Introduction to Linear Algebra

Row Reduction and Echelon Forms

Pangun Park

Chungnam National University

pgpark@[Link]

Pangun Park (CNU) 1 / 21

Row Echelon Form

Let’s come up with an algorithm for turning an arbitrary matrix into a “solved” matrix.
What do we mean by “solved”?

A matrix is in row echelon form if

All zero rows are at the bottom.
Each leading nonzero entry of a row is to the right of the leading entry of the row
above.
Below a leading entry of a row, all entries are zero.

Picture:
 
⋆ ⋆ ⋆ ⋆ ⋆
 0 ⋆ ⋆ ⋆ ⋆ ⋆ = any number
 
 0 0 0 ⋆ ⋆ ⋆ = any nonzero number
0 0 0 0 0

Definition
A pivot ⋆ is the first nonzero entry of a row of a matrix. A pivot column is a column
containing a pivot of a matrix in row echelon form.

Pangun Park (CNU) 2 / 21

Reduced Row Echelon Form

A matrix is in reduced row echelon form if it is in row echelon form, and in addition
The pivot in each nonzero row is equal to 1.
Each pivot is the only nonzero entry in its column.

Picture:
 
1 0 ⋆ 0 ⋆
0 1 ⋆ 0 ⋆ ⋆ = any number
 
0 0 0 1 ⋆ 1 = pivot
0 0 0 0 0

Note: Echelon forms do not care whether or not a column is augmented. Just ignore the
vertical line.

Question
Can every matrix be put into reduced row echelon form only using row operations?
Answer : Yes! Stay tuned.

Pangun Park (CNU) 3 / 21

Reduced Row Echelon Form: Continued

Why is this the “solved” version of the matrix?

 
1 0 0 1
 0 1 0 −2 
0 0 1 3

is in reduced row echelon form. It translates into

x= 1
y = −2
z= 3

which is clearly the solution.

Pangun Park (CNU) 4 / 21

Poll

Poll
Which of the following matrices are in reduced row echelon form?
! " ! "
1 0 0 0 0
A. B.
0 2 0 0 0
 
0
1  ) * ) *
C.  
 D. 0 1 0 0 E. 0 1 8 0
0
0
! "
1 17 0
F.
0 0 1

Answer : B, D, E, F
Note that A is in row echelon form though.

Pangun Park (CNU) 5 / 21

Reduced Row Echelon Form

Theorem
Every matrix is row equivalent to one and only one matrix in reduced row echelon form.

We’ll give an algorithm, called row reduction, which demonstrates that every matrix is
row equivalent to at least one matrix in reduced row echelon form.

Note : Like echelon forms, the row reduction algorithm does not care if a column is
augmented: ignore the vertical line when row reducing.

The uniqueness statement is interesting - it means that, no matter how you row reduce,
you always get the same matrix in reduced row echelon form. (Assuming you only do the
three legal row operations.) (And you don’t make any arithmetic errors.)

Maybe you can figure out why it’s true!

Pangun Park (CNU) 6 / 21

Row Reduction Algorithm

Step 1a Swap the 1st row with a lower one so a leftmost nonzero entry is in 1st row
(if necessary).
Step 1b Scale 1st row so that its leading entry is equal to 1.
Step 1c Use row replacement so all entries below this 1 are 0.
Step 2a Swap the 2nd row with a lower one so that the leftmost nonzero entry is in
2nd row.
Step 2b Scale 2nd row so that its leading entry is equal to 1.
Step 2c Use row replacement so all entries below this 1 are 0.
Step 3a Swap the 3rd row with a lower one so that the leftmost nonzero entry is in
3rd row.
etc.
Last Step Use row replacement to clear all entries above the pivots, starting with the
last pivot (to make life easier).
Example
 
0 −7 −4 2
 2 4 6 12 
3 1 −1 −2

Pangun Park (CNU) 7 / 21

Row Reduction: Example

   
0 −7 −4 2 R1 ←→ R2 2 4 6 12
 2 4 6 12   0 −7 −4 2
3 1 −1 −2 3 1 −1 −2
Step 1a: Row swap to make this nonzero. Step 1b: Scale to make this 1.
 
R 1 = R1 ÷ 2 1 2 3 6
 0 −7 −4 2
3 1 −1 −2
Step 1c: Subtract a multiple of
the first row to clear this.
 
R3 = R3 − 3R1 1 2 3 6
 0 −7 −4 2
0 −5 −10 −20
 
R2 ←→ R3 1 2 3 6
Optional: swap rows 2 and 3 to  0 −5 −10 −20 
make Step 2b easier later on.
0 −7 −4 2

Pangun Park (CNU) 8 / 21

Row Reduction: Example, continued

   
1 2 3 6 R2 = R2 ÷ −5 1 2 3 6
0 −5 −10 −20  0 1 2 4
0 −7 −4 2 0 −7 −4 2
Step 2a: This is already nonzero. Step 2c: Add 7 times
Step 2b: Scale to make this 1. the second row to clear this.
 
(There are no fractions because R3 = R3 + 7R2 1 2 3 6
of the optional step before.) 0 1 2 4
0 0 10 30

Note: Step 2 never messes up the first (nonzero) column of the matrix,
because it looks like this:
 
1 ⋆ ⋆ ⋆
“Active” row  0 ⋆ ⋆ ⋆ 
0 ⋆ ⋆ ⋆

Pangun Park (CNU) 9 / 21

Row Reduction: Example, continued
   
1 2 3 6 R3 = R3 ÷ 10 1 2 3 6
0 1 2 4 0 1 2 4
0 0 10 30 0 0 1 3
Step 3a: This is already nonzero.
Step 3b: Scale to make this 1.

Note: Step 3 never messes up the columns to the left.

Note: The matrix is now in row echelon form!
   
1 2 3 6 R2 = R2 − 2R3 1 2 3 6
0 1 2 4  0 1 0 −2 
0 0 1 3 0 0 1 3
 
Last step: Add multiples of R1 = R1 − 3R3 1 2 0 −3
the third row to clear these.  0 1 0 −2 
0 0 1 3
Last step: Add −2 times
the third row to clear this.
 
R1 = R1 − 2R2 1 0 0 1
0 1 0 −2 
0 0 1 3

Pangun Park (CNU) 10 / 21

Row Reduction: Example, continued

Success! The reduced row echelon form is

 
1 0 0 1 x = 1
 0 1 0 −2  −→ y = −2
0 0 1 3 z = 3

Pangun Park (CNU) 11 / 21

Recap

Get 
a 1 here Clear down Get a 1 here
      Clear down 
⋆ ⋆ ⋆ ⋆ 1 ⋆ ⋆ ⋆ 1 ⋆ ⋆ ⋆ 1 ⋆ ⋆ ⋆
⋆ ⋆ ⋆ ⋆ ⋆ ⋆ ⋆ ⋆ 0 ⋆ ⋆ ⋆ 0 1 ⋆ ⋆
       
⋆ ⋆ ⋆ ⋆ ⋆ ⋆ ⋆ ⋆ 0 ⋆ ⋆ ⋆ 0 ⋆ ⋆ ⋆
⋆ ⋆ ⋆ ⋆ ⋆ ⋆ ⋆ ⋆ 0 ⋆ ⋆ ⋆ 0 ⋆ ⋆ ⋆

(maybe these are already zero) Get a 1 

here  Clear down Matrix is in REF
   
1 ⋆ ⋆ ⋆ 1 ⋆ ⋆ ⋆ 1 ⋆ ⋆ ⋆ 1 ⋆ ⋆ ⋆
0 1 ⋆ ⋆ 0 1 ⋆ ⋆ 0 1 ⋆ ⋆ 0 1 ⋆ ⋆
       
0 0 0 ⋆ 0 0 0 ⋆ 0 0 0 1 0 0 0 1
0 0 0 ⋆ 0 0 0 ⋆ 0 0 0 ⋆ 0 0 0 0

Clear up Clear up Matrix is in RREF

     
1 ⋆ ⋆ ⋆ 1 ⋆ ⋆ 0 1 0 ⋆ 0
0
 1 ⋆ ⋆
0
 1 ⋆ 0
0
 1 ⋆ 0 Profit?
0 0 0 1 0 0 0 1 0 0 0 1
0 0 0 0 0 0 0 0 0 0 0 0

Pangun Park (CNU) 12 / 21

Row Reduction: Another example

The linear system

! "
2x + 10y = −1 2 10 −1
gives rise to the matrix
3 15 2
.
3x + 15y = 2
Let’s row reduce it: [interactive row reducer]
R1 = R1 ÷ 2 − 12
! " ! "
2 10 −1 1 5
(Step 1b)
3 15 2 3 15 2
R = R − 3R − 12
! "
2 2 1 1 5
7 (Step 1c)
0 0 2
2
R2 = R 2 ×
1 5 − 12
! "
7
(Step 2b)
0 0 1
1
R1 = R1 + R
2 2
! "
1 5 0
(Step 2c)
0 0 1

The row reduced matrix

! "
1 5 0 corresponds to the x + 5y = 0
0 0 1 inconsistent system 0 = 1.

Pangun Park (CNU) 13 / 21

Inconsistent Matrices

Question What does an augmented matrix in reduced row echelon form look like, if its
system of linear equations is inconsistent?

Answer
 
1 0 ⋆ ⋆ 0
 0 1 ⋆ ⋆ 0 
0 0 0 0 1

An augmented matrix corresponds to an inconsistent system of equations if and only if

the last (i.e., the augmented) column is a pivot column.

Pangun Park (CNU) 14 / 21

Another Example

The linear system

! "
2x + y + 12z = 1 2 1 12 1
gives rise to the matrix
1 2 9
.
x + 2y + 9z = −1 −1
Let’s row reduce it: [interactive row reducer]
R1 ←→ R2
! " ! "
2 1 12 1 1 2 9 −1
(Optional)
1 2 9 −1 2 1 12 1
R2 = R2 − 2R1
! "
1 2 9 −1
(Step 1c)
0 −3 −6 3
R2 = R2 ÷ −3
! "
1 2 9 −1
(Step 2b)
0 1 2 −1
R1 = R1 − 2R2
! "
1 0 5 1
(Step 2c)
0 1 2 −1

The row reduced matrix

! " #
1 0 5 1 corresponds to the x + 5z = 1
0 1 2 −1 linear system y + 2z = −1

Pangun Park (CNU) 15 / 21

Another Example: Continued

The system

x + 5z = 1
y + 2z = −1

comes from a matrix in reduced row echelon form. Are we done? Is the system solved?
Yes! Rewrite:

x = 1 − 5z
y = −1 − 2z

For any value of z, there is exactly one value of x and y that makes the equations true.
But z can be anything we want!
So we have found the solution set: it is all values x, y , z where

x = 1 − 5z
y = −1 − 2z for z any real number.
(z = z)

This is called the parametric form for the solution.

For instance, (1, −1, 0) and (−4, −3, 1) are solutions.
Pangun Park (CNU) 16 / 21
Free Variables
Definition
Consider a consistent linear system of equations in the variables x1 , . . . , xn . Let A be a
row echelon form of the matrix for this system.

We say that xi is a free variable if its corresponding column in A is not a pivot column.

Important
1 You can choose any value for the free variables in a (consistent) linear system.
2 Free variables come from columns without pivots in a matrix in row echelon form.

In the previous example, z was free because the reduced row echelon form matrix was
 
1 0 5 4
0 1 2 −1
In this matrix:
 
1 ⋆ 0 ⋆ ⋆
0 0 1 ⋆ ⋆
the free variables are x2 and x4 . (What about the last column?)
Pangun Park (CNU) 17 / 21
One More Example

The reduced row echelon form of the matrix for a linear system in x1 , x2 , x3 , x4 is

 
1 0 0 3 2
0 0 1 4 −1

The free variables are x2 and x4 : they are the ones whose columns are not pivot columns.
This translates into the system of equations

x1 +3x4 = 2 x1 = 2 − 3x4
−→
x3 +4x4 = −1 x3 = −1 − 4x4

What happened to x2 ? What is it allowed to be? Anything! The general solution is

(x1 , x2 , x3 , x4 ) = (2 − 3x4 , x2 , −1 − 4x4 , x4 )

for any values of x2 and x4 . For instance, (2, 0, −1, 0) is a solution (x2 = x4 = 0), and
(5, 1, 3, −1) is a solution (x2 = 1, x4 = −1).

The boxed equation is called the parametric form of the general solution to the system
of equations. It is obtained by moving all free variables to the =.

Pangun Park (CNU) 18 / 21

Yet Another Example

The linear system

x + y + z = 1 has matrix form 1 1 1 1 .

This is in reduced row echelon form. The free variables are y and z. The parametric form
of the general solution is

x = 1 − y − z.

Rearranging:

(x, y , z) = (1 − y − z, y , z),

where y and z are arbitrary real numbers. This was an example in the second lecture!

Pangun Park (CNU) 19 / 21

Poll

Poll
It it possible for a system of linear equations to have exactly two solution?

Pangun Park (CNU) 20 / 21

Summary
There are three possibilities for the reduced row echelon form of the augmented matrix of
a linear system.
1 The last column is a pivot column. In this case, the system is inconsistent. There

are zero solutions, i.e. the solution set is empty.

Picture:
 
1 0 0
 0 1 0 
0 0 1
2 Every column except the last column is a pivot column. In this case, the system has
a unique solution.
Picture:
 
1 0 0 ⋆
 0 1 0 ⋆ 
0 0 1 ⋆
3 The last column is not a pivot column, and some other column isn’t either. In this
case, the system has infinitely many solutions, corresponding to the infinitely many
possible values of the free variable(s).
Picture:
 
1 ⋆ 0 ⋆ ⋆
0 0 1 ⋆ ⋆
Pangun Park (CNU) 21 / 21