Transforming A Matrix To Reduced Row Echelon Form

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

Problem

Find an invertible matrix such that is in reduced row echelon form, where

[ ]

Solution
In order to understand the following process, you should have a good grasp of performing
elementary row operations on a matrix in order to transform it to the reduced row echelon (RRE)
form. For example, in the first step, you want the matrix

[ ]

to be transformed into

[ ]

by adding 2 times the first row to the second row and -1 time the first row to the third row. If you
don’t understand this, then you should review the Gauss-Jordan elimination algorithm. Numerous
tutorials are available online, including videos on YouTube.

When you transform a matrix to RRE form using the Gauss-Jordan elimination, what you are really
doing is multiplying the original matrix by a series of matrices, each one representing a step, such
as the one demonstrated above. In mathematical terms, if we let be the original matrix, be step
in the elimination process, and be the RRE form of , then

where is the single matrix that has the same effect as the series of elimination matrices . That
is,

Let’s begin with the original matrix:

[ ]

We will multiply by elimination matrices , adjusting the columns of from left to right putting
it in Row Echelon form, and then adjusting them from right to left, to arrive at the RRE form.
After the first elimination step , we want the resultant matrix to have

 in its first row: 1 time the first row of . This will have no change on the first row, since 1 is
the multiplicative identity.
 in its second row: 2 times the first row added to 1 time the second row. This will eliminate
the leading -2 in the second row.
 in row 3: -1 time the first row added to 1 time the third row.

So the first matrix to multiply by is

[ ]

Let’s analyze what the rows of really mean. Each row simply represents a combination of the
rows of . Row 1, ( ), essentially says, take one of row 1 of , and none of rows 2 and 3, add
them up, and the result will be row 1 of the new matrix.

[ ][ ] [ ]

In a more decomposed form, the resultant first row, , is


( )[ ] ( )[ ] ( )[ ]
[ ]

Similarly, the result of the second row, , is the sum of 2 times the first row and 1 time the second
row of .

[ ][ ] [ ]

( )[ ] ( )[ ] ( )[ ]
[ ]

The entire multiplication of the first step is shown below.

[ ][ ] [ ]

2
The next step is to eliminate the 2 in ( ) , i.e., the 2 in row 3 and column 2 of this first
intermediate matrix . Since we don’t need to make any modification to the first two rows, then
we only have this one elementary row operation to complete. We want to add -2 times the second
row to the third row, i.e., -2 times the second row plus 1 time the third row. This operation can be
expressed by the matrix

[ ]

Again, the rows of mean we want in our new matrix

 1 of the first row and none of the other two rows.


 1 of the second row and none of the other two rows.
 None of the first row, -2 of the second row, and 1 of the third row.

The result of the operation is

( ) [ ][ ] [ ]

In step three, we want to scale row 3 by a factor of .

( ) [ ][ ] [ ]
⁄ ⁄

We have reached the matrix in row echelon form. Observe that this is an upper triangular matrix.
Now we must transform it to reach the RRE form, by using similar steps to the ones taken above.

First, we want to eliminate the 12 and the 3 in, respectively, rows 2 and 1 of the third column.


( ) [ ][ ] [ ⁄ ]
⁄ ⁄

In the last step, we want to add one of row 2 to row 1.

⁄ ⁄
( ) [ ][ ⁄ ] [ ⁄ ]
⁄ ⁄

3
This is the original matrix in reduced row echelon form. We reached this matrix by multiplying it
by a series of elimination matrices. We seek, however, the single matrix that applies the same
transformations all in one step, so this matrix, , is the product of the elimination matrices.


( ( ( ( )))) ( ) [ ⁄ ]

The above equation demonstrates the application of the associative property of matrix arithmetic.
That is, we can regroup the multiplications in the equation as long as we don’t change the order of
the matrices being multiplied, since matrix arithmetic is not commutative. The result of this
multiplication, then, is

[ ][ ][ ][ ][ ]

[ ]

Note: to make the matrix more readable, the scalar was factored out to remove fractions.

As a last note, we can confirm that is invertible, because each of the elimination matrices is
invertible. This is true because since each simply finds combinations of the rows that have some
effect on the matrix, we can find another matrix that simply undoes that effect. Take for
example.

[ ]

Its effect on any matrix with three rows is adding -2 times the second row to 1 time the third row.
To undo that effect, we simply add back 2 times the second row to the first row.

[ ]

We multiply by , we get the identity matrix , confirming that is invertible.

[ ][ ] [ ] [ ][ ]

You might also like