Section 1.

Row Reduction
and Echelon Forms
 Row Reduction
and Echelon Forms
DEFINITION: The
DEFINITION Th process off simplifying
i lif i g a matrix
t i using
i g
row operations is called row reduction.
IIn thi
this section,
ti we refine
fi ththe method
th d off SSection
ti 11.11 iinto
t
a row reduction algorithm that will enable us to
analyze any system of linear equations.
equations
The algorithm we will discuss is known as Gaussian
elimination.
elimination
 Gaussian Elimination
Step 1: Find the left
left-most
most non
non-zero
zero column and select
a non-zero entry in that column. Interchange rows if
necessary so that this entry is in the first row.
Step 2: Divide the top row by the non-zero entry from
Step
p 1, there should now be a 1 in the top
p left entry.
y
This is called a “leading one” or a “pivot”.
Step 3: Eliminate all the entries beneath the pivot (in
the same column) by adding or subtracting multiples of
the top row.
Step 4: Cover up the top row and repeat Steps 1
through 3.
 Exercise 1
Use G
U Gaussian
i elimination
li i ti to t find
fi d a simpler
i l equivalent
i l t
system of linear equations. If possible, find the
solution(s) to the system.
system

⎧ x1 + x2 + 2 x3 = 9
⎪
⎨ 2 x1 + 4 x2 − 3 x3 = 1
⎪3 x + 6 x − 5 x = 0
⎩ 1 2 3
 Exercise 2
Use Gaussian elimination to find a simpler equivalent
system of linear equations. If possible, find the
solution(s) to the system.
system

⎧ x2 − 4 x3 = 8
⎪
⎨2 x1 − 4 x2 + 2 x3 = 10
⎪ 5 x − 8 x − 3x = 1
⎩ 1 2 3
 Row Echelon Form
DEFINITION: A matrix
DEFINITION t i is
i said
id tto be
b iin row echelon
h l
form if it satisfies the following three criteria:
1 The first non-zero
1. non zero entry in each row is a 1 (called a
leading one).
2 All rows containing only zeroes are grouped at the
2.
bottom of the matrix.
3. Each leading one is to the right of the leading one
in the row above it.
If Gaussian elimination is used properly to row reduce a
matrix, the end result will be in row echelon form.
 Exercise 3
Use G
U Gaussian
i elimination
li i ti to t row reduce
d the
th following
f ll i g
matrix. Check that resulting row equivalent matrix is in
row echelon form.
form

⎡1 1 1 1⎤
⎢ 2 3 0 0⎥
⎢ ⎥
⎢⎣4 3 − 1 2⎥⎦
 Exercise 4
Which
Whi h off the
th following
f ll i g matrices
ti are in
i row echelon
h l
form? If they are not, use Gaussian elimination to find
a row echelon form to which they are equivalent.
equivalent

⎡1 1 1 1 ⎤ ⎡0 0 0⎤
(A) ⎢0 1 0 0⎥ ⎢
(B) 1 2 1
⎥
⎢ ⎥ ⎢ ⎥
⎢⎣0 1 0 2⎥⎦ ⎢⎣0 0 1⎥⎦

⎡1 2⎤ ⎡0 1 2 2⎤
(C) ⎢ ⎥ (D) ⎢ ⎥
⎣0 4 ⎦ ⎣0 0 1 0 ⎦
 Uniqueness of
Row Echelon Form
NOTE: Using
NOTE Ui gG Gaussian
i elimination
li i ti tot row reduce
d a
matrix does not always result in the same row echelon
form matrix.
matrix That is,is the row echelon form is not
necessarily unique.
⎡1 1 1 1⎤
⎢0 1 0 0⎥
e.g. The matrix ⎢ ⎥ can be row reduced to either
⎢⎣0 1 0 2⎥⎦

⎡1 1 1 1⎤ ⎡1 1 1 1⎤
⎢0 1 0 0⎥ ⎢0 1 0 2⎥
⎢ ⎥ or ⎢ ⎥ depending on the steps taken.
taken
⎢⎣0 0 0 1⎥⎦ ⎣⎢0 0 0 1⎥⎦
 Gauss-Jordan
Gauss Jordan Elimination
IIn order
d tto row reduce
d a matrix
t i th
thatt iis iin row echelon
h l
form even further, we add the following additional
step to the Gaussian elimination algorithm:
Step 5: Beginning with the right-most leading one, use
row replacement operations to eliminate any non non-zero
zero
terms above that leading one. Repeat for each leading
one, as you move from right to left across the matrix.
With this additional step, this row reduction algorithm
is known as Gauss-Jordan elimination.
 Exercise 5
Use Gauss-Jordan
U G J d elimination
li i ti to t ffurther
th simplify
i lif the
th
following matrix, which is already in row echelon form.

⎡1 3 1 2 0⎤
⎢0 0 1 3 2⎥
⎢ ⎥
⎣⎢0 0 0 1 2⎥⎦
 Reduced Row Echelon Form
DEFINITION: A matrix
DEFINITION t i iis iin reduced
d d row
echelon form if it satisfies the following criteria:
11. The
Th matrix
t i is
i iin row echelon
h l form.
f
2. Each leading one is the only non-zero entry in its
column.
col

There are many different sequences

q of row operations
p
which can be applied to a matrix to find its RREF.
However, using Gauss-Jordan Elimination always leads
to the
h RREF off the
h matrix.
i
 Exercise 6
Determine
D t i whether
h th or nott each h off these
th row echelon
h l
form matrices are in reduced row echelon form.

⎡1 2 0 0 0 ⎤ ⎡1 0 1 0 1 ⎤
⎢
(A) 0 0 1 3 0 ⎥ ⎢0 1 5 0 2 ⎥
⎢ ⎥ (B) ⎢ ⎥
⎢⎣0 0 0 1 0 ⎥⎦ ⎢⎣0 0 0 1 3 ⎥⎦
 Uniqueness of the
Reduced Row Echelon Form
NOTE: For any given matrix
matrix, there is only one reduced
row echelon form matrix which can be obtained from it
through row reduction.
That is, the reduced row echelon form of a matrix is
unique.
unique
Thus, we refer to it as the reduced row echelon form
(RREF) off the
th matrix.
ti
 Exercise 7
Use Gauss-Jordan
U G J d elimination
li i ti to
t find
fi d the
th RREF off the
th
following matrix.
⎡4 8 12⎤
⎢0 1 4 ⎥
⎢ ⎥
⎢⎣1 2 1 ⎥⎦
 Exercise 8
Use Gauss-Jordan
U G J d elimination
li i ti to t find
fi d the
th solution(s)
l ti ( ) to
t
the following system of equations.

⎧ x1 − 5 x3 = 1
⎪
⎨ x2 + x3 = 4
⎪2 x − 3 x − 13 x = −10
⎩ 1 2 3
 Basic Variables
and Free Variables
DEFINITIONS:
Given a system of linear equations and its
corresponding augmented matrix,
matrix a free column is any
column in the coefficient part of its RREF that has no
leading one.
A free variable is a variable associated with a free
column. Non
Non-free
free variables are called basic variables.
Example: In Exercise 8, x1 and x2 are basic variables,
while x3 is a free variable.
variable
 Exercise 9
For each of the following systems of linear equations,
equations
find the RREF of its augmented matrix. Determine
which variables in the system
y are free and which are
basic.
(A) x + 2 x − x = 2 (B) ⎧ x + 2 x + x = 2
⎧ 1 1 2 3
⎪
2 3
⎪
⎨ x1 + 2 x2 + x3 = 4 ⎨ x1 − 2 x2 + x3 = 6
⎪x = 1 ⎪ x = −1
⎩ 3 ⎩ 2
 Existence and Uniqueness
Theorem
A system of linear equations is consistent if and only if
the row echelon form or row reduced echelon form of
the augmented
g matrix has no row of the form
[0 0 " 0 1] .
If a system of linear equations is consistent, then the
solution set contains either (i) a unique solution, when
th
there are no ffree variables,
i bl or (ii) infinitely
i fi it l many
solutions, when there is at least one free variable.
 Using Row Reduction to
Solve a Linear System
Step 1: Write the augmented matrix of the system.
Step 2: Use Gaussian elimination to obtain the row
echelon form. If the system is not consistent, stop.
Otherwise, go to Step 3.
Step
p 3: Continue row reduction to obtain the row
reduced echelon form of the augmented matrix.
Stepp 4: Write the system
y of equations
q corresponding
p g
to the matrix obtained in Step 3.
Stepp 5: Rewrite each non-zero equation
q from Step
p 4 so
that its one basic variable is expressed in terms of any
free variables appearing in the equation.
 Exercise 10
Solve the following systems of linear equations
using row reduction.
(A) {x1 − 2 x2 + x3 − 4 x4 = 3

(B) ⎧ x1 + 2 x2 + x3 = 2
⎪
⎨ x1 − 2 x2 + x3 = 6
⎪x + x = 0
⎩ 1 2

(C) ⎧ x1 + x2 = 6
⎨
⎩2 x1 + 2 x2 = 5