Lecture 3

Elementary Matrices & Row Reduced Echelon Form

Definition 1. Elementary row/column operations: Let A be an m×n matrix and R1 , . . . , Rm denote
the rows of A and C1 , . . . , Cn denote the columns of A. Then an elementary row(column) operation is a
map from Mm×n (F) to itself which is any one of the following three types:

1. Multiplying the i-th row(column) by a nonzero scalar λ ∈ F \ {0} denoted by Ri → λRi (Ci → λCi ).

2. Interchanging the i-th row(column) and the j-th row(column) denoted by Ri ↔ Rj (Ci ↔ Cj ).

3. For i 6= j, replacing the i-th row(column) by the sum of the i-th row(column) and µ multiple of the
j-th row(column) denoted by Ri → Ri + µRj (Ci → Ci + µCj ).

A row operation is a map from Mm×n (F) to itself which is a composition of finitely many elementary
row operations.

Remark 2. 1. Every elementary row operation is invertible.

1.The inverse of Ri → λRi is Ri → λ1 Ri ;
2. The inverse of Ri ↔ Rj is Ri ↔ Rj (self inverse, i.e., inverse of itself );
3. The inverse of Ri → Ri + µRj is Ri → Ri − µRj .

2. Let ρ be an elementary row operation and A ∈ Mm×n (F). Then ρ(A) = ρ(Im ) A, where Im is the
m × m identity matrix.

Definition 3. Elementary Matrix: Let Im denote the m × m identity matrix. A matrix obtained by
performing an elementary row operation on Im is called an elementary matrix. Therefore, there are three
types of elementary matrices:
1. Ei (λ), obtained by multiplying the i-th row by a nonzero scalar λ ∈ F \ {0} of Im .
2. Eij , obtained by interchanging the i-th row and the j-th row of Im .
3. Eij (µ), obtained by replacing the i-th row by the sum of the i-th row and µ multiple of the j-th row of
Im .

Example 4.    
" # 0 1 0 0 1 0
1 1    
M1 = M2 = 1 0 0 M3 = 0 0 1
0 1    
0 0 1 1 0 0
In the above, M1 , M2 are elementary matrices but M3 is not.

1
Remark 5. 1. Performing an elementary row operation on a matrix A is same as pre multiplication of
the respective elementary matrix to A.
2. Performing an elementary column operation on a matrix A is same as post multiplication of the re-
spective elementary matrix to A.

Definition 6. Row-equivalent matrices: Let A and B be two m × n matrices over a field F. Then B
is said to be row-equivalent to A if B is obtained from A by performing a finite sequence of elementary
row operations.

Theorem 7. If A and B are row equivalent matrices, the homogeneous systems of linear equations
Ax = 0 and Bx = 0 have exactly the same solutions.

Proof: It is given that A and B are row equivalent, that is, there exist elementary row operations,
ρ1 , ρ2 , . . . , ρk , such that B = ρk ◦ · · · ◦ ρ2 ◦ ρ1 (A), equivalently,

A = A0 → A1 → · · · → Ak = B.

It is enough to show that Aj x = 0 and Aj+1 x = 0 have the same solutions. In words, elementary operations
do not make any change to the solution set. Let k = 1. Then B = ρ(A) ⇒ B = ρ(Im )A. If Ax = 0, then
Bx = ρ(Im )Ax = ρ(Im )0 = 0. Similarly, if Bx = 0, then Ax = ρ−1 (B)x ⇒ Ax = ρ−1 (Im )Bx = 0.

Definition 8. Row-equivalent systems: The systems of linear equations Ax = b and Cx = d are said
to be row equivalent if their respective augmented matrices, (A|b) and (C|d) are row equivalent.

Theorem 9. Let Ax = b and Cx = d be two row equivalent linear systems. Then they have the same
solution set.

Proof: Let E1 , E2 , . . . , Ek be the elementary matrices such that E1 E2 · · · Ek (A|b) = (C|d). Suppose
y is a solution of Ax = b, i.e., Ay = b. Then Cy = E1 E2 · · · Ek Ay = E1 E2 · · · Ek b = d. Similarly, we can
prove that any solution of Cx = d is a solution of Ax = b..

Definition 10. Row echelon form: A form of a matrix satisfying the following properties is called row
echelon matrix.

1. Every zero-row of A (row which has all its entries 0) occurs below every non-zero row (which has a
non-zero entry);

2. Suppose the matrix has r nonzero rows (and remaining m−r rows are zero). If the leading coefficient
(the first non-zero entry) of i-th row (1 ≤ i ≤ r) occurs in the ki -th column, then k1 < k2 < · · · < kr ,

2
 that is, the leading coefficient of each row after the first is positioned to the right of the leading
coefficient of the previous row.

Definition 11. Row reduced echelon form: A form of a matrix satisfying the following properties is
called row reduced echelon form (in short RRE) or reduced echelon form:

1. The matrix is in row echelon form;

2. The leading coefficient of each row is 1;

3. All other elements in a column that contains a leading coefficient are zero.

Remark 12. 1. The process of computing row echelon form of a matrix by performing row operations
is called Gaussian elimination.

2. The process of computing row-reduced echelon form of a matrix by applying row operations is called
Gaussian-Jordan elimination.

3. Every matrix is row equivalent to a row-reduced echelon matrix. In fact, row-reduced echelon form
of a matrix is unique.


0 0 4 1
 
0 3 0 1
Example 13. Find the RRE form of 
 .
0 0 0 0


0 4 2 0
       
0 0 4 1 0 0 4 1 0 3 0 1 0 1 0 1/3
       
0 3 0 1 R3 ↔R4 0 3 0 1 R1 ↔R2 0 0 4 1 R1 → 31 R1 0 0 4 1 
Solution: 0 0 0 0 ∼ 0 4 2 0 ∼ 0 4 2 0 ∼ 0 4
       
       2 0 
0 4 2 0 0 0 0 0 0 0 0 0 0 0 0 0
     
0 1 0 1/3 0 1 0 1/3 0 1 0 1/3
     
0 0 4 1  R2 → 14 R2 0 0 1 1/4
R3 →R3 −4R1   R3 →R3 −2R2 0 0 1 1/4 
∼  ∼   ∼  
0
 0 2 −4/3 
0 0 2 −4/3
 
0 0 0 −11/6
 
0 0 0 0 0 0 0 0 0 0 0 0
   
0 1 0 1/3 0 1 0 0
   
R3 → −6 R 3 0 0 1 1/4 R1 →R1 −R3 , R2 →R2 − 14 R3 0 0 1 0
∼11  0 0
 ∼ 0 0 0 1 .
 
 0 1   
0 0 0 0 0 0 0 0

Remark 14. Consider a system of m linear equations in n unknowns Ax = 0. Let R be the RRE form
of A with r non-zero rows. The number of leading columns (column which contains a leading coefficient)

3
is r. We call the variables associated with leading columns leading variables or dependent variables.
The variables other than the dependent variables are called free variables or independent variables.
Note that, either there is no equation in which the free variable appears or it appears with at least one
another variable.

For instance, if we consider a homogeneous system of linear equation Ax = 0, where A is as in Example

13. Then the columns 2,3 and 4 are leading columns and hence, x2 , x3 and x4 are leading variables or
dependent variable, and x1 is an independent (free) variable which is not appearing in any of the reduced
equations.
 
1 2 0 3
 
0 0 1 1
Now consider a homogeneous system corresponding to the matrix  1 2 1 4 . Then the RRE

 
1 2 −1 2
 
1 2 0 3
 
0 0 1 1
form of A is 
0
 . Here, x1 , x3 are leading or dependent variables and x2 , x4 are free or inde-
 0 0 0
0 0 0 0
pendent variables.

Theorem 15. If A is an m × n matrix and m < n, then the homogeneous system of linear equations,
Ax = 0, has a non trivial solution.

Proof: Let R be the row-reduced echelon form of A. Then the systems Ax = 0 and Rx = 0 have
the same solution set. If r is the number of non-zero rows in R, then r ≤ m so that r < n, equivalently,
n − r > 0. Thus, there exists at least one independent variable (free variable) for the system Rx = 0,
and hence Rx = 0 has a non-trivial solution. 

Theorem 16. Let A ∈ Mn×n (F). The matrix A is row equivalent to the n × n identity matrix if and only
if the system of equations Ax = 0 has only the trivial solution.

Proof: If A is row equivalent to the n × n identity matrix In , then Ax = 0 and In x = 0 have only the
trivial solution. Conversely, suppose Ax = 0 has only the trivial solution and R is the RRE form of A.
Let r be the number of non-zero rows of R. Then r ≤ n. Note that Rx = 0 has only the trivial solution
(as Ax = 0 and Rx = 0 are row equivalent) so that r ≥ n. Hence, r = n and R = In .