MT-1004

Linear Algebra
Fall 2021
Lecture # 2

National University of Computer and Emerging Sciences

September 5, 2021
Row Equivalence

Important
The process of doing row operations to a matrix does
not change the solution set of the corresponding linear
equations!
Row Equivalence

Important
The process of doing row operations to a matrix does
not change the solution set of the corresponding linear
equations!

Definition
Two matrices are called row equivalent if one can be obtained from the other
by doing some number of elementary row operations.
Row Equivalence

Important
The process of doing row operations to a matrix does
not change the solution set of the corresponding linear
equations!

Definition
Two matrices are called row equivalent if one can be obtained from the other
by doing some number of elementary row operations.

So the linear equations of row-equivalent matrices have the same solution set.
A Bad Example
Example
Solve the system of equations
x + y =2
3x + 4y = 5
4x + 5y = 9
A Bad Example
Example
Solve the system of equations
x + y =2
3x + 4y = 5
4x + 5y = 9

Let’s try doing row operations:

A Bad Example
Example
Solve the system of equations
x + y =2
3x + 4y = 5
4x + 5y = 9

Let’s try doing row operations:

 
1 1 2
 3 4 5
4 5 9
A Bad Example
Example
Solve the system of equations
x + y =2
3x + 4y = 5
4x + 5y = 9

Let’s try doing row operations:

 
1 1 2
First clear these by  3 4 5
subtracting multiples
of the first row. 4 5 9
A Bad Example
Example
Solve the system of equations
x + y =2
3x + 4y = 5
4x + 5y = 9

Let’s try doing row operations:

 
1 1 2
First clear these by  3 4 5
subtracting multiples
of the first row. 4 5 9
 
1 1 2
 0 1 −1 
0 1 1
A Bad Example
Example
Solve the system of equations
x + y =2
3x + 4y = 5
4x + 5y = 9

Let’s try doing row operations:

 
1 1 2
First clear these by  3 4 5
subtracting multiples
of the first row. 4 5 9
 
1 1 2
Now clear this by  0 1 −1 
subtracting 0 1 1
the second row.
A Bad Example
Continued
 
1 1 2 translates into
x +y = 2
0 1 −1  y = −1
0 0 2 0= 2
A Bad Example
Continued
 
1 1 2 translates into
x +y = 2
0 1 −1  y = −1
0 0 2 0= 2

In other words, the original equations

x + y =2 x +y = 2
3x + 4y = 5 have the same solutions as y = −1
4x + 5y = 9 0= 2
A Bad Example
Continued
 
1 1 2 translates into
x +y = 2
0 1 −1  y = −1
0 0 2 0= 2

In other words, the original equations

x + y =2 x +y = 2
3x + 4y = 5 have the same solutions as y = −1
4x + 5y = 9 0= 2

But the latter system obviously has no solutions (there is no way to make them
all true), so our original system has no solutions either.
A Bad Example
Continued
 
1 1 2 translates into
x +y = 2
0 1 −1  y = −1
0 0 2 0= 2

In other words, the original equations

x + y =2 x +y = 2
3x + 4y = 5 have the same solutions as y = −1
4x + 5y = 9 0= 2

But the latter system obviously has no solutions (there is no way to make them
all true), so our original system has no solutions either.

Definition
A system of equations is called inconsistent if it has no solution. It is
consistent otherwise.
1.2: Row Echelon Form
Let’s come up with an algorithm for turning an arbitrary matrix into a “solved”
matrix.
1.2: 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”?
1.2: 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

1.2: 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

1. All zero rows are at the bottom.
1.2: 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

1. All zero rows are at the bottom.
2. Each leading nonzero entry of a row is to the right of the
leading entry of the row above.
1.2: 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

1. All zero rows are at the bottom.
2. Each leading nonzero entry of a row is to the right of the
leading entry of the row above.
3. Below a leading entry of a row, all entries are zero.
1.2: 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

1. All zero rows are at the bottom.
2. Each leading nonzero entry of a row is to the right of the
leading entry of the row above.
3. 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
1.2: 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

1. All zero rows are at the bottom.
2. Each leading nonzero entry of a row is to the right of the
leading entry of the row above.
3. 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 in row echelon form.
Reduced Row Echelon Form

A matrix is in reduced row echelon form if it is in row echelon

form, and in addition,
Reduced Row Echelon Form

A matrix is in reduced row echelon form if it is in row echelon

form, and in addition,
4. The pivot in each nonzero row is equal to 1.
Reduced Row Echelon Form

A matrix is in reduced row echelon form if it is in row echelon

form, and in addition,
4. The pivot in each nonzero row is equal to 1.
5. Each pivot is the only nonzero entry in its column.
Reduced Row Echelon Form

A matrix is in reduced row echelon form if it is in row echelon

form, and in addition,
4. The pivot in each nonzero row is equal to 1.
5. 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
Reduced Row Echelon Form

A matrix is in reduced row echelon form if it is in row echelon

form, and in addition,
4. The pivot in each nonzero row is equal to 1.
5. 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.
Reduced Row Echelon Form

A matrix is in reduced row echelon form if it is in row echelon

form, and in addition,
4. The pivot in each nonzero row is equal to 1.
5. 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?
Reduced Row Echelon Form

A matrix is in reduced row echelon form if it is in row echelon

form, and in addition,
4. The pivot in each nonzero row is equal to 1.
5. 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.
Reduced Row Echelon Form
Continued

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

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,
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.
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.

But what happens if there are fewer pivots than rows?

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.

But what happens if there are fewer pivots than rows? . . . parametrized
solution set (later).
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. 
0
 D. 0 1 0 0 E. 0 1 8 0
0
 
1 17 0
F.
0 0 1
Reduced Row Echelon Form
Theorem
Every matrix is row equivalent to one and only one matrix in reduced row
echelon form.
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.
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.
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—

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, nomatter how you row
reduce, you always get the same matrix in reduced row echelon form.
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, nomatter how you row
reduce, you always get the same matrix in reduced row echelon form.
(Assuming you only do the three legal row operations.)
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, nomatter 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.)
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, nomatter 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!

Row Reduction Algorithm
Step 1a Swap the 1st row with a lower one so a leftmost nonzero entry is in 1st
row (if necessary).

Example  
0 −7 −4 2
2 4 6 12 
3 1 −1 −2
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.

Example  
0 −7 −4 2
2 4 6 12 
3 1 −1 −2
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 above and below this 1 are 0.

Example  
0 −7 −4 2
2 4 6 12 
3 1 −1 −2
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 above and below this 1 are 0.
Step 2a Cover the first row, swap the 2nd row with a lower one so that the
leftmost nonzero (uncovered) entry is in 2nd row; uncover 1st row.

Example  
0 −7 −4 2
2 4 6 12 
3 1 −1 −2
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 above and below this 1 are 0.
Step 2a Cover the first row, swap the 2nd row with a lower one so that the
leftmost nonzero (uncovered) entry is in 2nd row; uncover 1st row.
Step 2b Scale 2nd row so that its leading entry is equal to 1.

Example  
0 −7 −4 2
2 4 6 12 
3 1 −1 −2
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 above and below this 1 are 0.
Step 2a Cover the first row, swap the 2nd row with a lower one so that the
leftmost nonzero (uncovered) entry is in 2nd row; uncover 1st row.
Step 2b Scale 2nd row so that its leading entry is equal to 1.
Step 2c Use row replacement so all entries above and below this 1 are 0.

Example  
0 −7 −4 2
2 4 6 12 
3 1 −1 −2
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 above and below this 1 are 0.
Step 2a Cover the first row, swap the 2nd row with a lower one so that the
leftmost nonzero (uncovered) entry is in 2nd row; uncover 1st row.
Step 2b Scale 2nd row so that its leading entry is equal to 1.
Step 2c Use row replacement so all entries above and below this 1 are 0.
Step 3a Cover the first two rows, swap the 3rd row with a lower one so that the
leftmost nonzero (uncovered) entry is in 3rd row; uncover first two rows.

Example  
0 −7 −4 2
2 4 6 12 
3 1 −1 −2
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 above and below this 1 are 0.
Step 2a Cover the first row, swap the 2nd row with a lower one so that the
leftmost nonzero (uncovered) entry is in 2nd row; uncover 1st row.
Step 2b Scale 2nd row so that its leading entry is equal to 1.
Step 2c Use row replacement so all entries above and below this 1 are 0.
Step 3a Cover the first two rows, swap the 3rd row with a lower one so that the
leftmost nonzero (uncovered) entry is in 3rd row; uncover first two rows.
etc.

Example  
0 −7 −4 2
2 4 6 12 
3 1 −1 −2
Row Reduction
Example
   
0 −7 −4 2 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.
 
1 2 3 6
 0 −7 −4 2
3 1 −1 −2
Step 1c: Subtract a multiple of
the first row to clear this.
 
1 2 3 6
0 −7 −4 2
0 −5 −10 −20
Optional: swap rows 2 and 3 to
make Step 2b easier later on.
 
1 2 3 6
0 −5 −10 −20 
0 −7 −4 2
Row Reduction
Example, continued
 
1 2 3 6
0 −5 −10 −20 
0 −7 −4 2
 
Step 2a: This is already nonzero. 1 2 3 6
Step 2b: Scale to make this 1. 0 1 2 4
(There are no fractions because 0 −7 −4 2
of the optional step before.)
Step 2c: Add multiples of
the second row to clear these.
 
1 0 −1 −2
0 1 2 4
0 −7 −4 2
 
1 0 −1 −2
0 1 2 4
0 0 10 30
Row Reduction
Example, continued
 
1 2 3 6
0 −5 −10 −20 
0 −7 −4 2
 
Step 2a: This is already nonzero. 1 2 3 6
Step 2b: Scale to make this 1. 0 1 2 4
(There are no fractions because 0 −7 −4 2
of the optional step before.)
Step 2c: Add multiples of
the second row to clear these.
 
1 0 −1 −2
0 1 2 4
0 −7 −4 2
 
1 0 −1 −2
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 ? ? ?
Row Reduction
Example, continued
 
1 0 −1 −2
0 1 2 4
0 0 10 30
Step 3a: This is already nonzero.
Step 3b: Scale to make this 1.
 
1 0 −1 −2
0 1 2 4
0 0 1 3
Step 3c: Add multiples of
the third row to clear these.
 
1 0 0 1
0 1 2 4
0 0 1 3
 
1 0 0 1
0 1 0 −2 
0 0 1 3
Row Reduction
Example, continued
 
1 0 −1 −2
0 1 2 4
0 0 10 30
Step 3a: This is already nonzero.
Step 3b: Scale to make this 1.
 
1 0 −1 −2
0 1 2 4
0 0 1 3
Step 3c: Add multiples of
the third row to clear these.
 
1 0 0 1
0 1 2 4
0 0 1 3
 
1 0 0 1
0 1 0 −2 
0 0 1 3
Note: Step 3 never messes up the columns to the left.
Row Reduction
Example, continued
 
1 0 −1 −2
0 1 2 4
0 0 10 30
Step 3a: This is already nonzero.
Step 3b: Scale to make this 1.
 
1 0 −1 −2
0 1 2 4
0 0 1 3
Step 3c: Add multiples of
the third row to clear these.
 
1 0 0 1
0 1 2 4
0 0 1 3
 
1 0 0 1
0 1 0 −2 
0 0 1 3
Note: Step 3 never messes up the columns to the left.
Success! The reduced row echelon form is
  
1 0 0 1 x = 1
 0 1 0 −2  =⇒ y = −2
Row Reduction
Example, continued
 
1 0 −1 −2
0 1 2 4
0 0 10 30
Step 3a: This is already nonzero.
Step 3b: Scale to make this 1.
 
1 0 −1 −2
0 1 2 4
0 0 1 3
Step 3c: Add multiples of
the third row to clear these.
 
1 0 0 1
0 1 2 4
0 0 1 3
 
1 0 0 1
0 1 0 −2 
0 0 1 3
Note: Step 3 never messes up the columns to the left.
Success! The reduced row echelon form is
  
1 0 0 1 x = 1
 0 1 0 −2  =⇒ y = −2
Row Reduction
Another example

The linear system

 
2x + 10y = −1 2 10 −1
gives rise to the matrix .
3x + 15y = 2 3 15 2
Row Reduction
Another example

The linear system

 
2x + 10y = −1 2 10 −1
gives rise to the matrix .
3x + 15y = 2 3 15 2
Let’s row reduce it:
Row Reduction
Another example

The linear system

 
2x + 10y = −1 2 10 −1
gives rise to the matrix .
3x + 15y = 2 3 15 2
Let’s row reduce it:
The row reduced matrix
 
1 5 0 corresponds to the x + 5y = 0
0 0 1 inconsistent system 0 = 1.
Inconsistent Matrices
Question
What does an augmented matrix in reduced row echelon form look like, if its
system of linear equations is inconsistent?
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
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.
Another Example
The linear system
 
2x + y + 12z = 1 2 1 12 1
gives rise to the matrix .
x + 2y + 9z = −1 1 2 9 −1
Another Example
The linear system
 
2x + y + 12z = 1 2 1 12 1
gives rise to the matrix .
x + 2y + 9z = −1 1 2 9 −1
Let’s row reduce it:
Another Example
The linear system
 
2x + y + 12z = 1 2 1 12 1
gives rise to the matrix .
x + 2y + 9z = −1 1 2 9 −1
Let’s row reduce it:
The row reduced matrix
  
1 0 5 1 corresponds to the x + 5z = 1
0 1 2 −1 linear system y + 2z = −1
Another Example
Continued

The system
x + 5z = 1
y + 2z = −1
comes from a matrix in reduced row echelon form.
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?
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!
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)
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.
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.
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.
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.
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.
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
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 .
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 .
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?)
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
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 :
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.
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
x3 + x4 = −1
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 + x4 = −1 x3 = −1 − 4x4.
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 + x4 = −1 x3 = −1 − 4x4.
What happened to x2 ? What is it allowed to be?
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 + x4 = −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).
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 + x4 = −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 right-hand side of the =.
Poll

Poll
Is it possible for a system of linear equations to
have exactly two solutions?
Summary
There are three possibilities for the reduced row echelon form of the augmented
matrix of a linear system.
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.
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:
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
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 ?
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 ? ?
With this knowledge you should be able to solve
Exercise 1.1 (1-32) & Exercise 1.2 (1-32)