Elementary Matrices and RREF: N N I Ij Ij
Elementary Matrices and RREF: N N I Ij Ij
1 Elementary Matrices
During solving a linear system, we have observed that we need to make row operations. We define
them now.
0 0 1 −1 1 0
3. Then E13 =
0 1 0 and E13 A = 0 −1 0.
1 0 0 1 2 5
When we solve a linear system, we multiply them on the left as different equations correspond to
different rows of the augmented matrix.
If we multiply these elementary matrices on the right then they correspond to column transforma-
tions. Note the change in the idea of Eij (c)!!
1 10 5 1 −3 5 5 2 1
AE2 (5) = 0 −5 0, AE12 (−5) = 0 −1
and AE13
0 =
0 −1 0 .
−1 5 0 −1 6 0 0 1 −1
Further, we observe
that
1 0 0 1 5 0 0 0 1
−1 −1 −1
(E2 (5)) = 0 1/5 0, (E12 (−5)) = 0 1
and (E13 ) = E13
0 =
0 1 0.
0 0 1 0 0 1 1 0 0
Is every elementary matrix (of arbitrary size n × n) invertible ? Think about it.
2. if the pivot of the (i + 1)-th row, if it exists, comes to the right of the pivot of the i-th row.
Important: Thus, observe that if A is in RREF and has r pivots then the matrix Ir is seating
inside the RREF as a submatrix.
So, look at the above examples and note that the matrices which are in RREF, have the above
property. For the matrices which are NOT in RREF, this property doesn’t hold.