Lecture -

9 ⑭4
-

properties of matrices
-
·
We saw how to define product of
and orders
two matrices A B of
mxn
and nxk ·

· We define addition of two matrices

A and B
of same order max

as follows :

and
If A =

[4] B =

[bij]
then A + B =

[C]
= dij+by for every
where 12ihm
& 12 jan .

ample
Ex- :
A =

[1 -]
,
,
B =

[0 %]
15 4

ther
[ ! 1is 10]
2+ 1 16
B 3 +

+ =

< +
4 04

-[in 19 I
3

Il 10
 ⑭
verifiable
facts :

A(BC)
Activity :
(AB)C =

write ABC
just
.

So we can

A(B+ C) Ac
Distributivity : AB
= +

(B c) D BD CD
- +
+ =

Non-commutativity
->
There exists
square
nxn matrices A and B such that
AB F BA .

Example : A =

[b] ,
B =

['43]
Then AB=
[ii] ,
BA=
[in 20] .
 ⑭
e
Upper triangular n
matrix :

An exh matrix A =

[4ij] is said to
the entries
be
upper triangular if all
that
diagonal
is
below the are zero
, ,

anj
=
0
for all ,j <in .

I
a
an
a as
I
-

, n -1

ass---. in an
.

A =

00

. ⑧
0 ----

An n -
1 En ,
n

L 0 ⑧ 0--- 0 On
We have seen that when we
perform
the Gaussian elimination the final
, 2

system of equations (in n

equations
with n
unknowns turns out to

be in
upper triangular form .

defines lower
Similarly,
to be the
one

one where
triangular
all the entries
matrix

above the di
agonal
all
ave

1_i< ;
zero
,
that

is aij=0 for <m .

 0
47
Example
(2] - ]
:

, .

Triangular
-
factorization :

Recall the system of linear equations

-
24+ v+ w - -
(i)
44 -
6 ii)
zw=
-

-
2u +

v +

- (iii)

which was written in matrix form as !

N
- x -

(=
21

][] [ ] = =

·
Recall the three elimination steps -

substract twice
Step 1 :
of equation cis
equation (ii) &
put into is
-
from .

5
.

24 + v +
w =

-
80 -2w =
12=

2u+7v + 2w= 9
elimination matrix
& the
corresponding
[ % ].
is
E=
 ⑲
S After Step 1 add (i) & (iii
;
2:
, equations
equation
and put it in
place of (iii) .

zu + v +
w = 5
-
Dv-zw=-12
80 +
3w =
14
elimination matrix
& the
corresponding
[ %]
is I 10
I -
=
·

I 0 I .

S After step
3 2 add (i) & (iii
;
S

equations
:
9
equation (iii)
and put it in
place of .

zu + v +
w = 5
-
Dv-zw=-12
w 2 =

elimination matrix
& the
corresponding
[ 10]
is 10 0
GE ·

0 I I
.

·
Note that A acts on x then E acts
,
An then Facts on EAx and
on ,
,

finally G acts on
FEAx .
 ⑭
let look
so , us at the matri
an
GFE
GFEA .

iii I
(!; ] : ) [00
aF
=

&
GFE -

[!; ][ ] [-2 =

-
:
11 I
I
lower triangular
&

[ zi](t] [8 oI
=

v =
GF EA =

upper
Also
, triangular
2

GFEAx= GFEb=
Ux =

G just thame
.

If we now
invert/reverse the steps 3
I
2
and 1 then we would get back to
E
the original system .
 would ⑤
These correspond to matrices

-
I
! ] /! E 6I 8
0
=

,
I

Now,

L= EFG =

I, i]
and
A =
EFG GFEA
=L