Scribd
Upload
35 views

3.4 Gramm - Schmidt Orthonormalization: U X U X

Uploaded by

Laxman Naidu N
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
35 views

3.4 Gramm - Schmidt Orthonormalization: U X U X

Uploaded by

Laxman Naidu N
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 10

104

3.4 GRAMM SCHMIDT ORTHONORMALIZATION



Let U
1
, U
2
, ., U
k
be k linearly independent vectors in C
n spanning a subspace W
.
The Gramm Schmidt process is the method to get an orthonormal set
k
,....., ,
2 1
such that the subspace spanned by U
1
, .., U
k
is the same as
the subspace spanned by
k
,.....,
1
thus providing an orthonormal basis for W.

The process goes as follows:

Let ;
1 1
U =


( )
1 1
1
1
1 1
,


= =
Note
1
1 =
(We have used the symbol x to denote the norm ) x , x ( of a vector x)

Next, let,

( )
1 1 2 2 2
, U U =

Note that

( )
1 2
,
= ) , ) , U (( ) , U (
1 1 1 2 1 2

= ) , )( , U ( ) , U (
1 1 1 2 1 2

= 1 ) , ( (since ) , U ( ) , U (
1 1 1 2 1 2
=
Hence we get,

.
1 2


Let

;
2
2
2

=
clearly
( ) 0 , , 1 , 1
2 1 1 2
= = =


Also

x =
1
U
1
+
2
U
2
then
( ) ( )
1 1 2 2 2 1 1
, U x + + =
( ) [ ]
1 1 2 2 2 2 1 1 1
, U x + + =



105
2 2 1 1
+ = x , where


( )
1 2 2 1 1 1
, U + =



2 2 2
=


Thus x subspace spanned by U
1
, U
2


x subspace spanned by
1
,
2
.

Thus
1
,
2
is an orthonormal basis for the subspace [U
1
,U
2
].

Having defined
1
,
2
,..,
i-1
we define
i
as follows:


( ) =

=
1 i
1 p
p p p i i
, U U
Clearly
( ) 0 ,
p i
=
1 p i-1

and

i
i
i

=


Obviously 1
i
= and
( )
, 0
i j
= for 1 1 j i

and x the subspace spanned by U
1
,U
2
,,U
i
which we denote by
[U
1
, U
2
, .., U
i
]
xthe subspace spanned by
i 2 1
, , , L which we denote by [
1
, ..,
i
].
Thus
1
,
2
, ..,
i
is an orthonormal basis for [U
1
, .., U
i
].
Thus at the k
th
stage we get an orthonormal basis
1
, .,
k
for [U
1
, .., U
k
].

Example:

Let
|
|
|
|
|
.
|

\
|
=
|
|
|
|
|
.
|

\
|

=
|
|
|
|
|
.
|

\
|
=
0
1
3
2
;
0
1
1
1
;
0
1
1
1
3 2 1
U U U


be l.i. vectors in R
4
. Let us find an orthonormal basis for the subspace
spanned by U1, U2, U
3
using the Gramm Schmidt process.


106
;
0
1
1
1
U
1 1
|
|
|
|
|
.
|

\
|
= =

( )
|
|
|
|
|
.
|

\
|
=

=
0
1
1
1
3
1
,
1 1
1
1


|
|
|
|
|
|
|
|
.
|

\
|
=

0
3
1
3
1
3
1
1


( )
1 1 2 2 2
, U U =


|
|
|
|
|
|
|
|
.
|

\
|
|
|
.
|

\
|
+
|
|
|
|
|
.
|

\
|

=
0
3
1
3
1
3
1
3
1
3
1
3
1
0
1
1
1



|
|
|
|
|
|
.
|

\
|

|
|
|
|
|
.
|

\
|

=
0
3
1
3
1
3
1
0
1
1
1



|
|
|
|
|
.
|

\
|

=
0
3
4
3
2
3
2
and
3
6 2
9
16
9
4
9
4
2
= + + =


107
|
|
|
|
|
|
.
|

\
|

=
|
|
|
|
|
.
|

\
|

= =
0
6
2
6
1
6
1
0
3
4
3
2
3
2
6 2
3
2
2
2



Thus


|
|
|
|
|
|
.
|

\
|

=
0
6
2
6
1
6
1
2



Finally,


( ) ( )
2 2 3 1 1 3 3 3
, , U U U =



|
|
|
|
|
|
.
|

\
|

|
|
.
|

\
|

|
|
|
|
|
|
.
|

\
|
|
|
.
|

\
|

|
|
|
|
|
.
|

\
|
=
0
6
2
6
1
6
1
6
3
0
3
1
3
1
3
1
3
6
0
1
3
2



|
|
|
|
|
.
|

\
|

|
|
|
|
|
.
|

\
|

|
|
|
|
|
.
|

\
|
=
0
1
2
1
2
1
0
2
2
2
0
1
3
2



|
|
|
|
|
.
|

\
|

=
0
0
2
1
2
1




108
2
1
2
1
4
1
4
1
3
= = + =


|
|
|
|
|
|
.
|

\
|

=
|
|
|
|
|
.
|

\
|

= =
0
0
2
1
2
1
0
0
2
1
2
1
2
3
3
3




Thus the required orthonormal basis for W, the subspace spanned by U
1
,U
2
, U
3

is
1
,
2
,
3
, where

|
|
|
|
|
|
.
|

\
|

=
|
|
|
|
|
|
.
|

\
|

=
|
|
|
|
|
|
.
|

\
|
=
0
0
2
1
2
1
;
0
6
2
6
1
6
1
;
0
3
1
3
1
3
1
3 2 1



Note that these
i
are mutually orthogonal and have, each, length one.

We now get back to Hermitian matrices. We had seen that the
eigenvalues of a Hermitian matrix are all real; and that the eigenvectors
corresponding to district eigenvalues are mutually orthogonal. We can further
show the following: (We shall not give a proof here, but illustrate with an
example).

Let A be any nxn Hermitian matrix. Let
( ) ( ) ( ) ( )
k
a
k
a a
C = .....
2 1
2 1
be its characteristic polynomial, where
1
,
2
,
..,
k
are its distinct eigenvalues and a
1
, .., a
k
are their algebraic multiplicities.
If W
i
is the characteristic subspace, (eigensubspace), corresponding to the
eigenvalue
i
; that is,


{ } x Ax x
i i
= = :

then it can be shown that dim is
i
= a
i
.

We choose any basis for
i
and orthonormalize

it by

G-S process and get an
orthonormal basis for
i
. If we now take all these orthonormal - basis vectors for

1
, . .,
k
and write them as the columns of a matrix P then

109
P
*
AP
will be a diagonal matrix.

Example :

|
|
|
.
|

\
|

=
3 1 2
1 3 2
2 2 6
A

Notice

A*

= A
1
= A
1
= A.

Thus the matrix A is Hermitian.

Characteristic Polynomial of A:

3 1 2
1 3 2
2 2 6

A I



( )
3 1 2
1 3 2
0 2 2 2
2 1
2


R R



( )
3 1 2
1 3 2
0 2 1
2

=




( )
3 5 0
1 7 0
0 2 1
2
1 2
1 3
2
2


R R
R R



( ) ( )( ) [ ] 5 3 7 2 =


110
( )[ ] 16 10 2
2
+ =

( )( )( ) 8 2 2 =

( ) ( ) 8 2
2
=


Thus
( ) ( ) ( ) 8 2
2
= C


1
= 2 a
1
= 2


2
= 8 a
2
= 1

The characteristic subspaces:

{ } x 2 Ax : x W
1
= =


( ) { } = = x I A x 2 :

i.e. We have to solve

(A 2I) x =

i.e.
|
|
|
.
|

\
|
=
|
|
|
.
|

\
|
|
|
|
.
|

\
|

0
0
0
1 1 2
1 1 2
2 2 4
3
2
1
x
x
x


2x
1
x
2
+ x
3
= 0

x
3
= - 2x
1
+ x
2


1
2 1 2
1 2
; ,
2
x
x x x x
x x
| |
|
=
|
|
+
\ .
arbitrary

1
: ; ,
2
x x scalars



| |
|
= =
`
|
|
+
\ . )


111

A basis for W
1
is

|
|
|
.
|

\
|
=
|
|
|
.
|

\
|

=
1
1
0
;
2
0
1
2 1
U U


We now orthonormalize this:

|
|
|
.
|

\
|

= =
2
0
1
1 1
U
5
1
=

1
1
1

=


|
|
|
|
|
.
|

\
|

=
5
2
0
5
1
1


( )
1 1 2 2 2
, U U =


|
|
|
|
|
.
|

\
|

|
|
.
|

\
|

|
|
|
.
|

\
|
=
5
2
0
5
1
5
2
1
1
0



|
|
|
|
|
.
|

\
|

+
|
|
|
.
|

\
|
=
5
4
0
5
2
1
1
0



|
|
|
|
.
|

\
|
=
5
1
1
5
2



112
5
30
25
30
25
1
1
25
4
2
= = + + =

|
|
|
|
|
.
|

\
|
=
|
|
|
|
.
|

\
|
= =
30
1
30
5
30
2
5
1
1
5
2
30
5
2
2
2




1
,
2
is an orthonormal basis for W
1
.

{ } x 8 Ax : x W
2
= =

( ) { } = = x I A x 8 :

So we have to solve

(A-8I) x = i.e.


|
|
|
.
|

\
|
=
|
|
|
.
|

\
|
|
|
|
.
|

\
|



0
0
0
5 1 2
1 5 2
2 2 2
3
2
1
x
x
x


This yields x
1
= -x
2
+ x
3
and therefore the general solution is


|
|
|
.
|

\
|
=
|
|
|
.
|

\
|

1
1
2 2


|
|
|
.
|

\
|
=
1
1
2
:
3
U Basis


Orthonormalize: only one step:

|
|
|
.
|

\
|
= =
1
1
2
3 3
U


113


|
|
|
|
|
.
|

\
|
=
|
|
|
.
|

\
|
= =
6
1
6
1
6
2
1
1
2
6
1
3
3
3



If
|
|
|
|
|
|
.
|

\
|

=
6
1
30
1
5
2
6
1
30
5
0
6
2
30
2
5
1
P

Then

P* = P
1
and

;
8 0 0
0 2 0
0 0 2
1 *
|
|
|
.
|

\
|
= = AP P AP P

a diagonal matrix.

You might also like