0% found this document useful (0 votes)
474 views5 pages

Homework 8 Solutions: 6.2 - Gram-Schmidt Orthogonalization Process

The document contains solutions to homework problems involving orthonormal bases and Gram-Schmidt orthogonalization. It provides the steps to apply the Gram-Schmidt process to generate orthonormal bases for given sets of vectors in inner product spaces. It also computes Fourier coefficients of vectors with respect to the generated orthonormal bases.

Uploaded by

CJ Jacobs
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
474 views5 pages

Homework 8 Solutions: 6.2 - Gram-Schmidt Orthogonalization Process

The document contains solutions to homework problems involving orthonormal bases and Gram-Schmidt orthogonalization. It provides the steps to apply the Gram-Schmidt process to generate orthonormal bases for given sets of vectors in inner product spaces. It also computes Fourier coefficients of vectors with respect to the generated orthonormal bases.

Uploaded by

CJ Jacobs
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

Homework 8 Solutions

Joshua Hernandez
November 18, 2009
6.2 - Gram-Schmidt Orthogonalization Process
2. Apply the Gram-Schmidt process to the given subset S of the inner product space V. Normalize the
vectors in the resulting basis to obtain an orthonormal basis , and compute the Fourier coecients
of the given vector relative to .
d. V = span(S), where S = {v
1
= (1, i, 0), v
2
= (1 i, 2, 4i)}, and x = (3 + i, 4i, -4).
Solution:
Orthogonalization: Dene v

1
= v
1
. We compute the inner products
v

2
= (1, i, 0)
2
= |1|
2
+|1|
2
+|0|
2
= 2
v
2
, v

1
= (1 i, 2, 4i), (1, i, 0) = (1 i) 1 + 2 i + 4i 0 = 1 3i,
and compute the next orthogonal vector
v

2
= v
2
proj
v

1
(v
2
) = v
2

v
2
, v

1
, v

1
= (1 i, 2, 4i)
13i
2
(1, i, 0) =
1
2
(1 + i, 1 i, 8i),
where
v

2
=
1
2
(1 + i, 1 i, 8i)
2
=
1
4
(|1 + i|
2
+|1 i|
2
+|8|
2
) = 17.
Normalization:
w
1
=
v

1
v

2
2
(1, i, 0), w
2
=
v

2
v

17
34
(1 + i, 1 i, 8i).
So = {

2
2
(1, i, 0),

17
34
(1+i, 1i, 8i)} is an orthonormal basis of span(S). Now we can compute
the Fourier coecients for x = (3 + i, 4i, -4):
c
1
= x, w
1
=
_
(3 + i, 4i, -4),

2
2
(1, i, 0)
_
=

2
2
_
(3 + i) 1 + 4i i 4 0

2
2
(7 + i),
c
2
= x, w
2
=
_
(3 + i, 4i, -4),

17
34
(1 + i, 1 i, 8i)
_
=

17
34
_
(3 + i) (1 + i) + 4i (1 i) 4 8i

17
34
[(3 + i)(1 i) + 4i(1 + i) 4(-8i)] =

17
34
((3 + 1 4) + (-3 + 1 + 4 + 32)i) =

17i.
Checking that these coecients work:
x = c
1
w
1
+ c
2
w
2
=

2
2
(7 + i)

2
2
(1, i, 0) +

17i

17
34
(1 + i, 1 i, 8i)
=
1
2
(7 + i, -1 + 7i, 0) +
1
2
(-1 + i, 1 + i, -8) = (3 + i, 4i, -4).
1
f. V = R
4
, S = {v
1
= (1, -2, -1, 3), v
2
= (3, 6, 3, -1), v
3
= (1, 4, 2, 8)}, and x = (-1, 2, 1, 1).
Solution:
Orthogonalization: Dene v

1
= v
1
. We compute the inner products
v

1
, v

1
= (1, -2, -1, 3), (1, -2, -1, 3) = 1
2
+ (-2)
2
+ (-1)
2
+ 3
2
= 15
v
2
, v

1
= (3, 6, 3, -1), (1, -2, -1, 3) = 3(1) + 6(-2) + 3(-1) 1(3) = -15,
and compute the next orthogonal vector
v

2
= (3, 6, 3, -1)
-15
15
(1, -2, -1, 3) = (4, 4, 2, 2).
Computing more inner products:
v

2
, v

2
= (4, 4, 2, 2), (4, 4, 2, 2) = 4
2
+ 4
2
+ 2
2
+ 2
2
= 40,
v
3
, v

1
= (1, 4, 2, 8), (1, -2, -1, 3) = 1(1) + 4(-2) + 2(-1) + 8(3) = 15,
v
3
, v

2
= (1, 4, 2, 8), (4, 4, 2, 2) = 1(4) + 4(4) + 2(2) + 8(2) = 40,
So
v

3
= v
3
proj
v

1
(v
3
) proj
v

2
(v
3
) = (1, 4, 2, 8)
15
15
(1, -2, -1, 3)
40
40
(4, 4, 2, 2) = (-4, 2, 1, 3),
and
v

2
, v

2
= (-4, 2, 1, 3), (-4, 2, 1, 3) = (-4)
2
+ 2
2
+ 1
2
+ 3
2
= 30.
Normalization:
w
1
=
v

1
v

15
15
(1, -2, -1, 3), w
2
=
v

1
v

10
20
(4, 4, 2, 2), w
2
=
v

1
v

30
30
(-4, 2, 1, 3).
So = {

15
15
(1, -2, -1, 3),

10
20
(4, 4, 2, 2),

30
30
(-4, 2, 1, 3)} is an orthonormal basis of span(S). Now
we can compute the Fourier coecients for x = (-1, 2, 1, 1):
c
1
= x, w
1
=

15
15
(-1, 2, 1, 1), (1, -2, -1, 3) = -3

15
15
=
-

15
5
,
c
2
= x, w
2
=

10
20
(-1, 2, 1, 1), (4, 4, 2, 2) = 8

10
20
=
2

10
5
,
c
3
= x, w
3
=

30
30
(-1, 2, 1, 1), (-4, 2, 1, 3) = 12

30
30
=
2

30
5
.
Checking that these coecients work:
c
1
w
1
+ c
2
w
2
+ c
3
w
3
=
-

15
5

15
15
(1, -2, -1, 3) +
2

10
5

10
20
(4, 4, 2, 2) +
2

30
5

30
30
(-4, 2, 1, 3)
=
-1
5
(1, -2, -1, 3) +
1
5
(4, 4, 2, 2) +
2
5
(-4, 2, 1, 3) = (-1, 2, 1, 1).
6. Let V be an inner product space, and let W be a nite-dimensional subspace of V. If x / W, prove that
there exists y V such that y W

, but x, y = 0.
Solution: By Theorem 6.6, we may write x = u + y, with u W and y W

.
x, y = u + y, y = u, y +y, y = y, y .
Note that y = 0, otherwise x = u W. Coercivity of the inner product forces x, y = y, y > 0.
11. Let A be an n n matrix with complex entries. Prove that AA

= I if and only if the rows of A form


2
an orthonormal basis for C
n
.
Solution: Denote by r
i
the ith row of A. Observe that
(AA

)
ij
=
n

k=1
A
ik
(A

)
kj
=
n

k=1
A
ik
A
jk
= r
i
, r
j
.
so
AA

= I (AA

)
ij
= r
i
, r
j
=
ij
{r
i
}
n
i=1
forms an orthonormal basis.
13c. Let V be an inner product space, and W be a nite-dimensional subspace of V. Prove that W = (W

.
Solution:
If x W and y W

, then x, y = y, x = 0, so x (W

.
If x / W, then, by problem 13, there exists y W

such that x, y = 0, so x / (W

.
16. a. Let V be an inner-product space, and let S = {v
1
, . . . , v
n
} be an orthonormal subset of V. Prove
that for any x V we have
x
2

i=1
| x, v
i
|
2
.
Solution: Dene W = span(S). This is a nite-dimensional subspace of V. Given x V, we
can write x = u + y, where u W and y W

(Theorem 6.6). Since S is an orthonormal basis


of W, we can write u =

n
i=1
u, v
i
v
i
, so that
x
2
= u
2
+y
2
=
_
_
n

i=1
u, v
i
v
i
_
_
2
+y
2
.
Applying problem 6.1.12,
=
n

i=1
| u, v
i
|
2
v
i

2
+y
2
=
n

i=1
| u, v
i
|
2
+y
2

i=1
| u, v
i
|
2
b. In the context of (a), prove that Bessels inequality is an equality if and only if x span(S).
Solution: In the context of (a), we have equality if and only if y
2
= 0, which occurs if and
only if y = 0, meaning x = u W = span(S).
18. Let V = C([1, 1]). Suppose that W
e
and W
o
denote the subspaces of V consisting of even and odd
functions, respectively. Prove that W
e
W
o
, where the inner product on V is dened by
f, g =
_
1
-1
f(t)g(t) dt.
3
Solution: Given f W
e
and g W
o
,
f(t), g(t) =
_
1
-1
f(t)g(t) dt =
_
0
-1
f(t)g(t) dt +
_
1
0
f(t)g(t) dt
= -
_
0
1
f(-t)g(-t) dt +
_
1
0
f(t)g(t) dt =
_
1
0
_
f(-t)g(-t) + f(t)g(t)
_
dt.
Applying the denition of odd and even functions,
=
_
1
0
_
-f(t)g(t) + f(t)g(t)
_
dt =
_
1
0
0 dt = 0.
19c. Let V = P(R) with the inner product f(x), g(x) =
_
1
0
f(t)g(t) dt and let W = P
1
(R) V. Find the
orthogonal projection of h(x) = 4 + 3x 2x
2
onto W.
Solution: Let = {v
1
= 1, v
2
= x}, a basis of W. Let v

1
= v
1
. Then
v

2
= 1, 1 =
_
1
0
1
2
dx = 1
v
2
, v

1
= x, 1 =
_
1
0
x(1) dx =
_
1
2
x
2

1
0
=
1
2
.
The next orthogonal basis vector:
v

2
= v
2
proj
v

1
(v
2
) = x
1/2
1
1 = -
1
2
+ x.
where
v
2

2
=

-
1
2
+ x, -
1
2
+ x
_
=
_
1
0
(-
1
2
+ t)
2
dt =
1
3
_
(-
1
2
+ x)
3

1
0
=
1
3
(
1
8

1
8
) =
1
12
.
Now we compute inner products with h(x) = 4 + 3x 2x
2
,
h(x), v

1
=

4 + 3x 2x
2
, 1
_
=
_
1
0
(4 + 3x 2x
2
) dx =
_
4x +
3
2
x
2

2
3
x
3

1
0
=
29
6
h(x), v

2
=

4 + 3x 2x
2
, -
1
2
+ x
_
=
_
1
0
(4 + 3x 2x
2
)(-
1
2
+ x) dx
=
_
1
0
(-2 +
5
2
x + 4x
2
2x
3
) dx =
_
-2x +
5
4
x
2
+
4
3
x
3

1
2
x
4

1
0
=
1
12
.
So
proj
W
(h) = proj
v

1
(h) + proj
v

2
(h) =
29/6
1
1 +
1/12
1/12
(-
1
2
+ x) = x +
13
3
.
0.2 0.4 0.6 0.8 1.0
0.4
0.2
0.2
0.4
0.6
0.8
1.0
(a) Orthogonal Basis
0.2 0.4 0.6 0.8 1.0
4.2
4.4
4.6
4.8
5.0
5.2
(b) Orthogonal Approximation
20c. In the context of (19c), nd the distance from h(x) to W.
4
Solution: The distance from h(x) to W is the distance between h(x) and proj
W
(h).
h(x) proj
W
(h) =
__
1
0
_
(4 + 3x 2x
2
) (
13
3
+ x)
_
2
dx
_1/2
=
__
1
0
_
-
1
3
+ 2x 2x
2
_
2
dx
_1/2
=
__
1
0
_
1
9

4
3
x +
16
3
x
2
8x
3
+ 4x
4
_
dx
_1/2
=
_
_
1
9
x
2
3
x
2
+
16
9
x
3
2x
4
+
4
5
x
5

1
0
_
1/2
=
_
(5 30 + 80 90 + 36)/45 =

5/15.
22. Let V = C([0, 1]) with the inner product f, g =
_
1
0
f(t)g(t) dt. Let W be the subspace spanned by the
linearly independent set {t,

t}.
a. Find an orthonormal basis for W
Solution: Let = {v
1
= t, v
2
=

t}, a basis of W. Let v

1
= v
1
. Then
v

2
= t, t =
_
1
0
t
2
dx =
_
1
3
t
3

1
0
=
1
3
v
2
, v

1
=

t, t
_
=
_
1
0

tt dt =
_
2
5
x
5
2
_
1
0
=
2
5
.
The next orthogonal basis vector:
v

2
= v
2
proj
v

1
(v
2
) =

t
2/5
1/3
t =

t
6
5
t.
where
v
2

2
=
_
1
0
(

t
6
5
t)
2
dt =
_
1
0
(t
12
5
t
3
2
+
36
25
t
2
) dt
=
_
(
1
2
t
2

24
25
t
5
2
+
12
25
t
3
_
1
0
=
1
50
(25 48 + 24) =
1
50
.
b. Let h(t) = t
2
. Use the orthonormal basis obtained in (a) to obtain the closest approximation of h
in W.
Solution: Now we compute inner products with h(t) = t
2
,
h(t), v

1
=

t
2
, t
_
=
_
1
0
t
3
dx =
_
1
4
t
4

1
0
=
1
4
h(t), v

2
=

t
2
,

t
6
5
t
_
=
_
1
0
(t
5
2

6
5
t
3
) dx =
_
2
7
t
7
2

3
10
t
4
_
1
0
=
-1
70
.
So
proj
W
(h) = proj
v

1
(h) + proj
v

2
(h) =
1/4
1/3
t +
-1/70
1/50
(

t
6
5
t) =
-5
7

t +
45
28
t.
0.2 0.4 0.6 0.8 1.0
0.2
0.2
0.4
0.6
0.8
1.0
(c) Orthogonal Basis
0.2 0.4 0.6 0.8 1.0
0.2
0.4
0.6
0.8
1.0
(d) Orthogonal Approximation
5

You might also like