Elementary Transformations Example.

The following operations w.r.t. a matrix are known as Using elementary transformations reduce
elementary transformations. the matrix A first to upper triangular
Interchange of any two ROWS, indicated by Rij or by
Ri Rj and then to identity matrix where
2 5 7
Multiplication of elements of any ROW by a non-zero A = 1 2 3
real number, indicated by Ri kRi
3 1 2
Addition of the constant multiple of jth ROW to ith ROW
indicated by Ri Ri + kRj
2 5 7
Similar COLUMN transformation are denoted by Solution: 1 2 3

Cij , kCi, Ci + kCj 3 1 2

Solution: 1 2 3 Solution:
1 0 0
Operating R1 R2 A = 2 7 Operating C2 C2-2C1, 1
5
A~ 0 1
3 1 2 C3 C3-3C1
0 0 2
Operating R2 R2-2R1, 1 2 3

1 0 0
1 A~ 0
A~ 0 Operating C3 C3-C2
0
1
and R3 R3-3R1 1
0 5 7
0 0 2
1 2 3
1 0 0
1
Operating R3 R3 +5R2 A~ 0 1
Operating C3 C3 / (-2) A~ 0 1 0
0 0 2
0 0 1
(Upper Triangular Form) (Identity matrix)

2 3 1 1
Definition: Two matrices A and B are said to be SOLUTIONS 1
1 2 4
equivalent, if one can be obtained from another by a (1) Let A = 3 1 3 2
sequence of elementary transformations and the same is
7
6 3 0
symbolically written as A~B.
1 1 2 4
Example: Using elementary transformations find the 2 3 1 1
ranks of the following matrices
Operating R12 3 1 3 2
(1) 2 3 1 1 (2) 1 2 3 (3) 1 2 1 0
1 1 2 4 2 4 3 0 6 3 0 7
2 4 7
1 1 2 4
3 1 3 2 1 0 2 8
3 6 10 Operating R2 R2 -2R1, 0
5 3 7
6 3 0 7
(5) 9 1 9 2 9 3 9 4 9 5 R3 R3-3R1, R4 R4-6R1 0 4 9 10

(4) 1 4 3 2 1 0 9 12 17
2 3 1 9 3 92 93 94 95 96
1 1 2 4

1 6 7 12 9
93 94 95 96 97 Operating R2 R2 -R3, 0 1 6 3

94 95 96 97 98 0 4 9 10
3 3 6 21 12
0 9 12 17

1
 1 1 2 4 (2) 1 2 3
Operating R3 R3-4R2 , 0 1 6 3 B = 2 4 7

R4 R4-9R2 0 0 33 22 3 6 10

0 0 66 44
Operating R2R2 -2R1, 1 2 3
1 1 2 4 B ~ 0 1
Operating , R4 R4-2R3 A ~ 0 1 6 3 R3R3-3R1 0
0 0 1
0 0 33 22
Though the matrix is in Upper Triangular Form
0 0 0 0
we observe that R2 and R3 are identical.
Observe that the matrix is in UPPER TRIANGULAR FORM
1 2 3
The rank of the matrix = the number of non zero rows
Operating R3 R3-R2 B ~ 0 0 1
(A) = 3 0 0 0

This form of the matrix where,

(3) Reduce to row -echelon
1 2 1 0
3 0
(i) First non-zero entry (or element) in every non-zero form and find rank of C for, C = 2 4
row is UNITY or can be made unity. 1 0 2 8

(ii) Elements in the column below these above 1 2 1 0

mentioned UNITIES are zeroes Or the number of Operating R2 R2+2R1,
C~ 0 8 5 0
zeroes before the non-zero element increases with row
R3 R3-R1

number. 0 2 1 8
(iii) All the zero rows are below the non-zero rows
1 2 1 0
is known as ROW ECHELON FORM
Operating R23 C ~ 0 2 1 8

The rank of the matrix = the number of non zero rows 0 8 5 0
(A) = 2

1 4 3 2 1
1 2 1 0 (4)
Operating R3 R3+4R2 C ~ 3 1 9 3
C =2
0 2 1 8 1 6 7 12 9

0 0 9 0 3 3 6 21 12
1 0 0 0 0
C2
C~
1 1 1 0 Operating C2 C2-4C1,
2
1 1 8
Operating 5 5 5 5
(2) 0 C3 C3-3C1 , C4 C4+2C1, C~
1 10 10 10 10

0 0 9 0 C5 C5 - C1
3

15 15 15 15
1 1 1 0 1 0 0 0 0
Operating R3
9 C ~ 0 1 1 8 Operating C3 C3 -C2, 2 5 0 0 0
C ~
0 0 1 0 C4 C4 -C2, C5 C5 -C2 1 10 0 0 0
This is in Row- echelon form.
3 15 0 0 0
The rank of the matrix = the number of non zero rows Lower Triangular form
The rank of the matrix = the number of non zero columns
(A) = 3
(A) = 2

2
 91 92 93 94 95 1 1 1 1 1
(5)
Find the rank by reducing 92 93 94 95 96 Operating R2 R2 -91R1,
0 1 2 3 4
the following matrix : D = R3 R3 -2R1, D~
93 94 95 96 97 0 0 0 0 0
R4 R4 -3R1
94 95 96 97 98 0 0 0 0 0
91 92 93 94 95
Operating R2R2 -R1, 1 1 1 1 1
R3R3 -R1 , D~
2 2 2 2 2
R4R4 -R1 This is in Row- echelon form.
3 3 3 3 3
The rank of the matrix = the number of non zero rows
1 1 1 1 1
91 92 93 94 95 (A) = 2
Operating R12 D~
2 2 2 2 2

3 3 3 3 3

1 2 1 0 1 2 1 0
(3) Reduce to row -echelon Operating R3 R3+4R2 C ~
form and find rank of C for, C = 2 4 3 0 0 2 1 8
1 0 2 8 0 0 9 0

C2 1 1 1 0
C~
1 1 0 1 1 8
2 Operating
(2) 0

Operating R2 R2+2R1,
C~ 0 8 5 0 0 0 9 0
R3 R3-R1

0 2 1 8 1 1 1 0
Operating R3
9 C ~ 0 1 1 8

1 1 0 0 0 1 0
2 This is in Row- echelon form.
Operating R23 C ~ 0 2 1 8
The rank of the matrix = the number of non zero rows
0 8 5 0
(A) = 3

1 4 3 2 1 91 92 93 94 95
(4) (5)
3 1 9 3 Find the rank of 92 9 3 9 4 9 5 9 6
C =2
1 6 7 12 9 the following matrix : D =
93 94 95 96 97
3 3 6 21 12
1 0 0 0 0 94 95 96 97 98
Operating C2 C2-4C1,
2 5 5 5 5 91 92 93 94 95
C3 C3-3C1 , C4 C4+2C1, C~
1 10 10 10 10

Operating R2R2 -R1, 1 1 1 1 1
C5 C5 - C1
3

15 15 15 15 R3R3 -R1 , D~
2 2 2 2 2
1 0 0 0 0
R4R4 -R1
Operating C3 C3 -C2, 2 5 0 0 0 3 3 3 3 3
C ~
C4 C4 -C2, C5 C5 -C2 1 10 0 0 0 1 1 1 1 1
91 92 93 94 95
3 15 0 0 0
Lower Triangular form Operating R12 D~
2 2 2 2 2
The rank of the matrix = the number of non zero columns
3 3 3 3 3
(A) = 2

3
Operating R2 R2 -91R1, 1 1 1 1 1 CONSISTENCY OF SYSTEM OF
0 1 2 3 4
R3 R3 -2R1, D~ SIMULTANEOUS LINEAR EQUATIONS
0 0 0 0 0
R4 R4 -3R1 Consider the following set of Linear equations.
0 0 0 0 0
a11 x1 + a12 x 2 + a13 x 3 + . . . + a1n xn = b1
a 21 x1 + a 22 x 2 + a 23 x 3 + . . . + a 2n xn = b 2
This is in Row- echelon form.
.... .... .... .... ....
The rank of the matrix = the number of non zero rows
.... .... .... .... ....
(A) = 2
.... .... .... .... ....
am1 x1 + am 2 x 2 + am 3 x 3 + . . . + amn xn = bm

From the knowledge of matrix multiplication, where a a a .... .... a

11 12 13 1n
this can be written in matrix form
a 21 a 22 a 23 .... .... a 2n
a11 a12 a13 .... .... a1n .... .... .... .... .... ....
x1 b1 A= ,
x b .... .... .... .... .... ....
a 21 a 22 a 23 .... .... a 2n
2 2
.... .... .... .... .... .... x3 b3 .... .... .... .... .... ....
an1 an 2 an 3 .... .... ann
=
.... .... .... .... .... .... ... ... x1 b1
... ... x b
.... .... .... .... .... .... 2
2

x3 b
am1 am 2 am 3 .... .... amn xn bm X = , and D = 2
... ...
... ...
i.e., AX = D
xn

b n

and
Here A is called coefficient matrix, X is C = [ A | D]
called variable vector and D is called
Right hand side column vector.
a11 a12 . . . a1n b1
If D = 0 i.e. b1, b2,, bm = 0 then the a
matrix equation becomes A X = 0, 21 a22 . . . a2 n b2
which is called a system of = .......................
HOMOGENEOUS linear equations.
.......................
If D 0 then the system A X = D is am1 am 2 . . . am n bm
called a system of
NON-HOMOGENEOUS linear equations.
is called the AUGMENTED MATRIX.

4
 To solve a non-homogeneous system If [A | D] (A) the system will have NO
of linear equations: solution i.e. the system will be
First, we calculate the rank of INCONSISTENT.
Augmented Matrix D.
After checking above three conditions if
If [A | D] = (A) = n (number of the system is consistent
variables) the system will be consistent
and will have UNIQUE solution. then its solution is calculated by, again
converting the Upper triangular /
If [A | D] = (A) < n (number of Echelon form of the Augmented Matrix D
variables) the system will be consistent (from which rank was found) into linear
but will have infinite number of equations and solving it.
solutions

Example 1 Test the consistency and solve Here 1 2 3 x 8

A = 2 3 0 X = y D = 5

x-2y+3z= 8
1 1 1 9
2x-3y = -5 ( *) z
x+ y+ z = 9 Then (*) in matrix form is AX = D
Solution:
To test the consistency, we calculate the 1 2 3 8
Augmented matrix of the system (*) 2 3 0 5
Consider [A/D]=
And then we find 1 1 1 9
i) the rank of matrix [A/D]
ii) the rank of matrix A (Augmented matrix)

Operating R2 R2 -2R1, R3 R3-R1

(*) and (**) are equivalent system
1 2 3 8
[A/D] ~ 0 1 6 21

0 3 2 1 x - 2y + 3z = 8 --------(i)
Operating R3 R3-3R2 1 2 3 8 (**) y - 6z = -21 --------(ii)
16z = 64 --------(iii)
[A/D] ~ 0 1 6 21 (**)
0 0 16 64 (iii) z=4
z = 4 in (ii) y=3
( (A) = 3 ) = ((A/D) = 3)
y = 3 and z = 4 in (i) x=2
System of equations (*) are consistent.
( (A) = 3 ) = ((A/D) = 3) = No. of unknowns Answer x = 2, y = 3 and z = 4
There exists a unique solution of (*)

5
Example 2 2 1 2 1
Test the consistency and solve (Augmented
Consider [A/D]= 1 2 1 2
2x+y+2z=1 matrix)
x+2y - z=2 ( *) 5 4 3 4
5x+4y+3z=4 Operating R12 1 2 1 2
To test the consistency, we calculate the Augmented matrix of the
system (*) And then we find [A/D] ~ 2 1 2 1
i) the rank of matrix [A/D] 5 4 3 4
ii) the rank of matrix A
Here Operating R2 R2 - 2R1, R3 R3 - 5R1
2 1 2 x 1
1 2 1
D = 2 1 2 1 2
A= X = y

[A/D] ~ 0 3 4 3

5 4 3 z 4

Then (*) in matrix form is AX = D 0 6 8 6

Operating R3 R3-2R2
x+ 2y - z = 2 --------(i)
(**)
1 2 1 2 -3y+ 4z = -3 --------(ii)

[A/D] ~ 0 3 4 3 (**) Let z = k be the parameters
0 0 0 0
4 5
y = 1+ k x=- k
3
( (A) = 2 ) = ((A/D) = 2) 3
4
System of equations (*) is consistent Answer: x = - 5
k y = 1+ k z=k
3 3
( (A) = 2 ) = ((A/D) = 2) < No.of unknowns=3
For k = 3 x = -5 y = 5 z = 3
There exist infinite solutions of (*) is one particular solution.
(*) and (**) are equivalent system

Example 5. For what values of and Operating, R3-R2,R1-R2

the following system 1 1 1 6
[A/D] ~
x+y+z=6 0 1 2 4

x + 2y + 3z = 10 0 1 1 6
x + 2y + z =

Operating R3-R2,
will have 1 1 1 6
0 1 2 4
i)Unique solution [A/D] ~
ii) Infinite solutions 0 0 3 10
1 1 1 6
1
iii)No solution 2 3 10
[A/D] ~ 1 1 1 6
1
[A/D] ~ 0 1 2 4
2

0 1 1 6

6
 System of equations will have i) Infinite solution if ( (A) = 2 ) = ((A/D = 2)

i) Unique solution if ( (A) = 3 ) = ((A/D = 3) < No. of unknowns=3

= No. of unknowns=3 hence = 3 and = 10

hence 3 0 ii) Infinite solution if ( (A) 2 ) ((A/D = 3)

Example 3. Test the consistency and solve Operating R2 R2 - 3R1, R3 R3 - 7R1

x-4y+7z=14
1 4 7 14
3x+8y -2 z=13 (*)
7x-8y+26z=5 [A/D] ~ 0 2 0 2 3 2 9

To test the consistency, we calculate the 0 2 0 2 3 9 3
Augmented matrix of the system (*) Operating R3 R3-R2
And then we find 1 4 7 14
i) the rank of matrix [A/D] 2 9
[A/D] ~ 0 2 0 2 3
ii) the rank of matrix A
0 0 0 6 4
1 4 7 14
( (A) = 2 ) ((A/D) = 3)
Consider [A/D]= 3 8 2 1 3
There exists no solution of system of simultaneous
7 8 2 6 5 linear equations(*).

System of Homogeneous Linear which can be written as

Equations a11 a12 . . . a1n x1 0
x 0
a11 x1 + a12 x 2 + . . . + a1n xn = 0 a a . . . a
21 22 2n
2
a 21 x1 + a 22 x 2 + . . . + a 2n xn = 0 . . . . . . . . . . . . . . . . . . . . = .

........................... . . . . . . . . . . . . . . . . . . . . .
a xn 0
............................ m1 am2 . . . amn
am1 x1 + am2 x 2 + . . . + amn xn = 0

7
 or AX = O 1. If
For such a system the [A] is always [A] = n (number of var iables)
equal to [A / D]. i.e | A | 0 then the system (1) will
Hence a system of Homogeneous linear have only Zero or Trivial or Unique
equations is always consistent. Solution.
There are following cases: 2. If
[A] < n (number of var iables)
i.e | A | = 0 then the system (1) will
have Non-Zero or Non-Trivial or
Infinite number of Solutions.

Example-1 Solve the system of i.e. AX = O , where the coefficient

equations matrix
x + 3 y 2z = 0 1 3 2
2x y + 4z = 0 A = 2 1 4

x 11y + 14z = 0 1 11 14
Solution: The given system of equation
can be written as
1 3 2 x 0
2 1 4 y = 0 Applying R 2 R 2 2R 1 ,
R 3 R 3 R1
1 11 14 z 0

1 3 2 which is in Echelon form and hence

A 0 7 8 [A] = 2 < 3 (number of var iables).
Therefore, the given system has non-
0 14 16 trivial solutions.
Applying R 3 R 3 2R 2 Converting last form of A into equations
we get
1 3 2 x + 3y 2z = 0
A 0 7 8 7y + 8z = 0

0 0 0

8
On assuming z = k , where k is arbitrary Example-2 Solve the system of
constant and applying back substitution, equations
we find x + 2 y + 3z = 0
8k 10k 3x + 4y + 4z = 0
that y = and x = .
7 7 7x + 10y + 12z = 0
10k 8k
Thus x = ,y= and z = k are
7 7 Solution: The given system of equation
the solutions of given system. can be written as

1 2 3 x 0 Applying R 2 R 2 3R 1 ,
3 4 4 y = 0 R 3 R 3 7R 1
1 2 3
7 10 12 z 0 A 0 2 5

i.e. AX = O , where the coefficient 0 4 9
matrix
Applying R 3 R 3 2R 2
1 2 3
1 2 3
A = 3 4 4
A 0 2 5
7 10 12
0 0 1

which is in Echelon form and hence Example-3 For what values of and
[A] = 3 = (number of var iables). , the following non-homogeneous
Therefore, the given system has trivial or system
zero or unique solution. x+y+z=6
x + 2y + 3z = 10
Thus x = 0 , y = 0 and z = 0 is the only
x + 2y + z =
solution of given system.
will have
(i) Unique solution
(ii) No solution
(iii) Infinite number of solutions.

9
 1 1 1 6

Solution: The given system of equation [A / D] = 1 2 3 10
can be written as
1 1 1 x 6 1 2
1 2 3 y = 10 Applying R 2 R 2 R 1 ,
R 3 R 3 R1
1 2 z
1 1 1 6
i.e. AX = D , where the augmented
matrix [A / D] = 0 1 2 4
0 1 1 6

Applying R 3 R 3 R 2 3 0 3.
1 1 1 6 (ii) For No solution, we require
[A / D] (A) , which is
[A / D] = 0 1 2 4
0 0 3 10 possible iff.:
3 = 0 and 10 0
which is in Echelon form. Or = 3 and 10
(i) For Unique solution, we require (iii) For Infinite number of solutions, we
require
[A / D] = (A) = 3 (number of vari [A / D] = (A) < 3 (number of variables)
, which is possible iff.: which is possible iff.:
3 = 0 and 10 = 0

Example-3 For what values of and

, the following non-homogeneous
system
2x + 3y + 5z = 9
7x + 3y -2z = 8
2x + 3y + z =
will have
(i) Unique solution
(ii) No solution
(iii) Infinite number of solutions.

10