0.67
Iteration - 2:
      
6 −2 2 1 7.34 1
−2 3 −1 0  −2.67 −0.36
AX (1) =   = λ(2) X (2)
   
 =  = 7.34 
 2 −1 3 
 0.67  4.01   0.55 
Iteration - 3:
      
6 −2 2 1 7.82 1
−2 3 −1 −0.36 −3.63 −0.46
AX (2) =   = λ(3) x(3)
   
 =  = 7.82 
 2 −1 3 
 0.55   4.01   0.51 
Iteration - 4:
      
6 −2 2 1 7.94 1
−2 3 −1 −0.46 −3.89 −0.49
AX (3) =   = λ(4) x(4)
   
 =  = 7.94 
 2 −1 3  0.51   3.99   0.5 
Iteration - 5:
      
6 −2 2 1 7.98 1
−2 3 −1 −0.49 −3.97 −0.5
AX (4)  = λ(5) x(5)
    
=  = = 7.98 
 2 −1 3 
 0.5   3.99   0.5 
Iteration - 6:
      
6 −2 2 1 8 1
−2 3 −1 −0.5 −4 −0.5
AX (4)  = λ(6) x(6)
    
=  = = 8
 2 −1 3 
 0.5   4   0.5 
Therefore, we can observe that the values of λ(5) ≈ λ(6) and x(5) ≈ x(6) . Now, stop
the iteration process and hence λ = 8 is the largest eigen value and the corresponding
eigen vector X = [1, −0.5, 0.5]T of the given matrix A. Problem - 03: Apply
Rayleigh’s power method to find the dominant eigen value and the corresponding
eigen vector of the matrix A.
 
25 1 2
 1 3 0
A=  2 0 −4

Solution: By data  
25 1 2
 1 3 0
A= 
 2 0 −4
and taking an initial eigen vector X (0) = [1, 0, 0]T

Iteration - 1:
 
25 1 2      
 1 3 1 25 1
0 
  0  1  0.04
AX (0) =   = λ(1) X (1)
 
 2 0 −4  0 =  2  = 25 
    
  0.08
Iteration - 2:
 
25 1 2      
1 25.2 1
 1 3 0
0.04 1.12 0.04
AX (1)  = λ(2) X (2)
   
= 2 0 −4  =  = 25.2 



 0.08  1.68  0.07
Iteration - 3:
 
25 1
Dr. Vijaya kumar, MSRIT, Bengalur 2 Engineering Mathematics - III [ME/CH 31]
     August 17, 2021
 93 / 112
Continued Unit - II Problems on Rayleigh’s Power Method ...
Problems: Applying Power method to obtain a largest eigen value and

corresponding eigen vector of the following matrices.
 
3 −1 1
 −1 5 −1
(1)  
 1 −1 3
by taking [0, 1, 0]T as initial eigen vector.

   
4 1 0 −2 2 −3
 1 20 1  2 1 −6
 0 1 4 (3) 
(2)    
−1 −2 0
Continued Unit - II on Diagonalization of a Square Matrix ...
Singular matrix: A square matrix A of order n is said to be a singular matrix if

|A| = 0.
Non - singular matrix: A square matrix A of order n is said to be a Non -
singular matrix if |A| = 6 0.
Similarity Transformation: If A and B are any two square matrices of order n, if
there exist a non - singular matrix P such that P −1 AP = B, then the matrix A is
said to be similar to the matrix B.
Similarly, if P −1 BP = A, then the matrix B is said to be similar to the matrix A.
Diagonalization of a square matrix: Suppose A is a square matrix of order n, if
there exist a non - singular matrix P such that P −1 AP = D, then the matrix A is
said to be a diagonalize matrix.
P −1 AP = D
Model matrix: If a matrix A is square matrix of order n, if there exist a non -
singular matrix P such that P −1 AP = D, then a non - sigular matrix P is called as a
Model matrix and whose columns are the eigen vectors of the matrix A, that is
P = [X1 , X2 , X3 , ..., Xn ]
Spectral matrix: If a matrix A is square matrix of order n, if there exist a non -

singular matrix P such that P −1 AP = D, then a diagonal matrix D is called as a
Spectral matrix and whose diagonal elements are the eigen values of the matrix A,
that is  
λ1 0 0 0 ... 0
 0 λ2 0 0 ... 0 
 
 0 0 λ3 0 ... 0 
 
D=  0 0 λ4 0 ... 0  
 .. .. .. .. ... .. 
 
 0 0 0 0 ... λn 
Power of a matrix: If A is the given square matrix of order n, then the power of
the matrix A is
We know that, by diagonalization of a matrix A is
P −1 AP = D
pre multiply by P and post multiply by P −1 on both sides, we get
(P )(P −1 AP )(P −1 ) = (P )D(P −1 )
(P P −1 )(A)(P P −1 ) = P DP −1
−1
but, P P = I and AI = IA = A, IAI = A
(I)(A)(I) = P DP −1
(IAI) = P DP −1
A = P DP −1
Therefore, consider the matrix A2
A2 = (A)(A) = (P DP −1 )(P DP −1 )
A2 = (P D)(P −1 P )(DP −1 )
A = (P D)(I)(DP −1 ) = (P D)(DP −1 )
2
A2 = P D2 P −1
similarly, continue like this we get
A3 = P D3 P −1
A4 = P D4 P −1
In general, for a positive integer m
Am = P Dm P −1
Working procedure to reduce to the diagonal form of the given matrix:

Step - 01: Write the characteristics equation for the given matrix A of the form
|A − λI| = 0, expanding this determinant find all eigen values of the given matrix A.
Step - 02: Obtain all eigen vectors corresponding each eigen values of the given
matrix A.
Step - 03: Verify the given matrix A is reduce to a diagonal for by using the
formula .
P −1 AP = D
where
1
P −1 = Adj(P )
|P |
Step - 04: Find the power of the given matrix A by using the formula .
Am = P Dm P −1
λm
 
1 0 0 0 ... 0
 0 λm2 0 0 ... 0 
λm
 
 0 0 3 0 ... 0 
Dm λm
 
=
 0 0 4 0 ... 0 

 .. .. .. .. ... .. 

 0 0 0 0 ... λm 
Continued Unit - II Problems on Diagonalization of a Square Matrix ...
Problem - 01: Diagonalize the matrix and hence find A6 .

 
2 2
 1 3
Solution: Given a matrix  

2 2
A= 1 3
We know that, the characteristics equation
|A − λI| = 0

2−λ 2

1 3 − λ = 0

Now, expanding the determinant, we get, the eigen values are 1 and 4
Let X = [x, y]T be the eigen vector and matrix equation is
[A − λI][X] = [0]
    
2−λ 2 x 0
 1 3 − λ  y =  0 (31)

x + 2y = 0
and
x + 2y = 0
both are same
x = −2y
x y
= = k1 , (k1 6= 0)
2 −1
x y
= k1 , x = 2k1 , = k1 , y = −k1
2 −1
Substitute x = 2k1 and y = −k1 in the eigen vector X, X1 = [2k1 , −k1 ]T and
X1 = [2, −1]T , if (k1 = 1) be the eigen vector corresponding to the eigen value λ1 = 1.
−2x + 2y = 0
and
x−y =0
both are same
x=y
x y
= = k2 , (k2 6= 0)
1 1
x y
= k2 , x = k2 , = k2 , y = k2
1 1
Substitute x = k2 and y = k2 in the eigen vector X, X2 = [k2 , k2 ]T and X2 = [1, 1]T ,
if (k2 = 1) be the eigen vector corresponding to the eigen value λ1 = 4.
∴ Model matrix
P = [X1 , X2 ]
 
2 1
P =  −1 1
Spectral matrix D is  
1 0
D= 0 4
We know that, by diagonalization of a matrix A is
P −1 AP = D
 −1     
2 1 2 2 2 1 1 0
−1
P AP =  −1 1  1 3  −1 1 =  0 4
To find A6 : We have,
A6 = P D6 P −1
  6  
2 1 1 0 2 1
A6 =  −1 1  0 4  −1 1
−1
16
  
2 1 0 2 1
6 6 
A = −1 1  0 4 −1 1
   −1  
2 1 1 0 2 1 1366 2730
6
A = −1 1  0 4096  −1 1 =  1365 2731
Problem - 02: Diagonalize the matrix and hence find A6 .

 
1 1 3
 1 5 1
 
 3 1 1
Solution: Given a matrix  

1 1 3
 1 5 1
A= 
 3 1 1
Continued Unit - II on Application of a first order and first degree
ODE’s using Matrix Method ...
Consider a simple first order and first degree ODE of the form
dy
= ay
dx
where a is any constant and y is unknown function to be determine. Find the general
solution of the ODE
Method - I: Using the method of separation of variables
dy
= adx
y
Integrating on both sides, we get
Z Z
dy
=a dx + c
y
log(y) = ax + c
y = eax+c = eax ec
y = c1 eax , c1 = ec
Therefore, y = c1 eax is the general solution of the above ODE.
Method - II: Using the method to solve second and higher order LDE
dy
= ay
dx
d
Dy = ay, D =
dx
Dy − ay = 0
(D − a)y = 0
The A. E. is f (m) = 0, m − a = 0, m = a, C.F = c1 eax and P.I = 0, because RHS is
zero. the general solution is y = c1 eax of the above ODE.
dy
Note: suppose dx = ay is the given first order and first degree ordinary differential
equation then the general solution is y = c1 eax .
Consider a set of m first order and first degree ordinary differential equation in n
unknowns as follows
y10 = a11 y1 + a12 y2 + a13 y3 + ... + a1n yn
y20 = a21 y1 + a22 y2 + a23 y3 + ... + a2n yn

y30 = a31 y1 + a32 y2 + a33 y3 + ... + a3n yn
...
...
yn0 = am1 y1 + am2 y2 + am3 y3 + ... + amn yn
Where, y1 = f1 (x), y2 = f2 (x), y3 = f3 (x), ..., yn = fn (x) are the unknown functions
to be determine.
The above system of ODE’s in matrix equation is
y 0 = Ay
where, A is the coefficient matrix of the given system of ODE.
y10
    
a11 a12 a13 .. ... a1n y1
 y20   a21 a22 a23 .. ... a2n   y2 
y30 
    

  
 a31 a32 a33 .. ... a3n 

 y3 


 ...  =
 
 .. .. .. .. ... .. 

 .. 

 ...   .. .. .. .. ... ..   .. 
0 
    
 yn  am1 am2 am3 .. ... amn 
 yn 
where, the coefficient matrix A is

 
a11 a12 a13 .. ... a1n
 a21 a22 a23 .. ... a2n 
 
 a31 a32 a33 .. ... a3n 
 
A=  .. .. .. .. ... .. 


 .. .. .. .. ... .. 

 am1 am2 am3 .. ... amn 
Continued Unit - II Problems on Application of a first order and first
degree ODE’s using Matrix Method ...
Problem - 1: Write the following system of ODE’s in matrix form and hence solve
using matrix method
y10 = −2y1 , y20 = 3y2 , y30 = y3
given that y1 (0) = 1, y2 (0) = −1 and y3 (0) = 2.
Solution: The given system of ODE’s
y10 = −2y1 , y20 = 3y2 , y30 = y3
and initial conditions y1 (0) = 1, y2 (0) = −1 and y3 (0) = 2.

The given system of ODE’s in matrix equation
y 0 = Ay
y10
    
−2 0 0 y1
0
 y2   0 3 0  y2 
=
y30  
  
 0 0 1  y3 
The general solutions for the given system of ODE’s are
y10 = −2y1 → y1 = c1 e−2x
y20 = 3y2 → y2 = c2 e3x
y30 = y3 → y3 = c3 ex
∴ y1 = c1 e−2x , y2 = c2 e3x and y3 = c3 ex are the general solutions for the given
system of ODE’s. Also given the initial conditions
y1 (0) = 1, when x = 0, y1 = 1
y1 = c1 e−2x → c1 = 1, y1 = e−2x
y2 (0) = −1, when x = 0, y2 = −1
y2 = c2 e3x → c2 = −1, y2 = −e3x
y3 (0) = 2, when x = 0, y3 = 2
y3 = c3 ex → c3 = 2, y3 = 2ex
Hence y1 = e−2x , y2 = −e3x and y3 = 2ex are the particular solutions for the given
system of ODE’s
Note: In the above system of ODE’s the coefficient matrix A is diagonal matrix and
hence it is easy to solve the given system of ODE’s. If the coefficient matrix A of the
given system of ODE’s is not in diagonal matrix then it reduce to diagonal matrix
using the concept of the diagonalization of the square matrix and solve the given
system of ODE’s. We have to taking y = P u and u0 = Du, where P is the model
matrix and D is the spectral matrix of the coefficient matrix A of the given system of
ODE’s.
P = [X1 , X2 , X3 ]
 
λ1 0 0
 0 λ2 0 
D=  0

0 λ3 
u01
   
u1
0
u2  u2 
u0 = 
 
,u = 
u03 

  u3 
using matrix method
x01 = x1 + x2 , x02 = 3x1 + 3x2
given that x1 (0) = 0, x2 (0) = 0 .
using matrix method
y10 = y1 + y2 , y20 = 4y1 − 2y2
given that y1 (0) = 1, y2 (0) = 6 .
using matrix method
y10 = 4y1 + 2y2 + 2y3 , y20 = 2y1 + 4y2 + 2y3 , y30 = 2y1 + 2y2 + 4y3

III SEM ME - CH 31-301 - Unit - II

Uploaded by

Copyright:

Available Formats

III SEM ME - CH 31-301 - Unit - II

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

III SEM ME - CH 31-301 - Unit - II

Uploaded by

Copyright:

Available Formats

Unit - II: LINEAR ALGEBRA

Consistency of a system of linear equations:

a11 x1 + a12 x2 + ... + a1n xn = b1

a21 x1 + a22 x2 + ... + a2n xn = b2

Geometrical Representation of a system of the linear equations for

Note - 01: Homogeneous system of linear equations is always consistent because it

From this we get, k = 11

The matrix equation of the given system of linear equations is

Reducing the augmented matrix [A : B] to row echelon form by applying the

Problem - 02: Test consistency and solve 5x + 3y + 7z = 4, 3x + 26y + 2z = 9,

The given system of a linear equations are in matrix equation

The augmented matrix of the given system of linear equations [A : B] is

Reducing the augmented matrix [A : B] to row echelon form by applying the

The augmented matrix [A : B] is in row echelon form, ρ(A) = 2 and ρ(A : B) = 2.

The given system of a linear equations are in matrix equation

The augmented matrix of the given system of linear equations [A : B] is

Reducing the augmented matrix [A : B] to row echelon form by applying the

The augmented matrix [A : B] is in row echelon form, ρ(A) = 2 and ρ(A : B) = 3.

The augmented matrix of the given system of linear equations [A : B] is

Reducing the augmented matrix [A : B] to row echelon form by applying the

The augmented matrix [A : B] is in row echelon form, ρ(A) = 2 and ρ(A : B) = 3,

The augmented matrix of the given system of linear equations [A : B] is

Reducing the augmented matrix [A : B] to row echelon form by applying the

Reducing the augmented matrix [A : B] to row echelon form by applying the

Now, converted to a system of linear equations

Reducing the augmented matrix [A : B] to row echelon form by applying the

Now, converted to a system of linear equations

solve by back substitution method

a11 x + a12 y + a13 z = b1 (12)

Let us choose an initial approximations are x0 = 0, y 0 = 0 and z 0 = 0.

−x−y−2z+10u = −9, −2x+10y−z−u = 15, −x−y+10z−2u = 27, 10x−2y−z−u = 3

Solution: By data the system of a linear equations are

Substitute x = x(1) = 0.3, y = y (1) = 1.56 and u = u(0) = 0 in Eqn. (20)

Substitute x = x(2) = 0.88692, y = y (2) = 1.952304 and u = u(1) = −0.1368 in Eqn.

Substitute x = x(3) = 0.9836, z = z (2) = 2.95656 and u = u(2) = −0.02476 in Eqn.

x(4) = 0.9968, y (4) = 1.9982, z (4) = 2.9987, u(4) = −0.00038

x(5) = 0.9994, y (5) = 1.9997, z (5) = 2.9998, u(5) = −0.0001

x(6) = 0.9999, y (6) = 1.9999, z (6) = 3.0000, u(6) = 0.0000

x + y + 54z = 110, 27x + 6y − z = 85, 6x + 15y + 2z = 72

Perform three iterations.

13x+5y−3z+u = 18, 2x+12y+z−4u = 13, 3x−4y+10z+u = 29, 2x+y−3z+9u = 31

Therefore, (A − λI)X = 0 becomes in the matrix form

∴ |A − λI| = 0 is called as a characteristics equation.

On expanding this determinant, we get a polynomial equation in λ of degree 3 is

A−1 (AX) = A−1 (λX)

(A−1 A)X = (λ)(A−1 X)

Again, multiplying by A on both sides, we get

Hence, λm is a eigen value of a Am .

The characteristics equation of a matrix A is

We know that, the characteristics equation of the given matrix is

Case - 01: Put λ = λ1 = 5 in Eqn. (18), we get

From this we get the system of a linear equations

The above equations are reduces to x − y = 0

Case - 02: Put λ = λ2 = −1 in Eqn. (18), we get

From this we get the system of a linear equations

The above equations are reduces to x − 4y = 0

We know that, the characteristics equation of the given matrix is

(2 − λ)[(2 − λ)2 − 0] − 0[0 − 0] + 1[0 − (2 − λ)] = 0