Lay, Lay, McDonald Linear Algebra and its Applications, 6th ed.

Page 0

Section 1.2

Row Reduction and Echelon Forms

Section 1.2,

Computational Linear Algebra

Brian Curtin
Page 0

University of South Florida

Section 1.2,
17
 Echelon Form
The leading entry of matrix row is its leftmost nonzero entry.
A matrix is in (row) echelon form if
1. All nonzero rows are above any rows of all zeros.
2. Each leading entry of a row is in a column to the right of the
Page 1

leading entry of the row above it.

L1 * *0 * * *0
Section 1.2,

L1 * * *0
0
L1
0 0
0 0 0 0 0 0 0
Page 1

A matrix is in reduced (row) echelon form whenever it is in echelon

form &
Section 1.2,

3. The leading entry in each nonzero row is 1.

18

4. Each leading is the only nonzero entry in its column

 Additional Vocabulary
Reduced echelon form makes it easy to read solutions to systems.
Often can answer question without full reduction.

By an echelon matrix, we mean a matrix in echelon form. The process

of putting a matrix in row echelon form is row reduction or Gaussian
Page 2

elimination.

By a reduced echelon matrix, we mean a matrix in reduced echelon

Section 1.2,

form. Process of putting a matrix in reduced row echelon form is

Gauss-Jordan elimination.
Page 2

Theorem
Uniqueness of Reduced Echelon Form: Each matrix is row equivalent
Section 1.2,

to one and only one reduced echelon matrix.

19

RREF(A) is the (unique) reduced row echelon form of A

 Row Reduction
Definition: A pivot position in a matrix A is a location in A that
corresponds to a leading 1 in the reduce echelon form of A
Definition: A pivot column of a matrix A is a column of A that
contains a pivot position
Page 3

Definition: A pivot of a matrix A is a nonzero number in pivot

position (that is used to create zeros via row operations).
Section 1.2,

Definition: The forward phase of row reduction refers to the process

described in steps 1–4 of row reduction, of producing a row
equivalent matrix that is in echelon form
Page 3

Definition: The backward phase refers to the process step 5 of row

reduction, making an echelon matrix a reduced echelon matrix.
Section 1.2,
20
 Example
1) First find leftmost nonzero column. It is a pivot column. Set pivot
position at the top row of this column.
 
0 1 1 →2 →3
 1 2 →1 0 2  Pivot position is top right of matrix.
Page 4


 2 4 1 →3 →2 
1 →4 →7 →1 →19
Section 1.2,

2) If necessary, swap rows to make the the entry in the pivot position
nonzero.
 
Page 4

1 2 →1 0 2
 0 1 1 →2 →3 
Swap rows 1 and 2 
 2 4

1 →3 →2 
Section 1.2,

1 →4 →7 →1 →19
21
 Example
3) Use row replacement to make all entries below pivot position 0.
 
1 2 →1 0 2
add →2↑ row 1 to row 3  0
 1 1 →2 →3 
add →1↑ row 1 to row 4  0 0 3 →3 →6 
Page 5

0 →6 →6 →1 →21

4) Cover everything in and left of the pivot column, in and above the
Section 1.2,

pivot row and apply 1–3 to the resulting submatrix. Repeat 4 until
there are no more nonzero rows.
The result is in ECHELON FORM!
Page 5

   
1 2 →1 0 2 1 2 →1 0 2
 0 1 1 →2 →3   ↓  0 1 1 →2 →3 

Section 1.2,

 
 0 0 3 →3 →6  R3 + 6R1  0 0 3 →3 →6 
22

0 →6 →6 →1 →21 0 0 0 →13 →39

echelon submatrix.
 Variables
5) Starting from rightmost pivot, work up and left.
a) Create zero entries above each pivot using replacement.
 b) If a pivot is not
 1, scale to make it 1.
1 0 0 0 →1
 0 1 0 0 2 
Page 6

 
 0 0 1 0 →1 
0 0 0 1 3
The result is in REDUCED ROW ECHELON FORM!
Section 1.2,

Page 6

Forward phase: 1–4

Backward phase: 5
Section 1.2,
23
 Variables
The basic or leading variables of a system of linear equations are those
corresponding to pivot columns in the coe!cient matrix of the system

The free variables of a system of linear equations are those which are
not basic, those corresponding to columns without pivots
Page 7

By a parametric description of a solution set, we mean a solution set

in which free variables act as parameters.
Section 1.2,

Page 7
Section 1.2,
24
 Variables Example
RRE matrices with fewer rows than columns (fewer equations than
variables)
' ( ' ( ' (
1 0 0 0 0 1 0 0 0 1 0 0
0 1 1 2 0 0 2 1 0 0 0 1
x1 = 0 x1 = anything 0=1
Page 8

x2 + x3 = 2 x2 = 0 no solution
x3 = 1/2
Section 1.2,

The variables in pivot columns are called basic

the other variables are called free
x2 = 2 → x3 : every choice of x3 gives a solution (determines x2 ).
Page 8
Section 1.2,
25
 Number of solutions

Theorem
Existence and Uniqueness Theorem: A linear system is consistent if
and only if the rightmost column of the augmented matrix is not a pivot
column–that is, if and only if an echelon form of the augmented matrix has
Page 9

no row of the form

[0 · · · 0 b] with b nonzero.
Section 1.2,

If a linear system is consistent, then the solution set contains either

(i) a unique solution, when there are no free variables, or
(ii) infinitely many solutions, when there is at least one free variable.
Page 9
Section 1.2,
26
 MATLAB

The RREF command returns reduced row echelon form of a matrix.

Enter an augmented matrix and ask MATLAB to reduce.
Page 10

 
1 1 0 0
A1 =  1 →2 →3 →3 ;
1 →1 2 6
Section 1.2,

R1=rref(A1)
 
1 0 0 1
R1 =  0 1 0 →1 
Page 10

0 0 1 2
Technology can take over the tedious tasks, but it is up to a human to
translate the problem for the machine, and interpret the output.
Section 1.2,
27