Module 1: Matrices and Linear Algebra

Lesson 2

Rank of a Matrix and Solution of a Linear System

2.1 Introduction

In this lecture we shall discuss about the rank of matrices, the consistency of
systems of linear equations, and finally present the Gauss elimination method
for solving the linear systems. For this we need an important form of matrices
called echelon form which is obtained by applying elementary row (or column)
operations.

2.2 Echelon Form of a Matrix

Echelon form of a matrix is useful in solving system of linear equations, finding

rank of a matrix and checking many more results in linear algebra.

An m × n matrix A is said to be in (row) echelon form if

(i) All the zero rows of A are at the bottom.

(ii) For the non-zero rows of A, as the row number increases, the number of zero
entries at the beginning of the row also increases.

In the echelon form of a matrix some people consider one more condition that
the 1st non-zero entry in a non-zero row is equal to 1. However this condition is
not required for us and therefore not included in the definition of echelon form
of a matrix. One finds row echelon form of a matrix by applying elementary
row operations. By applying elementary column operations, one gets column
echelon form of the matrix.

Example 2.2.1: Find the row echelon form of

Rank of a Matrix and Solution of a Linear System

1 3 5 
A = 1 4 3  .
1 1 9 
 

We keep 1st row as it. Then we make 1st entry of the second row zero by
applying elementary row operations. So replacing 2nd row R2 by R2 − R1 one
gets
1 3 5 
 
 0 1 −2  .
1 1 9 
 

Then we make at least 1st two entries of the 3rd row of the above matrix zero.
For this we replace R3 by R3 − R1 in the above matrix and get

1 3 5 
 
 0 1 −2  .
 0 −2 4 
 

Finally by replacing R3 by R3 + 2R2 one gets the echelon form of A and is given
by

1 3 5 
 
 0 1 −2  .
0 0 0 
 

2.3 Rank of a Matrix

The rank of a matrix has several equivalent definitions. Here we take the rank
of a matrix A as the number of non-zero rows in the row echelon form of A. It
is also defined as the number of nonzero columns in the column echelon form of
the matrix. Whatever way the definition may be given the rank of a matrix will
Rank of a Matrix and Solution of a Linear System

be the same, is a fixed number. Therefore the rank of a matrix A has the
following properties.

(1) Matrix A and its transpose have the same rank, that is, rank(A) = rank(AT).

(2) If A is a matrix of size m × n then rank (A) is at the most min{m, n}.

(3) If B is a sub-matrix of A then rank (B) is less than or equal to rank(A).

Example 2. 3.1: Here we find rank of the matrix

 2 −2 3 4 −1
 
−1 1 2 5 2 
A=  .
 0 0 −1 −2 3 
 
 1 −1 2 3 0 

Here we find echelon form of the matrix A. First row will be kept as it is.
Replacing R4 by R4 + R2 and then R2 by 2R2+ R1 the matrix will be

 2 −2 3 4 −1
 
 0 0 7 14 3  .
 0 0 −1 −2 3 
 
0 0 4 8 2 

Replacing R4 by R4 + 4R3 and then replacing R3 by 7R3 + R2 the matrix will be

 2 −2 3 4 −1 
 
0 0 7 14 3 
.
0 0 0 0 24 
 
0 0 0 0 14 

Finally replacing R4 by R4 − R3 the resultant matrix will be in echelon form

as given below:
Rank of a Matrix and Solution of a Linear System

 2 −2 3 4 −1 
 
0 0 7 14 3 
.
0 0 0 0 24 
 
0 0 0 0 0

Now there are three non-zero rows in the echelon form of the given matrix A.
Therefore rank of A is equal to 3.

2.4 Solution of a Linear System

Recall that a system of m linear equations in n variables x1, x2, . . . , xn will be of

the form
a11x1 + a12x2 + . . . + a1nxn = b1
a21x1 + a22x2 + . . . + a2nxn = b2
. . . .
am1x1 + am2x2 + . . . + amnxn = bm

where aij’s and bi’s are real or complex numbers.

By using matrix notation this system can be expressed as Ax = b, where A is the

m × n matrix

 a11 a12  a1n 

 
 a 21 a 22  a 2n 
A= ,
  
 
 a m1 a m2  a mn 

 b1 
 x1   
  b
x is the n × 1 matrix x =    , b is the m × 1 matrix  2  .
x    
 n  
 bm 
Rank of a Matrix and Solution of a Linear System

A system of linear equations has either (i) no solution or (ii) exactly one
solution or (iii) infinitely many solutions. The system is said to be consistent if
it has at least one solution, that is (ii) or (iii) of the above hold, and is
inconsistent if it has no solution.

The following theorem gives conditions for existence of solution of the system
Ax = b.

Theorem 2.4.1: Let Ax = b be a system of m linear equations on n variables.

Let the augmented matrix ϵA of A be (A b). Then

(i) The system is consistent if rank A = rank ϵA.

(ii) The system has a unique solution if rank A = rank ϵA = n.

(iii) The system has infinitely many solutions if rank A = rank ϵA =k < n.

Remark 2.4.1: Recall that if b = 0 then the system Ax = 0 is called

homogeneous. In this case A = ϵA and so from the above theorem a
homogeneous system is always consistent. In fact (x1, x2, . . . , xn) = (0, 0, . . . ,
0) is always a solution of Ax = 0.

2.5 Gauss Elimination Method for Solving a System

Gauss-Elimination method is a matrix method constantly used to solve large

systems of linear equations. The main steps in this method are as follow:

1. Consider the augmented matrix of the system.

2. Convert the augmented matrix in to row echelon form. Decide whether the
system is consistent or not. If yes then go to the next step, stop otherwise.
Rank of a Matrix and Solution of a Linear System

3. Write the system of equations corresponding to the matrix in echelon form

obtained in step 2. Now this system is either solvable by back-substitution or
having some free variables (variables which do not occur at the beginning of
any equation of the system in this step) for which we assign arbitrary
real/complex value and then solve the system.

We explain the above method through some examples.

Example 2.5.1: Consider the system of linear equations

2x − 2y + 3z + 4u = − 1

− x + y + 2z + 5u = 3

− z − 2u = 3

x − y + 2z + 3u = 0

The augmented matrix of this system is

 2 −2 3 4 −1
 
 −1 1 2 5 2 
 0 0 −1 −2 3 .
 
 1 −1 2 3 0 

Notice that this is the same matrix A appears in Example 3.1. So its row echelon
form will be

2 −2 3 4 −1 
 
0 0 7 14 3 
0 0 0 0 24  .
 
0 0 0 0 0 
Rank of a Matrix and Solution of a Linear System

Observe that the rank of the co-efficient matrix is 2 and that of the augmented
matrix is 3. Therefore according to Theorem 4.1(i) the given system is
inconsistent.

Example 2.5.2: Here we shall solve the system

2x + y − 2z = 10

3x + 2y + 2z = 1

5x + 4y + 3z = 4

2 1 −2 10 
 
The augmented matrix is  3 2 2 1 .
5 4 
 4 3 

2 1 −2 10 
 
Row echelon form of this matrix is  0 1 10 −28  .
0 −14 42 
 0

Notice that 1st three columns is the row echelon form of the co-efficient matrix
and its rank is equal to three which is same as the rank of the augmented matrix.
Therefore the system is consistent and since the number of variables is also
equal to three from Theorem 4.1(ii) the system has a unique solution.

The system corresponding to the echelon form of the augmented matrix is:

2x + y − 2z = 10

y + 10z = − 28

− 14z = 42
Rank of a Matrix and Solution of a Linear System

From the last equation we get z = − 3. Then by back substitution we get y = 2

and x = 1 from the 2nd and 1st equations respectively. Hence (1,2,-3) is the
unique solution of the given system.

Example 2.5.3: Here we shall solve the system

x + 2y − 3z = 6

2x − y + 4z = 2

4x + 3y − 2z = 14

 1 2 −3 6 
 
Augmented matrix of the system is  2 −1 4 2 
 4 3 −2 14 
 

 1 2 −3 6 
 
and its row echelon form is  0 −5 10 −10  .
0 0 0 0 


The rank of the co-efficient matrix and the rank of the augmented matrix are
same and is equal to 2 which is less than the number of variables. Therefore the
system has infinite number of solutions. From the row echelon form of the
augmented matrix the system will be

x + 2y − 3z = 6

− 5y + 10z = − 10

Here z is the free variable. So it can take any real value. Let z = α, α is a real
number. Then from the second equation of the above system y = 2 + 2α and
Rank of a Matrix and Solution of a Linear System

then from the first equation x = 2 – α. Hence the set of all solutions of the
system is
{(2 – α, 2 + 2α, α) : α ϵ }.

2.6 Conclusions

In this lecture we have observed that homogeneous systems are always

consistent. We shall see in an other lecture that the set of all solutions of a
homogeneous system has linearity property and therefore these systems are of
special interest. In a subsequent lecture we shall learn about the linearity
property of sets, which is the basic thing of the subject Linear Algebra.

Keywords: Echelon form of matrices, rank of a matrix, solution of linear

system, Gauss elimination method.

Suggested Readings:

Linear Algebra, Kenneth Hoffman and Ray Kunze, PHI Learning pvt. Ltd.,
New Delhi, 2009.

Linear Algebra, A. R. Rao and P. Bhimasankaram, Hindustan Book Agency,

New Delhi, 2000.

Linear Algebra and Its Applications, Fourth Edition, Gilbert Strang, Thomson
Books/Cole, 2006.

Matrix Methods: An Introduction, Second Edition, Richard Bronson, Academic

press, 1991.