Sei sulla pagina 1di 10

Physics 411 HW#1 Solutions

Problem 1 Read the Gram-Schmidt procedure in problem A.4 of Griffiths (p. 440). Consider the following sets of complex vectors.
..
.
Which sets are linearly independent, span the space of complex 3-dimensional vectors (C3 ),
neither, or both? Explain why. Are any of these sets of vectors a basis for C3 ? If so, check
if the vectors are orthonormal. If you find they are not orthonormal, then use the GramSchmidt procedure to generate an orthonormal basis.
Solution
Let us consider general strategies for determining the linear dependence and span before
looking at any particular set. For a small set of vectors, denote it {|v1 i , |v2 i , . . . , |vn i},
linear dependence is most easily shown by explicitly constructing one of the vectors in terms
of the others. We could write this as
X
aj |vj i
|vi i =
j6=i

where aj is some complex number. The sum may not even have to go over all of the vectors
in the set (an example of this comes up in set C). The condition for linear independence is
X
aj |vj i = 0 aj = 0.
j

All the aj are zero! This says that there is no vector which is equivalent to the sum of others
in the set. Essentially you show that the only solution is the trivial solution.
How about span? We need to write out an arbitrary vector in whatever vector space we are
in and show that it can be written as a sum over multiples of the vectors in our set. As you
will see below, this is less work than it sounds. On to the actual problem!

0
0
0
A = 0 , i , 1

i
1
0
Linear Independence: This set of vectors is not linearly independent. To show its linearly
dependent we can write the second vector in terms of the first and third:



0
0
0
i = i 0 + i 1
1
i
0
1

Span: This set of vectors does not span the space C3 . This is immediately obvious since
every vector of the set has a zero in the first entry. We could never generate the following
arbitrary vector from this set when 6= 0.



0
0
0
6= a 0 + b i + c 1

i
1
0

0
2i
1
B = 1 , 5 , 0

0
i
3+i
Linear Independence: This set of vectors is linearly independent. Remembering the
condition stated earlier, we have to show that the only solution to the following is the trivial
solution (a = b = c = 0):


1
0
2i
0
a 1 + b 5 + c 0 = 0
0
i
3+i
0
This breaks up into three separate equations:
a + 2ic = 0 a = 2ic
a + 5b = 0 a = 5b
ib + (3 + i)c = 0 b = (3i 1)c

(1)
(2)
(3)

Substituting (1) and (3) into (2) we have 5(3i 1)c = 2ic, and the equality only holds
when c = 0! It then follows immediately that a = b = c = 0, and thus the set of vectors are
linearly independent.
Span: This set of vectors does span the space C3 . Given an arbitrary vector we need only
show that a, b, and c can be solved for in terms of , , and .


1
0
2i
= a 1 + b 5 + c 0

0
i
3+i
Rewriting this as a system of equations:
a + 2ic =
a + 5b =
ib + (3 + i)c =
2

(4)
(5)
(6)

Subtracting (4) from (5) we get = 5b 2ic. Now solve for b in (6), b = i + (3i 1)c.
Substituting this result into the first we can solve for c.
c=

+ 5i
13i 5

We are done! Now c is known in terms of , and , and since a and b are written in terms
of c they are also known in terms of , and .



0
0
0
i
C = 0 , 2i + 1 , i , 0

0
0
4i
3
Linear Independence: This set of vectors is not linearly independent, and we dont need
to use all of the vectors in the set to show this. To show its linearly dependent we can write
the third vector in terms of the second and fourth:



0
0
0
i = i 2i + 1 + 4i 0
2i + 1
3
0
3
4i
Span: This set of vectors does span the space C3 . Given an arbitrary vector we need only
show that a, b, and c can be solved for in terms of , , and .

i
0
0
= a 0 + b 2i + 1 + c 0

0
0
3
You can read off what a, b, and c must be in terms of , , and .
a = i

b=
2i + 1

c=
3
Would it ever be possible to have a set of 4 vectors in C3 which are both linearly independent
and spanning?

0
1+i

0
D=
, 1 + 2i

0
0
3

Linear Independence: This set of vectors is linearly independent. Remembering the


condition stated earlier, we have to show that the only solution to the following is the trivial
solution (a = b = c = 0):


1+i
0
0

0
a
+ b 1 + 2i = 0
0
0
0
It is clear that both a and b must be zero to solve this system.
Span: This set of vectors does not span the space C3 . Since both vectors have a zero in the
third entry we could never generate an arbitrary vector from the set when 6= 0.

1+i
0
6= a 0 + b 1 + 2i

0
0

In summary, we obtained the following results:


Set Lin. Indep.
A
no
B
yes
C
no
D
yes

Spanning
no
yes
yes
no

Since the vectors of set B are both linearly independent and spanning they form a basis.
We can calculate the inner product of two of the vectors, and if it is non-zero, then we know
they are not orthogonal to each other. Furthermore, we can find the norm of one of the
vectors to show that it is not normalized. Lets first take the inner product of the first and
second vector of B.

 0
1 1 0 5 = 5 6= 0
i
Now let us show that the second vector of B is not normalized.

 0
0 5 i 5 = 25 + 1 = 26 6= 1
i

The norm is the square root of the inner product, but it is clear this vector is not normalized.
Now we want to produce an orthonormal set of vectors from the vectors in B. We do this
by applying the Gram-Schmidt procedure.
To parallelize ourselves with Griffiths notation make the following definitions:

1

|e1 i = 1
0

|e2 i = 5
i

2i
|e3 i = 0
3+i

Also, we will define our orthonormal vectors in the same manner as Griffiths (i.e. |e01 i, |e02 i,
and |e03 i).
Lets begin.

1
1
|e1 i
= 1
|e01 i =
ke1 k
2 0
On to step two:

0
0
0

|e2 i he1 |e2 i |e1 i = 5


i

0

= 5
i
5
2
5

=
2
i



0
1


1
1

1 1 0 5
1
2
2
i
0

1
1

{5} 1
2
0

If we divide this vector by its norm, then we will have |e02 i. The norm, which you should
verify, turns out to be 254 .

5
1
|e02 i = 5
54 2i
Before going in to the calculation of |e03 i lets first compute the projection of |e3 i on to |e01 i
and |e02 i

2i


1
he01 |e3 i = 1 1 0 0 = 2i
2
3+i

2i

(16i + 2)
1

5 5 2i 0 =
he02 |e3 i =
54
54
3+i
Step 3:



2i
1
5
(16i + 2)
5
|e3 i he01 |e3 i |e01 i he02 |e3 i |e02 i = 0 i 1
54
3+i
0
2i

26i + 10
1
26i 10
=
54
50i + 130
If we divide this vector by its norm, then we will have |e03 i. The norm, which you should
verify, turns out to be 582
.
9

26i + 10
1
26i 10
|e03 i =
6 582 50i + 130
It can be shown that these are in fact orthonormal. If you didnt do so in your homework,
then I suggest you verify it now.

Problem 2 Consider the set of functions {|f0 i , |f1 i , |f2 i} defined on the the interval [L/2, +L]
such that hx|f0 i = 1, hx|f1 i = ix, and hx|f2 i = x2 + 3i. Given the inner product hg|f i defined in class, use the Gram-Schmidt procedure on these functions, beginning with |f0 i and
progressing to |f2 i, to create an orthonormal basis {|0 i , |1 i , |2 i} for the set of polynomial
functions of order 2 with complex coefficients on the specified interval. Express your final
answer as functions of the real coordinate x, namely hx|i i, and also provide |i i as linear
combinations of |fi i.
Solution
Begin by constructing |0 i from |f0 i:
6

|f0 i
|0 i = p
hf0 |f0 i
ZL
hf0 |f0 i =

ZL
hf0 |xihx|f0 i dx =

L/2

1 dx =

3L
2

L/2

2
|f0 i
3L
r
2
hx|0 i =
3L

|0 i =

Now we construct |1 i:
|1 i =

|f1 i h0 |f1 i |0 i
k...k

where the denominator is the norm of the numerator. We need to calculate the coefficient
appearing in front of |0 i and then the norm of the numerator.

ZL
h0 |f1 i =

ZL r
h0 |xihx|f1 i dx =

L/2

2
3
(ix)dx = i
3L
4

L3
6

L/2

v
!
! r
r
r
u
3
3
u
3
L
3
L
9L3
k...k = t hf1 | + i
h0 |
|f1 i i
|0 i =
4 6
4 6
32
To complete the calculation above you will need to know hf1 |f1 i.
ZL
hf1 |f1 i =

ZL
hf1 |xihx|f1 i dx =

L/2

3
(ix)(ix)dx = L3
8

L/2

We can put everything together now and write:


r


32
L
|f1 i i |f0 i
|1 i =
9L3
4
r


32
L
hx|1 i = i
x
9L3
4
7

Now we construct |2 i:
|2 i =

|f2 i h1 |f2 i |1 i h0 |f2 i |0 i


k...k

Now calculate h1 |f2 i, h0 |f2 i, and the norm of the numerator.

ZL
h1 |f2 i =

ZL
h1 |xihx|f2 i dx =

L/2

r
i

32
9L3

r


9L5
L
x
(x2 + 3i)dx = i
4
128

L/2

ZL
h0 |f2 i =

ZL r
h0 |xihx|f2 i dx =

L/2

2 2
(x + 3i)dx =
3L

2
3L

3L3
9L
+i
8
2

L/2

v
!
r
r
u
 3

5
u
9L
2
3L
9L
k...k = t hf2 | i
h1 |
i
h0 |
128
3L
8
2
!
r
r


9L5
2 3L3
9L
|f2 i + i
|1 i
+i
|0 i
128
3L
8
2
r
27L5
k...k =
640
Putting everything together:
r

 2



L
L
|2 i =
+ 3i |f0 i
|f2 i + i |f1 i
2
8
r


L
L2
640
2
hx|2 i =
x x
27L5
2
8
640
27L5

You should check the orthonormality of these results to ensure we have constructed an
orthonormal basis.

Problem 3 Create 3 3 matrices (in the Cartesian basis) that represent rotation operators
on vectors for (a) rotations of the vector by angle about the +zaxis and (b) rotations of
angle about the y axis. Show that for arbitrary angles these rotation matrices do not
commute: [Rz ()] [Ry ()] 6= [Ry ()] [Rz ()]. Prove that [Rz ()]n = [Rz (n)].
Solution
8

To construct the rotation matrices we consider what happens to each of the unit vectors in
R3 under the rotations (T (
ei ) is the vector which ei is transformed into by the rotation). We
write the rotation matrix around the positive z-axis by the following matrix:

ex ) T (
ey ) T (
ez )
Rz+ () = T (

cos() sin() 0
Rz+ () = sin() cos() 0
0
0
1
Rotation around the negative y-axis ( is measured from the positive x axis towards the
positive z axis):

cos() 0 sin()
1
0
Ry () = 0
sin() 0 cos()
Now we evaluate the commutator to show that reversing the order of rotations about these
two axes is not equivalent.

[Rz+ (), Ry ()] = Rz+ ()Ry () Ry ()Rz+ ()

cos() cos() sin() cos() sin()


= sin() cos() cos() sin() sin()
sin()
0
cos()

cos() cos() cos() sin() sin()


sin()
cos()
0
cos() sin() sin() sin() cos()
You can see that for arbitrary angle the off-diagonal terms will not cancel.
It is easiest to show [Rz ()]n = [Rz (n)] using the exponential form of the sine and cosine.
Additionally, since the rotation matrix is block diagonal we can focus on the 2 2 block in

the upper left.




cos() sin()
Rz () =
sin() cos()
ei +ei
2
ei ei
2i

i ei

2i
ei +ei
2





1 i 1 i
1 i 1 i
+ e
= e
i 1
i 1
2
2
Lets now look at [Rz ()]2 . Since we broke up our matrix into the sum of two different
matrices we have to consider four different products of matrices.








1 2i 1 i
1 2i
1 2i 1 i
1 i
2 2i
e
= e
= e
i 1
i 1
2i 2
i 1
4
4
2







1 2i 1 i
1 2i 2 2i
1 2i 1 i
1 i
e
= e
= e
i 1
i 1
2i 2
i 1
4
4
2





1 i i 1 i
1 0 0
1 i
e e
=
i 1
i 1
4
4 0 0





1 i i 1 i
1 0 0
1 i
e e
=
i 1
i 1
4
4 0 0
So, the cross terms vanish and the only change

1 2i 1
Rz ()Rz () = e
i
2
e2i +e2i
2
e2i e2i
2i

=

=

in the other two products is 2.





1 2i 1 i
i
+ e
1
i 1
2
!
2i
2i
e e
2i
e2i +e2i
2

cos(2) sin(2)
sin(2) cos(2)

Moreover, to any power this result holds: the cross terms vanish and n (n + 1). For the
mathematically inclined this can be proven by induction, but the techniques used in that
proof are the same as above.

10

Potrebbero piacerti anche