Hilbert Space Problems

Problems and Solutions
in
Hilbert space theory,
wavelets
and
generalized functions
by
Willi-Hans Steeb
International School for Scientic Computing
at
University of Johannesburg, South Africa
Preface
The purpose of this book is to supply a collection of problems in Hilbert
space theory, wavelets and generalized functions.
Prescribed books for problems.
1) Hilbert Spaces, Wavelets, Generalized Functions and Modern Quantum
Mechanics
by Willi-Hans Steeb
Kluwer Academic Publishers, 1998
ISBN 0-7923-5231-9
2) Classical and Quantum Computing with C++ and Java Simulations
by Yorick Hardy and Willi-Hans Steeb
Birkhauser Verlag, Boston, 2002
ISBN 376-436-610-0
3) Problems and Solutions in Quantum Computing and Quantum Informa-
tion, second edition
by Willi-Hans Steeb and Yorick Hardy
World Scientic, Singapore, 2006
ISBN 981-256-916-2
http://www.worldscibooks.com/physics/6077.html
The International School for Scientic Computing (ISSC) provides certi-
cate courses for this subject. Please contact the author if you want to do
this course or other courses of the ISSC.
e-mail addresses of the author:
steebwilli@gmail.com
steeb_wh@yahoo.com
Home page of the author:
http://issc.uj.ac.za
v
Contents
Notation ix
1 General 1
2 Finite Dimensional Hilbert Spaces 8
3 Hilbert Space L
2
() 19
4 Hilbert Space
2
(N) 41
5 Fourier Transform 44
6 Wavelets 48
7 Operators 52
8 Generalized Functions 56
Bibliography 68
Index 70
vii
Notation
:= is dened as
belongs to (a set)
/ does not belong to (a set)
intersection of sets
union of sets
empty set
N set of natural numbers
Z set of integers
Q set of rational numbers
R set of real numbers
R
+
set of nonnegative real numbers
C set of complex numbers
R
n
n-dimensional Euclidean space
space of column vectors with n real components
C
n
n-dimensional complex linear space
space of column vectors with n complex components
H Hilbert space
i
1
'z real part of the complex number z
z imaginary part of the complex number z
[z[ modulus of complex number z
[x +iy[ = (x
2
+y
2
)
1/2
, x, y R
T S subset T of set S
S T the intersection of the sets S and T
S T the union of the sets S and T
f(S) image of set S under mapping f
f g composition of two mappings (f g)(x) = f(g(x))
x column vector in C
n
x
T
transpose of x (row vector)
0 zero (column) vector
| . | norm
x y x
y scalar product (inner product) in C

n
x y vector product in R
3
A, B, C mn matrices
det(A) determinant of a square matrix A
tr(A) trace of a square matrix A
rank(A) rank of matrix A
A
T
transpose of matrix A
ix
A conjugate of matrix A
A
conjugate transpose of matrix A

A
conjugate transpose of matrix A

(notation used in physics)
A
1
inverse of square matrix A (if it exists)
I
n
n n unit matrix
I unit operator
0
n
n n zero matrix
AB matrix product of mn matrix A
and n p matrix B
A B Hadamard product (entry-wise product)
of mn matrices A and B
[A, B] := AB BA commutator for square matrices A and B
[A, B]
+
:= AB +BA anticommutator for square matrices A and B
AB Kronecker product of matrices A and B
AB Direct sum of matrices A and B
jk
Kronecker delta with
jk
= 1 for j = k
and
jk
= 0 for j ,= k
eigenvalue
real parameter
t time variable
H Hamilton operator
The Pauli spin matrices are used extensively in the book. They are given
by
x
:=
_
0 1
1 0
_
,
y
:=
_
0 i
i 0
_
,
z
:=
_
1 0
0 1
_
.
In some cases we will also use
1
,
2
and
3
to denote
x
,
y
and
z
.
x
Chapter 1
General
Problem 1. Let H be a Hilbert space with scalar product , ). Let
u, v H.
(i) Show that
[u, v)[ |u| |v| .
(ii) Show that
|u +v| |u| +|v| .
Problem 2. Consider a Hilbert space H with scalar product , ). The
scalar product implies a norm via |f|
2
:= f, f), where f H.
(i) Show that
|f +g|
2
+|f g|
2
= 2(|f|
2
+|g|
2
) .
(ii) Assume that f, g) = 0, where f, g H. Show that
|f +g|
2
= |f|
2
+|g|
2
.
Problem 3. Let f, g H. Use the Schwarz inequality
[f, g)[
2
f, f)g, g) = |f|
2
|g|
2
to prove the triangle inequality
|f +g| |f| +|g| .
1
2 Problems and Solutions
Problem 4. Consider a complex Hilbert space H and [
1
), [
2
) H. Let
c
1
, c
2
C. An antilinear operator K in this Hilbert space H is characterized
by
K(c
1
[
1
) +c
2
[
2
)) = c
1
K[
1
) +c
2
K[
2
) .
A comb is an antilinear operator K with zero expectation value for all states
[) of a certain complex Hilbert space H. This means
[K[) = [LC[) = [L[
) = 0
for all states [) H, where L is a linear operator and C is the complex
conjugation.
(i) Consider the two-dimensional Hilbert space H = C
2
. Find a unitary
2 2 matrix such that
[UC[) = 0 .
(ii) Consider the Pauli spin matrices with
0
= I
2
,
1
=
x
,
2
=
y
,
3
=
z
. Find
3
=0
3
=0
[
C[)g
,
[
C[)
where g
,
= diag(1, 1, 0, 1).
Problem 5. Let P be a nonzero projection operator in a Hilbert space
H. Show that |P| = 1.
Problem 6. A family,
j
jJ
of vectors in the Hilbert space, H, is
called a frame if for any f H there exist two constants A > 0 and
0 < B < , such that
A|f|
2
jJ
j
[f)[
2
B|f|
2
.
Consider the Hilbert space H = R
2
and the family of vectors
_
0
=
_
1
1
_
,
1
=
_
1
1
__
.
Show that we have a tight frame.
Problem 7. Let T : X Y be a linear map between linear spaces (vector
spaces) X, Y . The null space or kernel of the linear map T, denoted by
kerT, is the subset of X dened by
kerT := x X : Tx = 0 .
General 3
The range of T, denoted by ranT, is the subset of Y dened by
ranT := y Y : there exists x X such that Tx = y .
Let P be a projection operator in a Hilbert space H. Show that ranP is
closed and
H = ranP kerP
is the orthogonal direct sum of ranP and kerP.
Problem 8. Let H be an arbitrary Hilbert space with scalar product
, ). Show that if is a bounded linear functional on the Hilbert space H,
then there is a unique vector u H such that
(x) = u, x) for all x H.
Problem 9. Let H be an arbitrary Hilbert space. A bounded linear op-
erator A : H H satises the Fredholm alternative if one of the following
two alternatives holds:
(i) either Ax = 0, A
x = 0 have only the zero solution, and the linear

equations Ax = y, A
x = y have a unique solution x H for every y H;

(ii) or Ax = 0, A
x = 0 have nontrivial, nite-dimensional solution spaces

of the same dimension, Ax = y has a (nonunique) solution if and only if
y u for every solution u of A
u = 0, and A
x = y has a (nonunique)
solution if and only if y u for every solution u of Au = 0.
Give an example of a bounded linear operator that satises the Fredholm
alternative.
Problem 10. Let (M, d) be a complete metric space (for example a
Hilbert space) and let f : M M be a mapping such that
d(f
(m)
(x), f
(m)
(y)) kd(x, y), x, y M
for some m 1, where 0 k < 1 is a constant. Show that the map f has
a unique xed point in M.
Problem 11. Let H be a Hilbert space and let f : H H be a monotone
mapping such that for some constant > 0
|f(u) f(v)| |u v| u, v H.
Show that for any w H, the equation
u +f(u) = w
has a unique solution u.
Problem 12. Let f, g H. Find all solutions to the equations
f, g)g, f) = i .
Problem 13. Let f, g H. Show that
|f +g|
2
+|f g|
2
= 2(|f|
2
+|g|
2
)
where the norm is implied by the scalar product of the Hilbert space.
Problem 14. Show that
f, g) =
1
4
|f +g|
2
1
4
|f g|
2
or
f, g) =
1
4
|f +g|
2
1
4
|f g|
2
+
i
4
|f +ig|
2
i
4
|f ig|
2
depending on whether we are dealing with a real and complex Hilbert space.
Problem 15. Given a Hilbert space H and a Hilbert subspace ( of H.
The Hilbert space projection theorem states that for every f H, there
exists a unique g G such that
(i) f g G
(ii) |f g| = inf
hG
|f h|
where the space (
is dened by
(
:= k H : k[u) = 0 for all u ( .

Show that if g is the minimizer of |f h| over all h (, then it is true
that f g (
.
Problem 16. Let
n
nZ
be an orthonormal basis in a Hilbert space
H. Then any vector f H can be written as
f =
nZ
f,
n
)
n
.
Now suppose that
n
nZ
is also a basis for H, but it is not orthonormal.
Show that if we can nd a so-called dual basis
n
nZ
satisfying
n
[
m
) = (n m)
General 5
then for any vector f H, we have
f =
nZ
f[
n
)
n
.
Here (n m) denotes the Kronecker delta with (n m) = 0 if n = m
and 1 otherwise.
Problem 17. Let (X
1
, | |
1
) and (X
2
, | |
2
) be two normed spaces. Show
that the product vector spaces X = X
1
X
2
is also a normed vector space
if we dene
|x| := max(|x
1
|
1
, |x
2
|
2
)
with x = (x
1
, x
2
).
Problem 18. Let A be a linear bounded self-adjoint operator in a Hilbert
space H. Let u, v H and C. Consider the equation
Au u = v .
(i) Show that for nonreal (i.e. it has an imaginary part) v cannot vanish
unless u vanishes.
(ii) Show that for nonreal we have
|(AI)
1
v|
1
[[
|v| .
Problem 19. Let E be the exterior of the unit disc
z C : [z[ > 1
and T the unit circle
z C : [z[ = 1 .
Let H
2
(E) be the Hardy spacee of square integrable functions on T, analytic
in the region E. The inner product for f(z), g(z) H
2
(E) is dened by
f, g) =
1
2
_

f(e
i
)
g(e
i
)d =
1
2i
_
T
f
(1/z
)g(z)
dz
z
.
Let f(z) = z
2
and g(z) = z + 1. Find the scalar product f, g).
Problem 20. Let O = u
1
.u
2
, . . . be an orthonormal set in a innite
dimensional Hilbert space. Show that if
x =
j=1
c
j
u
j
then
|x|
2
=
j=1
[c
j
[
2
.
Problem 21. Two Cauchy sequences x
k
and y
k
are said to be
equivalent if for all > 0, there is a k(epsilon) such that for all j k()
we have d(x
j
, y
j
) < . One writes x
k
y
k
. Obviously, is an
equivalence relationship. Show that equivalent Cauchy sequences have the
same limit.
Problem 22. Consider the sequence x
k
, k = 1, 2, . . . in R dened
by k
k
= 1/k
2
for all k = 1, 2, . . .. Show that this sequence is a Cauchy
sequence.
Problem 23. Let H be a Hilbert space and o be a sub Hilbert space.
Show that any element u of H can be decomposed uniquely
u = v +w
where v is in o and w is in o
.
Problem 24. Let u, v
1
, v
2
be elements of a Hilbert space. Show that
2|u v
1
|
2
+ 2|u v
2
|
2
= |2
_
u
v
1
+v
2
2
_
|
2
+|v
1
v
2
|
2
.
Problem 25. Let P be the set of prime numbers. We dene the set
S := (p, q) : p, q P p q .
Show that
d((p
1
, q
1
), (p
2
, q
2
)) := [p
1
q
1
p
2
q
2
[
denes a metric.
Problem 26. Consider the vector space of all continuous functions de-
ned on [a, b]. We dene a metric
d(f, g) := max
axb
[f(x) g(x)[ .
Let a = , b = , f(x) = sin(x) and g(x) = cos(x). Find d(f, g).
General 7
Problem 27. The nn matrices over R form a vector space. Show that
d(A, B) :=
n
j=1
n
k=1
[a
jk
b
jk
[
denes a metric.
Problem 28. Let n 1. Consider the continuous function
f
n
(t) =
_
_
_
0 0 t < 1/2 1/n
1/2 +
n
2
(t 1/2) 1/2 1/n t 1/2 + 1/n
1 1/2 + 1/n t 1
Show that the sequence f
n
(t) is not a Cauchy sequence for the uniform
norm, but with any of the L
p
norms (1 p < ) it is a Cauchy sequence.
Problem 29. The sequence space consists of the set of all (bounded or
unbounded) sequences of complex
x = (
1
,
2
, . . .)
Thus we have a vector space. Can we dene a metric in this vector space
which is implied by a norm?
Chapter 2
Finite Dimensional
Hilbert Spaces
Problem 1. Consider the Hilbert space R
4
and the vectors
_
_
_
1
0
0
0
_
_
_,
_
_
_
1
1
0
0
_
_
_,
_
_
_
1
1
1
0
_
_
_,
_
_
_
1
1
1
1
_
_
_ .
(i) Show that the vectors are linearly independent.
(ii) Use the Gram-Schmidt orthogonalization process to nd mutually or-
thogonal vectors.
4
. Show that the vectors (Bell
basis)
1
2
_
_
_
1
0
0
1
_
_
_,
1
2
_
_
_
1
0
0
1
_
_
_,
1
2
_
_
_
0
1
1
0
_
_
_,
1
2
_
_
_
0
1
1
0
_
_
_
are linearly independent. Show that they form a orthonormal basis in the
Hilbert space R
4
.
4
. Find all pairwise orthogonal
vectors (column vectors) x
1
, . . . , x
p
, where the entries of the column vectors
8
Finite Dimensional Hilbert Spaces 9
can only be +1 or 1. Calculate the matrix
p
i=1
x
i
x
T
i
and nd the eigenvalues and eigenvectors of this matrix
Problem 4. A sequence f
n
(n N) of elements in a normed space E
is called a Cauchy sequence if, for every > 0, there exists a number M
,
such that |f
p
f
q
| < for p, q > M
. Consider the Hilbert space R. Show

that
s
n
=
n
j=1
1
(j 1)!
, n 1
is a Cauchy sequence.
Problem 5. Two Cauchy sequences x
k
and y
k
are said to be equiv-
alent if for all > 0, there is a k() such that for all j k() we have
d(x
j
, y
j
) < . One writes x
k
y
k
. Obviously, is an equivalence
relationship. Show that equivalent Cauchy sequences have the same limit.
Problem 6. Consider the sequence x
k
, k = 1, 2, . . . in R dened
by x
k
= 1/k
2
for all k = 1, 2, . . .. Show that this sequence is a Cauchy
sequence.
Problem 7. Consider the Hilbert space C
2
and the vectors
[0) =
_
i
i
_
, [1) =
_
1
1
_
.
Normalize these vectors and then calculate the probability [0[1)[
2
.
n
. Let x, y R
n
. Show that
|x +y|
2
+|x y|
2
2(|x|
2
+|y|
2
) .
Note that
|x|
2
:= x, y) .
Problem 9. Let [0), [1) be an orthonormal basis in the Hilbert space C
2
.
Let
[) = cos(/2)[0) +e
i
sin(/2)[1)
where , R.
(i) Find [).
(ii) Find the probability [0[)[
2
. Discuss [0[)[
2
as a function of .
(iii) Assume that
[0) =
_
1
0
_
, [1) =
_
0
1
_
.
Find the 2 2 matrix [)[ and calculate the eigenvalues.
2
. Show that the vectors
_
1
2
_
1
1
_
,
1
2
_
1
1
__
are linearly independent. Find
1
2
_
1
1
_
2
_
1
1
_
,
1
2
_
1
1
_
2
_
1
1
_
,
1
2
_
1
1
_
2
_
1
1
_
,
1
2
_
1
1
_
2
_
1
1
_
.
Show that these four vectors form a basis in R
4
. Consider the 4 4 matrix
Q which is constructed from the four vectors given above, i.e. the columns
of the 4 4 matrix are the four vectors. Find Q
T
. Is Q invertible? If so
nd the inverse Q
1
. What is the use of the matrix Q?
4
. Let A be a symmetric 44
matrix over R. Assume that the eigenvalues are given by
1
= 0,
2
= 1,
3
= 2 and
4
= 3 with the corresponding normalized eigenfunctions
u
1
=
1
2
_
_
_
1
0
0
1
_
_
_, u
2
=
1
2
_
_
_
1
0
0
1
_
_
_, u
3
=
1
2
_
_
_
0
1
1
0
_
_
_, u
4
=
1
2
_
_
_
0
1
1
0
_
_
_ .
Find the matrix A by means of the spectral theorem.
Problem 12. Show that the 2 2 matrices
A =
1
2
_
1 0
0 1
_
, B =
1
2
_
0 1
1 0
_
,
C =
1
2
_
0 i
i 0
_
, D =
1
2
_
1 0
1
_
form an orthonormal basis in the Hilbert space M
2
(C).
Problem 13. Consider the Hilbert space H of the 2 2 matrices over
the complex numbers with the scalar product
A, B) := tr(AB
), A, B H.
Show that the rescaled Pauli matrices
j
:=
1
j
, j = 1, 2, 3
1
=
1
2
_
0 1
1 0
_
,
2
=
1
2
_
0 i
i 0
_
,
3
=
1
2
_
1 0
0 1
_
plus the rescaled 2 2 identity matrix
0
=
1
2
_
1 0
0 1
_
form an orthonormal basis in the Hilbert space H.
Problem 14. Let A, B be two n n matrices over C. We introduce the
scalar product
A, B) :=
tr(AB
)
trI
n
=
1
n
tr(AB
) .
This provides us with a Hilbert space.
The Lie group SU(N) is dened by the complex n n matrices U
SU(N) := U : U
U = UU
= I
n
, det(U) = 1 .
The dimension is N
2
1. The Lie algebra su(N) is dened by the n n
matrices X
su(N) := X : X
= X , trX = 0 .
(i) Let U SU(N). Calculate U, U).
(ii) Let A be an arbitrary complex nn matrix. Let U SU(N). Calculate
UA, UA).
(iii) Consider the Lie algebra su(2). Provide a basis. The elements of the
basis should be orthogonal to each other with respect to the scalar product
given above. Calculate the commutators of these matrices.
Problem 15. Let

H = S
x
be a Hamilton operator, where
S
x
:=

2
_
_
0 1 0
1 0 1
0 1 0
_
_
and is the frequency.
(i) Find exp(i

Ht/)(0), where (0) = (1, 1, 1)
T
/
3.
(ii) Calculate the time evolution of
S
z
:=
_
_
1 0 0
0 0 0
0 0 1
_
_
using the Heisenberg equation of motion. The matrices S
x
, S
y
, S
z
are the
spin-1 matrices, where
S
y
:=

2
_
_
0 i 0
i 0 i
0 i 0
_
_
.
Problem 16. Consider the linear operator
A =
_
_
2 0 0
0 0 1
0 1 0
_
_
in the Hilbert space R
3
. Find
|A| := sup
x=1
|Ax|
using the method of the Lagrange multiplier.
4
. Show that the Bell basis
u
1
=
1
2
_
_
_
1
0
0
1
_
_
_, u
2
=
1
2
_
_
_
1
0
0
1
_
_
_, u
3
=
1
2
_
_
_
0
1
1
0
_
_
_, u
4
=
1
2
_
_
_
0
1
1
0
_
_
_
forms an orthonormal basis in this Hilbert space.
3
. Let x R
3
, where x is
considered as a column vector. Find the matrix xx
T
. Show that at least
one eigenvalue is equal to 0.
Problem 19. (i) Consider the Hilbert space C
4
. Show that the matrices
1
=
1
2
(I
2
I
2
+
x
x
),
2
=
1
2
(I
2
I
2
x
)
are projection matrices in C
4
.
(ii) Find
1
2
. Discuss.
(iii) Let e
1
, e
2
, e
3
, e
4
be the standard basis in C
4
. Calculate
1
e
j
,
2
e
j
, j = 1, 2, 3, 4 .
and show that we obtain 2 two-dimensional Hilbert spaces under these
projections.
Problem 20. Consider the 3 3 matrix
A =
_
_
2 0 2
1 0 0
0 0 1
_
_
.
(i) The matrix A can be considered as an element of the Hilbert space of
the 3 3 matrices with the scalar product A, B) := tr(AB
T
). Find the
norm of A with respect to this Hilbert space.
(ii) On the other hand A can be considered as a linear operator in the
Hilbert space R
3
. Find die norm
|A| := sup
x=1
|Ax|, x R
3
.
(iii) Find the eigenvalues of A and AA
T
. Compare the result with (i) and
(ii).
3
. Find the spectrum (eigen-
values and normalized eigenvectors) of matrix
A =
_
_
1 2 3
1 2 3
1 2 3
_
_
.
Find |A| := sup
x=1
|Ax|, where |.| denotes the norm and x R
3
.
Problem 22. Find the spectrum (eigenvalues and normalized eigenvec-
tors) of the 3 3 matrix
A =
_
_
3 3 3
3 3 3
3 3 3
_
_
.
Find |A|, where |.| denotes the norms
|A|
1
:= sup
x=1
|Ax|
|A|
2
:=
_
tr(AA
).
Compare the norms with the eigenvalues. Find exp(A).
Problem 23. Consider the Hilbert space M
4
(C) of all 44 matrices over
C with the scalar product A, B) := tr(AB
), where A, B M
4
(C). The
-matrices are given by
1
=
_
_
_
0 0 0 i
0 0 i 0
0 i 0 0
i 0 0 0
_
_
_,
2
=
_
_
_
0 0 0 1
0 0 1 0
0 1 0 0
1 0 0 0
_
_
_
3
=
_
_
_
0 0 i 0
0 0 0 i
i 0 0 0
0 i 0 0
_
_
_,
4
=
_
_
_
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
_
_
_
and
5
=
1
4
=
_
_
_
0 0 1 0
0 0 0 1
1 0 0 0
0 1 0 0
_
_
_ .
We dene the 4 4 matrices
jk
:=
i
2
[
j
,
k
], j < k
where j = 1, 2, 3, k = 2, 3, 4 and [ , ] denotes the commutator.
(i) Calculate
12
,
13
,
14
,
23
,
24
,
34
.
(ii) Do the 16 matrices
I
4
,
1
,
2
,
3
,
4
,
5
,
5
1
,
5
2
,
5
3
,
5
4
,
12
,
13
,
14
,
23
,
24
,
34
form a basis in the Hilbert space M
4
(C)? If so is the basis orthogonal?
Problem 24. Find the spectrum (eigenvalues and normalized eigenvec-
tors) of matrix
A =
_
_
1 1 1
1 1 1
1 1 1
_
_
.
Find |A|, where |.| denotes the norm.
Problem 25. Let A and B be two arbitrary matrices. Give the denition
of the Kronecker product. Let u
j
(j = 1, 2, . . . , m) be an orthonormal basis
m
. Let v
k
(k = 1, 2, . . . , n) be an orthonormal basis in
the Hilbert space R
n
. Show that u
j
v
k
(j = 1, 2, . . . , m), (k = 1, 2, . . . , n)
is an orthonormal basis in R
m+n
.
A =
1
2
_
1 0
0 1
_
, B =
1
2
_
0 1
1 0
_
,
C =
1
2
_
0 i
i 0
_
, D =
1
2
_
1 0
0 1
_
form an orthonormal basis in the Hilbert space M
2
(C).
A =
_
1 0
0 0
_
, B =
_
1 1
0 0
_
, C =
_
1 1
1 0
_
, D =
_
1 1
1 1
_
form a basis in the Hilbert space M
2
(R). Apply the Gram-Schmidt tech-
nique to obtain an orthonormal basis.
Problem 28. Consider the 3 3 matrices over the real numbers
A =
_
_
2 0 2
1 0 0
0 0 1
_
_
.
(i) The matrix A can be considered as an element of the Hilbert space of
the 3 3 matrices over the real nunbers with the scalar product
B, C) := tr(BC
T
) .
Find the norm of A with respect to this Hilbert space.
(ii) On the other hand the matrix A can be considered as a linear operator
3
. Find the norm
|A| := sup
x=1
|Ax|, x R
3
.
(iii) Find the eigenvalues of A and A
T
A. Compare the result with (i) and
(ii).
2
. The Pauli spin matrices
x
,
y
,
z
act as linear operators in this Hilbert space. Let
H =
z
be a Hamilton operator, where
z
=
_
1 0
0 1
_
and is the frequency. Calculate the time evolution (intial value problem)
of
x
=
_
0 1
1 0
_
i.e.
i
d
x
dt
= [
x
,

H](t).
The matrices
x
,
y
,
z
are the Pauli matrices, where
y
=
_
0 i
i 0
_
.
4
. Consider the Hamilton
operator
H :=
_
_
_
0 0 0 i
0 0 i 0
0 i 0 0
i 0 0 0
_
_
_ .
Find the time-evolution of the operator
3
=
_
_
_
0 0 i 0
0 0 0 i
i 0 0 0
0 i 0 0
_
_
_
using the Heisenberg equation of motion
i
d
3
dt
= [
3
,

H](t) .
Problem 31. Let M be any n n matrix. Let x = (x
1
, x
2
, . . .)
T
. The
linear operator A is dened by
Ax = (w
1
, w
2
, . . .)
T
where
w
j
:=
n
k=1
M
jk
x
k
, j = 1, 2, . . . , n
w
j
:=x
j
, j > n
and T(A) =
2
(N). Show that A is self-adjoint if the n n matrix M is
hermitian. Show that A is unitary if M is unitary.
n
. Let u
j
, j = 1, 2, . . . , n, and
v
j
, j = 1, 2, . . . , n be orthonormal bases in C
n
, where u
j
, v
j
are considered
as column vectors. Show that
U =
n
j=1
u
j
v
j
is a unitary n n matrix.
2
. Given the vectors
u
1
=
_
0
1
_
, u
2
=
_
3/2
1/2
_
, u
3
=
_
3/2
1/2
_
.
The three vectors u
1
, u
2
, u
3
are at 120 degrees of each other and are
normalized, i.e. |u
j
| = 1 for j = 1, 2, 3. Every given two-dimensional
vector v can be written as
v = c
1
u
1
+c
2
u
2
+c
3
u
3
, c
1
, c
2
, c
3
R
in many dierent ways. Given the vector v minimize
1
2
(c
2
1
+c
2
2
+c
2
3
)
subject to the two constraints
v c
1
u
1
c
2
u
2
c
3
u
3
= 0.
Problem 34. Let A, H be n n hermitian matrices, where H plays the
role of the Hamilton operator. The Heisenberg equations of motion is given
by
dA(t)
dt
=
i
[H, A(t)] .
with A = A(t = 0) = A(0). Let E
j
(j = 1, 2, . . . , n
2
) be an orthonormal
basis in the Hilbert space H of the n n matrices with scalar product
X, Y ) := tr(XY
), X, Y H.
Now A(t) can be expanded using this orthonormal basis as
A(t) =
n
2
j=1
c
j
(t)E
j
and H can be expanded as
H =
n
2
j=1
h
j
E
j
.
Find the time evolution for the coecients c
j
(t), i.e. dc
j
/dt, where j =
1, 2, . . . , n
2
.
Problem 35. The sequence space consists of the set of all (bounded or
unbounded) sequences of complex numbers
x = (x
1
, x
2
, . . .)
Thus we have a vector space. Can we dene a metric in this vector space
which is not implied by a norm?
Chapter 3
Hilbert Space L
2
()
Problem 1. A basis in the Hilbert space L
2
[0, 1] is given by
B :=
_
e
2ixn
: n Z
_
.
Let
f(x) =
_
2x 0 x < 1/2
2(1 x) 1/2 x < 1
Is f L
2
[0, 1]? Find the rst two expansion coecients of the Fourier
expansion of f with respect to the basis given above.
Problem 2. (i) Consider the Hilbert space L
2
[1, 1]. Consider the se-
quence
f
n
(x) =
_
_
_
1 if 1 x 1/n
nx if 1/n x 1/n
+1 if 1/n x 1
where n = 1, 2, . . .. Show that f
n
(x) is a sequence in L
2
[1, 1] that is a
Cauchy sequence in the norm of L
2
[1, 1].
(ii) Show that f
n
(x) converges in the norm of L
2
[1, 1] to
sgn(x) =
_
1 if 1 x < 0
+1 if 0 < x 1
.
(iii) Use this sequence to show that the space C[1, 1] is a subspace of
L
2
[1, 1] that is not closed.
19
Problem 3. Let f L
2
(R). Give the denition of the Fourier transform.
Let us call the transformed function

f. Is

f L
2
(R)? What is preserved
under the Fourier transform?
Problem 4. Consider the Hilbert space L
2
[a, b], where a, b R and b > a.
Find the condition on a and b such that
cos(x), sin(x)) = 0
where , ) denotes the scalar product in L
2
[a, b].
Hint. Since b > a, we can write b = x +, where > 0.
2
[0, 1]. The Legendre polyno-
mials are dened as
P
0
(x) = 1, P
n
(x) =
1
2
n
n!
d
n
dx
n
(x
2
1).
Show that the rst rst four elements are given by
P
0
(x) = 1, P
1
(x) = x, P
2
(x) =
1
2
(3x
2
1), P
3
(x) =
1
2
(5x
3
3x) .
Normalize the four elements. Show that the four elements are pairwise
orthonormal.
Problem 6. Let R be a bounded region in n-dimensional space. Consider
the eigenvalue problem
u = u, u(q R) = 0
where R denotes the boundary of R.
(i) Show that all eigenvalues are real and positive
(ii) Show that the eigenfunctions which belong to dierent eigenvalues are
orthogonal.
Problem 7. Consider the inner product space
C[a, b] = f(x) : f is continuous on x [a, b]
with the inner product
f, g) :=
_
b
a
f(x)g
(x)dx.
This implies a norm
f, f) =
_
b
a
f(x)f
(x)dx = |f|
2
.
Hilbert Space L
2
() 21
Show that C[a, b] is incomplete. This means nd a Cauchy sequence in
the space C[a, b] which converges to an element which is not in the space
C[a, b].
2
[, ]. Given the function
f(x) =
_
_
_
1 0 < x
0 x = 0
1 < 0
Obviously f L
2
[, ]. Find the Fourier expansion of f. The orthonor-
mal basis B is given by
B :=
_
k
(x) =
1
2
exp(ikx) k Z
_
.
Find the approximation a
0
0
(x) +a
1
1
(x) +a
1
1
(x), where a
0
, a
1
, a
1
are the Fourier coecients.
Problem 9. Consider the linear operator A in the Hilbert space L
2
[0, 1]
dened by Af(x) := xf(x). Find the matrix elements
P
i
, AP
j
)
for i, j = 0, 1, 2, 3, where P
i
are the (normalized) Legrende polynomials. Is
the matrix A
ij
symmetric?
2
[0, 2). Let
g(x) = cos(x), f(x) = x.
Find the conditions on the coecients of the polynomial
p(x) = a
3
x
3
+a
2
x
2
+a
1
x +a
0
such that
g(x), p(x)) = 0, f(x), p(x)) = 0 .
Solve the equations for a
3
, a
2
, a
1
, a
0
.
2
(R). Give the denition
and an example of an even function in L
2
(R). Give the denition and an
example of an odd function in L
2
(R). Show that any function f L
2
(R)
can be written as a combination of an even and an odd function.
Problem 12. The Chebyshev polynomials T
n
(x) of the 1-st kind are
dened for x [1, 1] and given by
T
n
(x) = cos(narccos x), n = 0, 1, 2, . . .
The Chebyshev polynomials U
n
(x) of the 2-nd kind are dened for x
[1, 1] and given by
U
n
(x) =
sin((n + 1) arccos x)
1 x
2
, n = 0, 1, 2, . . . .
Consider the Hilbert spaces
H
1
= L
2
_
[1, 1],
dx
1 x
2
_
, H
2
= L
2
_
[1, 1],
2
1 x
2
dx
_
which bases are formed by the Chebyshev polynomials of the 1-st and 2-nd
type
(1)
n
(x) =
2T
n
(x), n 1,
(1)
0
= T
0
(x) = 1
(2)
n
(x) =U
n
(x), n 0
Find a recursion relation for
(1)
n
and
(2)
n
.
2
[, ]. Obviously cos(x)
L
2
[, ]. Find the norm | cos(x)|. Find nontrivial functions f, g
L
2
[, ] such that
f(x), cos(x)) = 0, g(x), cos(x)) = 0
and
f(x), g(x)) = 0 .
2
[0, 1]. Find a non-trivial
polynomial p
p(x) = ax
3
+bx
2
+cx +d
such that
p, 1) = 0, p, x) = 0, p, x
2
) = 0 .
Problem 15. Consider the set of polynomials
1, x, x
2
, . . . , x
n
, . . . .
Use the Gram-Schmidt procedure and the inner product
f, g) =
_
b
a
f(x)g(x)(x)dx, (x) > 0
to obtain the rst four orthogonal polynomials when
Hilbert Space L
2
() 23
(i) a = 1, b = 1, (x) = 1 (Legrendre polynomials)
(ii) a = 1, b = 1, (x) = (1 x
2
)
1/2
(Chebyshev polynomials)
(iii) a = 0, b = +, (x) = e
x
(Laguerre polynomials)
(iv) a = , b = +, (x) = e
x
2
(Hermite polynomials)
Problem 16. Consider the function
f(x) =
j=0
1
2
j
cos(jx) .
Is f an element of L
2
[, ]?
2
([0, 1]). The shifted Legrendre
polynomials , dened on the interval [0, 1], are obtained from the Legren-
dre polynomial by the transformation y = 2x 1. The shifted Legrendre
polynomials are given by the recurrence formula
P
j
(x) =
(2j + 1)(2x 1)
j + 1
P
j
(x)
j
j + 1
P
j1
(x) j = 1, 2, . . .
and P
0
(x) = 1, P
1
(x) = 2x 1. They are elements of the Hilbert space
L
2
([0, 1]). A function u in the Hilbert space L
2
([0, 1]) can be approximated
in the form of a series with n + 1 terms
u(x) =
n
j=0
c
j
P
j
(x)
where the coeents c
j
R, j = 0, 1, . . . , n. Consider the Volterra integral
equation of rst kind
_
x
0
y(t)
(x t)
dt = f(x), 0 t x 1
with 0 < < 1 and f L
2
([0, 1]). Consider the ansatz
y
n
(x) = a
0
x
+
n
j=0
c
j
P
j
(x) .
to nd an approximate solution to the Volterra integral equation of rst
kind ( = 1/2)
_
x
0
y(t)
x t
dt = f(x)
where
f(x) =
2
105
x(105 56x
2
+ 48x
3
) .
Problem 18. The Fock space T is the Hilbert space of entire functions
with inner product given by
f[g) :=
1
_
C
f(z)g(z)e
|z|
2
dxdy, z = x +iy
where C denotes the complex numbers. Therefore the growth of functions
in the Hilbert space T is dominated by exp([z[
2
/2). Let f, g T with
Taylor expansions
f(z) =
j=0
a
j
z
j
, g(z) =
j=0
b
j
z
j
.
(i) Find f[g) and |f|
2
.
(ii) Consider the special that f(z) = sin(z) and g(z) = cos(z). Calculate
f[g).
(iii) Let
/(z, w) := e
zw
, z, w C
Calculate f(z)[/(z, w)).
2
[0, ]. Let | . | be the norm
induced by the scalar product of L
2
[0, ]. Find the constants a, b such that
| sin(x) (ax
2
+bx)|
is a minimum.
2
(R). Let
f
n
(x) =
x
1 +nx
2
, n = 1, 2, . . . .
(i) Find |f
n
(x)| and
lim
n
|f
n
(x)| .
(ii) Does the sequence f
n
(x) converge uniformely on the real line?
Problem 21. Let n = 1, 2, . . .. We dene the functions f
n
L
2
[0, ) by
f
n
(x) =
_
n for n x n + 1/n
0 otherwise
(i) Calculate the norm |f
n
f
m
| implied by the scalar product. Does the
sequence f
n
converge in the L
2
[0, ) norm?
(ii) Show that f
n
(x) converges pointwise in the domain [0, ) and nd the
limit. Does the sequence converge pointwise uniformly?
Hilbert Space L
2
() 25
(iii) Show that f
n
(n = 1, 2, . . .) is an orthonormal system. Is it a basis
in the Hilbert space L
2
[0, )?
Problem 22. Consider the function f L
2
[0, 1]
f(x) =
_
x for 0 x 1/2
1 x for 1/2 x 1
A basis in the Hilbert space is given by
B :=
_
1,
2 cos(nx) : n = 1, 2, . . .
_
.
Find the Fourier expansion of f with respect to this basis. From this
expansion show that
2
8
=
k=0
1
(2k + 1)
2
.
Problem 23. A particle is enclosed in a rectangular box with impene-
trable walls, inside which it can move freely. The Hilbert space is
L
2
([0, a] [0, b] [0, c])
where a, b, c > 0. Find the eigenfunctions and the eigenvalues. What can
be said about the degeneracy, if any, of the eigenfunctions?
2
function
f(x) = ax
3
+bx
2
+cx +d .
such that
f(x), x) = 0, f(x), x
2
) = 0, f(x), x
3
) = 0
where , ) denotes the scalar product
2
function f such that
f(x), x) = 0, f(x), x
2
) = 0, f(x), x
3
) = 0
where , ) denotes the scalar product
2
[0, 1] and the polynomials
1, x, x
2
, x
3
, x
4
.
Apply the Gram-Schmidt orthogonalization process to these polynomials.
2
(T). Let f L
2
(T). Give an
example of a bounded linear functional.
2
(R). Show that the Hilbert
space is the direct sum of the Hilbert space / of even functions and the
Hilbert space ^ of odd functions. Give an example of such functions in
this Hilbert space.
Problem 29. Let a > 0. Consider the Hilbert space L
2
[0, a]. Let
Af(x) := xf(x)
for f L
2
[0, a]. Find the norm of the operator A. We dene
|A| := sup
f=1
|Af| .
2
[0, 2]. Let
g(x) = cos(x), f(x) = x.
Find the conditions on the coecients of the polynomial
p(x) = a
3
x
3
+a
2
x
2
+a
1
x +a
0
such that
g(x), p(x)) = 0, f(x), p(x)) = 0 .
Solve the equations for a
3
, a
2
, a
1
, a
0
.
Problem 31. Consider the function f : R R
f(x) =
1 cos(2x)
x
.
Using LHospital rules we have f(0) = 0. Is f L
2
(R)?
2
[1, 1]. The Legendre poly-
nomials are given by
P
j
(x) :=
1
2
j
j!
d
j
dx
j
(x
2
1)
j
.
Hilbert Space L
2
() 27
Find the scalar product
P
j
(x), P
k
(x)) .
Problem 33. Consider the Hilbert space H = L
2
(T). This is the vector
space of 2-periodic functions. Then
u(x) =
1
2
is a constant function which is normalized, i.e. |u| = 1. Show that the
projection operator P
u
dened by
P
u
f := u, f)u
maps a function f to its mean. This means
P
u
f = f), f) =
_
2
0
f(x)dx.
2
[, ] and the vector space
of continiuous real-valued functions C[, ] on the interval [, ]. Let
k > 0 and
f
k
(x) =
_
_
_
0 if x 0
kx if 0 x 1/k
1 if 1/k x
The sequence of functions f
k
belongs to the vector space C[, ].
(i) Show that f
n
in the norm of the Hilbert space L
2
[, ], where
(x) :=
_
0 if x 0
1 if 0 < x
so that the sequence f
k
is a Cauchy sequence in the Hilbert space
L
2
[, ].
(ii) Show that |g| > 0 for every g C[, ]. Conclude that C[, ]
is not a Hilbert space.
Problem 35. Let be the unit disk. A Hilbert space of analytic functions
can be dened by
H :=
_
f(z) analytic, [z[ < 1 : sup
a<1
_
|z|=a
[f(z)[
2
ds <
_
and the scalar product
f, g) := lim
a1
_
|z|=a
f(z)g(z)ds .
Let c
n
(n = 0, 1, 2, . . .) be the coecients of the power-series expansion of
the analytic function f. Find the norm of f.
Problem 36. Let C
n
denote the complex Euclidean space. Let z =
(z
1
, . . . , z
n
) C
n
and w = (w
1
, . . . , w
n
) C
n
then the scalar product
(inner product) is given by
z w := zw
= zw
T
where z = (z
1
, . . . , z
n
). Let E
n
denote the set of entire functions in C
n
.
Let F
n
denote the set of f E
n
such that
|f|
2
:=
1
n
_
C
n
[f(z)[
2
exp([z[
2
)dV
is nite. Here dV is the volume element (Lebesgue mesure)
dV =
n
j=1
dx
j
dy
j
=
n
j=1
r
j
dr
j
d
j
with z
j
= r
j
e
ij
. The norm follows from the scalar product of two functions
f, g F
n
f, g) :=
1
n
_
C
n
f(z)g(z) exp([z[
2
)dV .
Let
z
m
:= z
m1
1
z
mn
n
where the multiindex m is dened by m! = m
1
! m
n
! and [m[ =
n
j=1
m
j
.
z
m
, z
p
) .
Problem 37. Let be a complex-valued dierentiable function of in
the interval [0, 2] and (0) = (2), i.e. is an element of the Hilbert
space L
2
([0, 2]). Assume that (normalization condition)
_
2
0
()()d = 1 .
Calculate
i
_
2
0
()
d
d
()d
where denotes the imaginary part.
Hilbert Space L
2
() 29
Problem 38. The Legrendre polynomials are dened on the interval
[1, 1] and dened by the recurrence formula
L
j
(x) =
2j + 1
j + 1
xL
j
(x)
j
j + 1
L
j1
(x) j = 1, 2, . . .
and L
0
(y) = 1, L
1
(x) = x. They are elements of the Hilbert space
L
2
([1, 1]). Calculate the scalar product
L
j
(x), L
k
(x))
for j, k = 0, 1, . . .. Discuss.
Problem 39. Let f
n
: [1, 1] [1, 1] be dened by
f
n
(x) =
_
_
_
1 for 1 x 0
1 nx for 0 x 1/n
0 for 1/n x 1
Show that f
n
L
2
[1, 1]. Show that f
n
is a Cauchy sequence.
2
([1, 1]). The Chebyshev
polynomials are dened by
T
n
(x) := cos(ncos
1
x), n = 0, 1, 2, . . . .
They are elements of the Hilbert space L
2
([1, 1]). We have
T
0
(x) = 1, T
1
(x) = x, T
2
(x) = 2x
2
1, T
3
(x) = 4x
3
3x.
Calculate the scalar products
T
0
, T
1
), T
1
, T
2
), T
2
, T
3
) .
Calculate the integrals
_
1
1
T
m
(x)T
n
(x)
1 x
2
dx
for (m, n) = (0, 1), (m, n) = (1, 2), (m, n) = (2, 3).
2
[0, 1] with the scalar
product , ). Let f : [0, 1] [0, 1]
f(x) :=
_
2x if x [0, 1/2)
2(1 x) if x [1/2, 1]
Thus f L
2
[0, 1]. Calculate the moments
k
, k = 0, 1, 2, . . . dened by
k
:= f(x), x
k
)
_
1
0
f(x)x
k
dx.
(ii) Show that
k=0
[
k
[
2
<
_
1
0
[f(x)[
2
dx.
Problem 42. Let a, b R and < a < b < +. Let f be a function
in the class C
1
(i.e., the derivative df/dt exists and is continuous) on the
interval [a, b]. Thus f is also an element of the Hilbert space L
2
([a, b]).
Show that
lim
_
b
a
f(t) sin(t)dt = 0 . (1)
Problem 43. Consider the Lie group
SU(1, 1) =
__

_
[[
2
[[
2
= 1
_
.
The elements of this Lie group act as analytic automorphism of the disk
:= [z[ < 1
under
z zg =
z +
z +
where (zg)h = z(gh). Let n 2. We dene
H
n
:= f(z) analytic in , |f|
2
=
_
[f(z)[
2
(1 [z[
2
)
n2
dxdy <
and
U
n
(g)f(z) :=
1
(
z +)
n
f(( z +)/(
z +)) .
Then H
n
is a Hilbert space, i.e., the analytic functions in
L
2
(, (1 [z[
2
)
n2
dxdy)
form a closed subspace. U
n
is a representation, i.e.,
U
n
(gh) = U
n
(g)U
n
(h)
Hilbert Space L
2
() 31
and U
n
(e) = I, where e is the identity element in SU(1, 1) (2 2 unit
matrix).
Show that
1
(1 [z[
2
)
2
dx dy
is invariant z zg.
Problem 44. Consider the problem of a particle in a one-dimensional
box. The underlying Hilbert space is L
2
(a, a). Solve the Schrodinger
equation
i
t
=

H
as follows: The formal solution is given by
(t) = exp(i

Ht/)(0) .
Expand (0) with respect to the eigenfunctions of the operator

H. The
eigenfunctions form a basis of the Hilbert space. Then apply exp(i

Ht/).
Calculate the probability
P = [, (t))[
2
where
(q) =
1
a
sin
_
q
a
_
and
(q, 0) =
1
a
sin
_
q
a
_
.
Problem 45. Let f L
2
(R
n
). Consider the following operators
T
y
f(x) =f(x y), translation operator
M
k
f(x) =e
ixk
f(x), modulation operator
D
s
f(x) =[s[
n/2
f(s
1
x), s R 0 dilation operator
where x k = k
1
x
1
+ +x
n
k
n
.
(i) Find |T
y
f|, |M
k
|, |D
s
f|, where | | denotes the norm in L
2
(R
n
).
(ii) Find the adjoint operators of these three operators.
Problem 46. Consider the vector space
H
1
(a, b) := f(x) L
2
(a, b) : f
(x) L
2
(a, b)
with the norm g H
1
(a, b))
|g|
1
:=
_
|g|
2
0
+|g/x|
2
0
.
Consider the Hilbert space L
2
(, ) and f(x) = sin(x). Find the norm
|f|
1
.
Problem 47. Let f H
1
(a, b). Then for a x < y b we have
f(y) = f(x) +
_
y
x
f
(s)ds .
(i) Show that f C[a, b].
(ii) Show that
[f(y) f(x)[ |f|
1
_
[y x[ .
2
[0, ). The Laguerre poly-
nomials are dened by
L
n
(x) = e
x
d
n
dx
n
(x
n
e
x
), n = 0, 1, 2, . . . .
The rst ve Laguerre polynomials are given by
L
0
(x) =1
L
1
(x) =1 x
L
2
(x) =2 4x +x
2
L
3
(x) =6 18x + 9x
2
x
3
L
4
(x) =24 96x + 72x
2
16x
3
+x
4
.
Show that the function
n
(x) =
1
n!
e
x/2
L
n
(x)
form an orthonormal system in the Hilbert space L
2
[0, ).
2
[, ]. A basis in this
Hilbert space is given by
B =
_
1
2
e
ikx
: k Z
_
.
Find the Fourier expansion of
f(x) = 1 .
Hilbert Space L
2
() 33
Problem 50. (i) Consider the functions
f(x) =
1
1 +x
2
, g(x) =
x
1 +x
2
.
Obviously f, g L
2
(R). Calculate the scalar product
f, g) =
_

f(x)g(x)dx.
(ii) Let > 0. Consider the functions
f(t) =
sin(t)
t
, g(t) =
1 cos(t)
t
.
Obvioulsy f(0) = 1, g(0) = 0 and f, g L
2
(R). Calculate the scalar
product
f, g) =
_

f(t)g(t)dt .
2
[0, 1]. Let T
n
be the n + 1-
dimensional real linear space of all polynomial of maximal degree n in the
variable x, i.e.
T
n
= span 1, x, x
2
, . . . , x
n
.
The linear space T
n
can be spanned by various systems of basis functions.
An important basis is formed by the Bernstein polynomials
B
n
0
(x), B
n
1
(x), . . . , B
n
n
(x)
of degree n with
B
n
i
(x) := x
i
(1 x)
ni
, i = 0, 1, . . . , n.
The Bernstein polynomials have a unique dual basis
D
n
0
(x), D
n
1
(x), . . . , D
n
n
(x)
which consists of the n + 1 dual basis functions
D
n
i
(x) =
n
j=0
c
ij
B
n
j
(x) .
The dual basis functions satisfy
D
n
i
(x), B
n
j
(x)) =
ij
.
(i) Calculate the scalar product
B
m
i
(x), B
n
j
(x)) .
(ii) Find the coecients c
ij
.
Problem 52. Consider Fourier series and analytic (harmonic) functions
on the disc
D := z C : [z[ 1 .
A Fourier series can be viewed as the boundary values of a Laurent series
n=
c
n
z
n
.
Suppose we are given a function f on T. Find the harmonic extension u of
f into D. This means
u = 0 and u = f on D = T
where :=
2
/x
2
+
2
/y
2
.
Problem 53. Consider the compact abelian Lie group U(1)
U(1) = e
2i
: 0 < 1 .
The Hilbert space L
2
(U(1)) is the space L
2
([0, 1]) consisting of all measure-
able funcrions f() with period 1 such that
_
1
0
[f()[
2
d < .
Now the set of functions
e
2im
: m Z
form an orthonormal basis for the Hilbert space L
2
([0, 1]). Thus every
f L
2
([0, 1]) can be expressed uniquely as
f() =
+
m=
c
m
e
2im
, c
m
=
_
1
0
f()e
2im
d .
Calculate
_
1
0
[f()[
2
.
Hilbert Space L
2
() 35
Problem 54. The Hilbert space L
2
(R) is the vector space of measur-
able functions dened almost everywhere on R such that [f[
2
is integrable.
H
1
(R) is the vector space of functions with rst derivatives in L
2
(R). Give
two examples of such a function.
2
[, ]. The set of functions
_
1
2
e
inx
_
nZ
is an orthonormal basis for L
2
[, ]. Let
K(x, t) =
1
2
e
itx
.
For t xed nd the Fourier expansion of this function.
Problem 56. Consider the vector space C([0, 1]) of continouos functions.
We dene the triangle function
(x) :=
_
2x 0 x 1/2
2 2x 1/2 < x 1
.
Let
0
(x) := x and
n
(x) := (2
j
x k)
where j = 0, 1, 2, . . ., n = 2
j
+k and 0 k < 2
j
. The functions
1,
0
,
1
, . . .
are the Schauder basis for the vector space C([0, 1]). Let f C([0, 1]).
Then
f(x) = a +bx +
n=1
c
n
n
(x) .
(i) Find the Schauder coecients a, b, c
n
.
(ii) Consider g : [0, 1] [0, 1]
g(x) = 4x(1 x) .
Find the Schauder coecients for this function.
Problem 57. Let s be a nonnegative integer. Let x R and h
n
(n =
0, 1, 2, . . . be
h
n
(x) =
(1)
n
2
n/2
n!
4
exp(x
2
/2)
d
n
e
x
2
dx
n
.
Ths h
n
for an orthonormal basis in the Hilbert space L
2
(R). Consider the
sequence
f
s
(x) =
1
s + 1
s
n=0
e
in
h
n
(x)
where s = 0, 1, 2, . . .. Show that the sequence converges weakly but not
strongly to 0.
Problem 58. Let C
nN
be the vector space of all nN complex matrices.
Let Z C
nN
. Then Z

Z
T
, where
T
denotes transpose. One denes a
Gaussian measure on C
nN
by
d(Z) :=
1
nN
exp(tr(ZZ
))dZ
where dZ denotes the Lebesgue measure on C
nN
. The Fock space T(C
nN
)
consists of all entire functions on C
nN
which are square integrable with
respect to the Gaussian measure d(Z). With the scalar product
f[g) :=
_
C
nN
f(Z)g(Z)d(Z), f, g F(C
nN
)
one has a Hilbert space. Show that this Hilbert space has a reproducing
kernel K. This means a continuous function K(Z, Z
) : C
nN
C
nN
C
such that
f(Z) =
_
C
nN
K(Z, Z
)f(Z
)d(Z
)
for all Z C
nN
and f T(C
nN
).
2
[0, ) and the function f
L
2
[0, )
f(x) = exp(u
1/4
) sin(u
1/4
) .
Find
_

0
f(x)x
n
dx, n = 0, 1, 2, . . . .
2
[0, 2) with the scalar prod-
uct
f
1
, f
2
) =
1
2
_
2
0
f
1
(e
i
)f
2
(e
i
)d.
(i) Let f
1
(z) = z and f
2
(z) = z
2
. Find f
1
, f
2
).
(ii) Let f
1
(z) = z
2
and f
2
(z) = sin(z). Find f
1
, f
2
).
Hilbert Space L
2
() 37
2
(R
2
) with the basis
mn
(x
1
, x
2
) = NH
m
(x
1
)H
n
(x
2
)e
(x
2
1
+x
2
2
)/2
where m, n = 0, 1, . . . and N is the normalization factor. Consider the
two-dimensional potential
V (x
1
, x
2
) =
a
4
(x
4
1
+x
4
2
) +cx
1
x
2
.
(i) Find all linear transformation T : R
2
R
2
such that
V (Tx) = V (x) .
(ii) Show that these 2 2 matrices form a group. Is the group abelian.
(iii) Find the conjugacy classes and the irreducible representations.
(iv) Consider the Hilbert space L
2
(R
2
) with the orthogonal basis
mn
(x
1
, x
2
) = H
m
(x
1
)e
x
2
1
/2
H
n
(x
2
)e
x
2
2
/2
where m, n = 0, 1, 2, . . .. Find the invariant subspaces from the projection
operators of the irreducible representations.
2
[, ]. A basis in this
Hilbert space is given by
B =
_
1
2
e
ikx
: k Z
_
Find the Fourier expansion of
f(x) = 1 .
2
[0, 1]. Let T
n
be the n + 1-
dimensional real linear space of all polynomial of maximal degree n in the
variable x, i.e.
T
n
= span 1, x, x
2
, . . . , x
n
.
The linear space T
n
can be spanned by various systems of basis functions.
An important basis is formed by the Bernstein polynomials B
n
0
(x), B
n
1
(x), . . . , B
n
n
(x)
of degree n with
B
n
i
(x) := x
i
(1 x)
ni
, i = 0, 1, . . . , n.
The Bernstein polynomials have a unique dual basis D
(
0
x), D
n
1
(x), . . . , D
n
n
(x)
which consists of the n + 1 dual basis functions
D
n
i
(x) =
n
j=0
c
ij
B
n
j
(x) .
The dual basis functions satisfy
D
n
i
(x), B
n
j
(x)) =
ij
.
(i) Find the scalar product
B
m
i
(x), B
n
j
(x)) .
(ii) Find the coecients c
ij
.
Problem 64. Let V be a metric vector space. A reproducing kernel
Hilbert space on V is a Hilbert space H of functions on V such that for each
x V , the point evaluation functional
x
(f) := f(x), f H
is continouos. A reproducing kernel Hilbert space H possesses a unique
reproducing kernel K which is a function on V V characterized by the
properties that for all f H and x V , K(x, ) H and
f(x) = f, K(x, ))
H
where , )
H
denotes the inner product on H. The reproducing kernel K
uniquely determines the reproducing kernel Hilbert space H. The repro-
ducing kernel Hilbert space of a reproducing kernel K is denoted by H
K
.
The Paley-Wiener space is dened by
S := f C(R
d
) L
2
(R
d
) : supp

f [, ]
d
is a reproducing kernel Hilbert space. The Fourier transform of f L

1
(R
d
)
is given by
f(k) :=
1
(
2)
2d
_
R
2d
f(x)e
ixk
dx, k R
d
where x k = x
1
k
1
+ +x
d
k
d
is the inner product in R
d
. The norm on the
vector space S inherits from that in L
2
(R
d
). Show that the reproducing
kernel for the Paley-Wiener space S is the sinc function
sinc(x, y) :=
d
j=1
sin((x
j
y
j
))
(x
j
y
j
)
, x, y R
d
.
Problem 65. Can one construct an orthonormal basis in the Hilbert
space L
2
(R) starting from ( > 0)
e
|x|/
, xe
|x|/
, x
2
e
|x|/
, x
3
e
|x|/
, . . . .
Hilbert Space L
2
() 39
2
[1, 1]. Normalize the func-
tion f(x) = x in this Hilbert space.
Problem 67. Show that (Mehlers formula)
exp((u
2
+v
2
2uvz)/(1z
2
)) = (1z
2
)
1/2
exp((u
2
+v
2
))
n=0
z
n
n!
H
n
(u)H
n
(v)
where H
n
are the Hermite polynomials.
2
(R). Let j, k = 1, 2, . . ..
Consider the functions
f
j
(x) = x
j
e
|x|
, f
k
(x) = x
k
e
|x|
.
(f
j
(x), f
k
(x)) :=
_

f
j
(x)

f
k
(x)dx =
_

f
j
(x)f
k
(x)dx.
Discuss.
2
(R) and the one-dimensional
Schrodinger equation (eigenvalue equation)
_
d
2
dx
2
+V (x)
_
u(x) = Eu(x)
where the potential V is given by
V (x) = x
2
+
ax
2
1 +bx
2
where b > 0. Insert the ansatz
u(x) = e
x
2
/2
v(x)
and nd the dierential equation for v. Discuss. Make a polynomial ansatz
for v.
2
(R). Let g > 0. Consider
the one-dimensional Schrodinger equation (eigenvalue equation)
_
d
2
dx
2
+x
2
+
x
2
1 +gx
2
_
u(x) = Eu(x) .
Find a solution of the second order dierential equation by making the
ansatz
u(x) = A(1 +gx
2
) exp(x
2
/2) .
2
[1/2, 1/2]. Show that
the following sets
B
1
:=
k
(x) = exp(2ikx), k Z
B
2
:=
k
(x) =
2 sin(2kx), k N
each form an orthonormal basis in this Hilbert space.
(ii) Expand the step function
f(x) =
_
1 for x [1/2, 0]
1 for x [0, 1/2]
with respect to the basis B
1
and with respect to the basis B
2
. Show that
the two expansions are equivalent. Recall that
2 sin(x) sin(y) cos(x y) cos(x +y) .
Problem 72. Consider the problem of a free particle in a one-dimensional
box [a, a]. The underlying Hilbert space is L
2
[a, a]. An orthonormal
basis in L
2
[a, a] is given by
B = u
(+)
k
(q), u
()
k
(q) : k N
where
u
(+)
k
=
1
a
cos
_
(k 1/2)q
a
_
, u
()
k
=
1
a
sin
_
kq
a
_
.
The formal solution of the initial value problem of the Schrodinger equation
i
t
=

H
is given by
(t) = exp(i

Ht/)(0) .
Let
(q, 0) =
1
a
sin(q/a), (q) =
1
a
sin(q/a) .
Find exp(i

Ht/) and P = [, (t))[
2
.
Chapter 4
Hilbert Space
2
(N)
Problem 1. Consider the Hilbert space
2
(N). Let x = (x
1
, x
2
, . . .)
T
be
an element of
2
(N). We dene the linear operator A in
2
(N) as
Ax = (x
2
, x
3
, . . .)
T
i.e. x
1
is omitted and the n+1st coordinate replaces the nth for n = 1, 2, . . ..
Then for the domain we have T(A) =
2
(N). Find A
y and the domain of

A
, where y = (y
1
, y
,
. . .). Is A unitary?
2
(N) and x = (x
1
, x
2
, . . .)
T
. The
linear bounded operator A is dened by
A(x
1
, x
2
, x
3
, . . . , x
2n
, x
2n+1
, . . .)
T
= (x
2
, x
4
, x
1
, x
6
, x
3
, x
8
, x
5
, . . . , x
2n+2
, x
2n1
, . . .)
T
.
Show that the operator A is unitary. Show that the point spectrum of A is
empty and the continuous spectrum is the entire unit circle in the -plane.
2
(N). Suppose that S and T
are the right and left shift linear operators on this sequence space, dened
by
S(x
1
, x
2
, . . .) = (0, x
1
, x
2
, . . .), T(x
1
, x
2
, x
3
, . . .) = (x
2
, x
3
, x
4
, . . .) .
Show that T = S
.
41
Problem 4. Find the spectrum of the innite dimensional matrix
A =
_
_
_
_
_
_
0 1 0 0 0 . . .
1 0 1 0 0 . . .
0 1 0 1 0 . . .
0 0 1 0 1 . . .
.
.
.
.
.
.
_
_
_
_
_
_
In other words
a
ij
=
_
_
_
1 if i = j + 1
1 if i = j 1
0 otherwise
Problem 5. Let P
j
(j = 0, 1, 2, . . .) be the Legendre polynomials
P
0
(x) = 1, P
1
(x) = x, P
2
(x) =
1
2
(3x
2
1), . . . .
Calculate the innite dimensional matrix A = (A
jk
)
A
jk
=
_
+1
1
P
j
(x)
dP
k
(x)
dx
dx
where j, k = 0, 1, . . .. Consider the matrix A as a linear operator in the
Hilbert space
2
(N
0
). Is |A| < ?
Problem 6. Let Z be the set of integers. Consider the Hilbert space
2
(Z
2
). Let (m
1
, m
2
) Z
2
. Let f(m
1
, m
2
) be an element of
2
(Z
2
). Con-
sider the unitary operators
Uf(m
1
, m
2
) := e
2im2
f(m
1
+1, m
2
), V f(m
1
, m
2
) := e
2im1
f(m
1
, m
2
+1) .
They are the so-called magnetic translation operators with phase and
respectively. Find the spectrum of U and V . Find the commutator [U, V ].
The so-called Harper operator which is self-adjoint is dened by
H := U +U
+V +V
.
Find the spectrum of

H. Consider the case , irrational and , rational.
Problem 7. The spectrum (

H) of a linear operator

H is dened as the
set of all for which the resolvent
R() = (I

H)
1
Hilbert Space
2
(N) 43
does not exist. If the linear operator

H is self-adjoint, the spectrum is a
subset of the real axis. The Lebesgue decomposition theorem states that
=
pp

ac

sing
where
pp
is the countable union of points (the pure point spectrum),
ac
is absolutely continuous with respect to Lebesgue measure and
sing
is
singular with respect to Lebesgue measure, i.e. it is supported on a set of
measure zero. Consider the Hilbert space
2
(Z) and the linear operator
H = I
2
I
2
3
I
2
I
2

where
1
is at position 0. Find the spectrum of this linear operator.
Chapter 5
Fourier Transform
2
(R). Find the Fourier trans-
form of the function
f(x) =
_
_
_
1 if 1 x 0
e
x
if x 0
0 otherwise
Problem 2. (i) Find the Fourier transform for
f
(x) =

2
exp([x[), > 0.
Discuss large and small.
(ii) Calculate
_

(x)dx.
Problem 3. Find the Fourier transform of the hat function
f(t) =
_
1 [t[ for 1 < t < 1
0 otherwise
Problem 4. Let f L
2
(R) and f L
1
(R. Assume that f(x) = f(x).
Can we conclude that

f(k) =

f(k)?
44
Fourier Transform 45
2
(R). Find the Fourier trans-
form of
f(x) = e
a|x|
, a > 0 .
2
(R). Let a > 0. Dene
f
a
(x) =
_
1
2a
[x[ < a
0 [x[ > a
Calculate _
R
f
a
(x)dx
and the Fourier transform of f
a
. Discuss the result in dependence of a.
2
(R). Let
() =
_
1 if 1/2 [[ 1
0 otherwise
and
() = e
||
, > 0 .
(i) Calculate the inverse Fourier transform of

() and

(), i.e.
(t) =
1
2
_
R
e
it

()
(t) =
1
2
_
R
e
it
()
(ii) Calculate the scalar product (t)[(t)) by utilizing the identity
2(t)[(t)) =
()[
()) .
2
(R) and the function f
L
2
(R)
f(x) =
_
1 if [x[ < 1
0 if [x[ 1
Calculate f f and verify the convolution theorem
f f =

f

f .
Problem 9. Let
f() =
_
(1
2
) for [[ 1
0 for [[ > 1
Find f(t).
Problem 10. Let a > 0. Find the Fourier transform of the function
f
a
: R R
f
a
(x) =
_
_
_
x/a
2
+ 1/a for a x 0
x/a
2
+ 1/a for 0 x a
0 otherwise
Problem 11. Let a > 0. Find the Fourier transform of
f
a
(t) =
1
a
e
a|t|
.
Discuss the cases a large and a small. Is f
a
L
2
(R).
Problem 12. Show that the Fourier transform of the rectangular window
of size N
w
n
=
_
1 for 0 n N 1
0 otherwise
is
W(e
i
) =
sin(N/2)
sin(/2)
e
i(N1)/2
.
2
(R). Let T > 0. Consider
the function in L
2
(R)
f(t) =
_
Acos(t) for T < t < T
0 otherwise
where A is a positive constant. Calculate the Fourier transform.
Problem 14. Let > 0. Show that the Fourier transform of the Gaussian
function
g
(x) =
1
2
exp
_
x
2
2
2
_
is again a Gaussian function
g
(k) = e
2
k
2
/2
.
We have
_
(x)dx = 1. Is
_

k
(k)dk = 1 ?
Fourier Transform 47
Problem 15. Show that the analytic function f : R R
f(x) = sech(x)
is an element of L
2
(R) and L
1
(R). Find the Fourier transform of the func-
tion.
Problem 16. Let a > 0. Find the Fourier transform of
2f
a
(x) +
sin(ax)
ax
where f
a
is the function with 1 for [x[ a and 0 otherwise.
Problem 17. Consider the Hermite-Gauss functions
f
n
(x) =
2
1/4
2
n
n!
H
n
(
2x) exp(x
2
), n = 0, 1, 2, . . .
where H
n
is the nth Hermite polynomial. They for an orthonormal basis
2
(R). Do the Fourier transform of the functions form
an orthonormal basis in the Hilbert space L
2
(R).
Chapter 6
Wavelets
2
[0, 1] and the function f(x) =
x
2
in this Hilbert space. Project the function f onto the subspace of L
2
[0, 1]
spanned by the functions (x), (x), (2x), (2x 1), where
(x) :=
_
1 for 0 x < 1
0 otherwise
(x) :=
_
_
_
1 for 0 x < 1/2
1 for 1/2 x < 1
0 otherwise
.
This is related to the Haar wavelet expansion of f. The function is called
the father wavelet and is called the mother wavelet.
Problem 2. Consider the function H L
2
(R)
H(x) =
_
_
_
1 0 x 1/2
1 1/2 x 1
0 otherwise
Let
H
mn
(x) := 2
m/2
H(2
m
x n)
where m, n Z. Draw a picture of H
11
, H
21
, H
12
, H
22
. Show that
H
mn
(x), H
kl
(x)) =
mk
nl
, k, l Z
where . ) denotes the scalar product in L
2
(R) Expand the function
f(x) = exp([x[)
48
Wavelets 49
with respect to H
mn
. The functions H
mn
form an orthonormal basis in
L
2
(R).
2
[0, 1] and the Haar scaling
function (father wavelet)
(x) =
_
1 if 0 x < 1
0 otherwise
Let n be a positive integer. We dene
g
k
(x) :=
n(nx k), k = 0, 1, . . . , n 1 .
(i) Show that the set of functions g
0
, g
1
, . . . , g
n1
is an orthonormal set
2
[0, 1].
(ii) Let f be a continuous function on the unit interval [0, 1]. Thus f
L
2
[0, 1]. Form the projection f
n
on the subspace S
n
of the Hilbert space
L
2
[0, 1] spanned by g
0
, g
1
, . . . g
n1
, i.e.
f
n
=
n1
k=0
f, g
k
)g
k
.
Show that f
n
(x) f(x) pointwise in x as n .
Problem 4. The continuous wavelet transform
Wf(a, b) =
1
a
_
+
f(t)
_
t b
a
_
dt, (a, b R, a > 0)
decomposes the function f L
2
(R) hierarchically in terms of elementary
components ((tb)/a). They are obtained from a single analyzing wavelet
applying dilations and translations. Here

denotes the complex conju-
gate of and a is the scale and b the shift parameter. The function has
to be chosen so that it is well localized both in physical and Fourier space.
The signal f(t) can be uniquely recovered by the inverse wavelet transform
f(t) =
1
C
_
+
_
+
0
Wf(a, b)
_
t b
a
_
da
a
db
if (t) (respectively its Fourier transform

() satises the admissibility
condition
C
=
_
+
0
[
()[
2
d < .
Consider the function
(t) = te
t
2
/2
.
Does satises the admissibility condition?
Problem 5. Consider the function H L
2
(R)
H(x) =
_
_
_
1 0 x < 1/2
1 1/2 x 1
0 otherwise
Let
H
mn
(x) := 2
m/2
H(2
m
x n)
where m, n Z. Draw a picture of H
11
, H
21
, H
12
, H
22
. Show that
H
mn
(x), H
kl
(x)) =
mk
nl
, k, l Z
where . ) denotes the scalar product in L
2
(R). Expand the function
f(x) = exp([x[)
with respect to H
mn
. The functions H
mn
form an orthonormal basis in
L
2
(R).
2
(R). Let L
2
(R) and
assume that satises
_
R
(t)(t k)dt =
0,k
i.e. the integral equals 1 for k = 0 and vanishes for k = 1, 2, . . .. Show that
for any xed integer j the functions
jk
(t) := 2
j/2
(2
j
t k), k = 0, 1, 2, . . .
form an orthonormal set.
Problem 7. Consider the function : R R
(x) :=
_
1 for x [0, 1]
0 otherwise
Find (x) := (2x) (2x 1). Calculate
_

(x)dx.
Problem 8. Consider the Littlewood-Paley orthonormal basis of wavelets.
The mother wavelet of this set is
L(x) :=
1
x
(sin(2x) sin(x)) .
Wavelets 51
Then
L
mn
(x) =
1
2
m/2
L(2
m
wx n), m, n Z
generates an orthonormal basis in the Hilbert space L
2
(R). Apply the rule
of LHospital to nd L(0).
Chapter 7
Operators
Problem 1. Show that an isometric operator need not be a unitary
operator.
2
[0, 1]. Show that the linear
operator T : L
2
[0, 1] L
2
[0, 1] dened by
Tf(x) = xf(x)
is a bounded self-adjoint linear operator without eigenvalues.
Problem 3. Show that if two bounded self-adjoint linear operators S and
T on a Hilbert space H are positive semi-denite and commute (ST = TS),
then their product ST is positive semi-denite. We have to show that
STf, f) 0 for all f H.
Problem 4. Let a > 0. Consider the Hilbert space L
2
[a, a]. Consider
the Hamilton operator
H =

2
2m
d
2
dx
2
+V (x)
where
V (x) =
_
0 for [x[ a
otherwise
52
Operators 53
Solve the Schrodinger equation, where the initial function (t = 0) = (x)
is given by
(x) =
_
x/a
2
+ 1/a for a x 0
x/a
2
+ 1/a for 0 x a
Normalize . Calculate the probability to nd the particle in the state
(x) =
1
a
sin
_
x
a
_
after time t. A basis in the Hilbert space L
2
[a, a] is given by
_
1
a
sin
_
nx
a
_
,
1
a
cos
_
(n 1/2)x
a
_
n = 1, 2, . . .
_
.
Problem 5. Show that in one dimensional problems the energy spectrum
of the bound state is always non-degenerate. Hint. Suppose that the oppsite
is true. Let u
1
and u
2
be two linearly independent eigenfunctions with the
same energy eigenvalues E, i.e.
d
2
u
1
dx
2
+
2m
2
(E V )u
1
= 0,
d
2
u
2
dx
2
+
2m
2
(E V )u
2
= 0 .
Problem 6. A particle is enclosed in a rectangular box with imprenetra-
ble walls, inside which it can move freely. The Hilbert space is L
2
([0, a]
[0, b] [0, c]). Find the eigenfunctions and eigenvalues. What can be said
about the degeneracy, if any, of the eigenfunctions.
Problem 7. Conside the Hilbert space L
2
[0, 1] and the linear operator
T : L
2
[0, 1] L
2
[0, 1] dened by
(Tf)(x) := xf(x) .
Show that T is self-adjoint and positive denite. Find its positive square
root.
2
(N) and the linear operator T
dened by
T : (x
1
, x
2
, x
3
, . . .) (0, 0, x
3
, x
4
, . . .) .
Is T bounded? Is T self-adjoint? If so is T positive?
Problem 9. In classical mechanics we have
L = r p, T = r F
where T is the torque, F = V (V potential depending only on r) and
dL
dt
= T.
In quantum mechanics with p i, r r and wave function we
have
L = i
_
R
3
d
3
x
(r )
and
T =
_
R
3
d
3
x
(r V )
since F = V . and
obey the Schrodinger equation

i
t
=

2
2m
2
+V
i
t
=

2
2m
+V
.
Show that
dL
dt
= T.
Problem 10. Let

H be a bounded self-adjoint Hamilton operator with
normalized eigenfunctions
j
(j I) which form an orthonormal basis in
the underlying Hilbert space. We can write
(t) =
jI
c
j
e
iEjt/
j
where E
j
are the eigenvalues of

H. Find P(t) = (t = 0)[(t)).
Problem 11. Consider the Hamilton operator
H =

2
2m
d
2
dx
2
+D(1 e
x
)
2
+eExcos(t)
where > 0. Find the quantum Liouville equation for this Hamilton
operator.
Problem 12. Consider the Schrodinger equation
i
t
=
_
1
2m
+V (x)
_
Operators 55
Find the coupled system of partial dierential equations for
:=
, v :=
_
_
.
2
(R). Let f L
2
(R) and
R. We dene the operator U() as
U()f(x) := e
i/2
f(xe
i
) .
Is the operator U() unitary?
2
(R). Let k Z. For k = 0
we dene s
0
= 0, for k 1 we dene
s
k
:= 1 +
1
2
+
1
3
+ +
1
k
and for k < 0 we dene s
k
= s
k
. Let > 0. Dene the indicator
functions W
k
as
W
k
(x) :=
_
1 for s
k
< x/ s
k+1
0 otherwise
Let u L
2
(R). Dene the linear operator O as
(Ou)(x) := g(x)u(x)
where
g(x) =
x
kZ
_
s
k
+s
k+1
2
_
W
k
(x) .
(i) Show that O is a bounded self-adjoint operator for any > 0.
(ii) Show that the norm of O
|O| = sup
u=1
|Ou|
is given by 1/2.
Chapter 8
Generalized Functions
Problem 1. Consider the function H : R R
H(x) :=
_
_
_
1 0 x 1/2
1 1/2 x 1
0 otherwise
Find the derivative of H in the sense of generalized functions. Obviously
H can be considered as a regular functional
_
R
H(x)(x)dx.
Find the Fourier transform of H. Draw a picture of the Fourier transform.
Problem 2. Let C
m
[a, b] be the vector space of m-times dierentiable
functions and the m-th derivative is continuous over the interval [a, b] (b >
a). We dene an inner product (scalar product) of such two functions f
and g as
f, g)
m
:=
_
b
a
_
fg +
df
dx
dg
dx
+ +
d
m
f
dx
m
d
m
g
dx
m
_
dx.
Given (Legendre polynomials)
f(x) =
1
2
(3x
2
1), g(x) =
1
2
(5x
3
3x)
and the interval [1, 1], i.e. a = 1 and b = 1. Show that f and g are
orthogonal with respect to the inner product f, g)
0
. Are they orthogonal
with respect to f, g)
1
?
56
Generalized Functions 57
Problem 3. Let P be the parity operator, i.e.
Pr := r .
Obviously, P = P
1
. We dene
O
P
u(r) := u(P
1
r) u(r) .
The vector r can be expressed in spherical coordinates as
r = r(sin cos , sin sin, cos )
where
0 < 2 ; 0 < .
(i) Calculate P(r, , ).
(ii) Let
Y
lm
(, ) =
(1)
l+m
2
l
l!
_
2l + 1
4
(l m)!
(l +m)!
_
1/2
(sin)
m
d
l+m
d(cos )
l+m
(sin)
2l
e
im
be the spherical harmonics. Find
O
P
Y
lm
.
Problem 4. In the Hilbert space H =
2
(N
0
) Bose annihilation and
creation operators denoted by b and b
are dened as follows: They have a

common domain
T(b) = T(b
) =
_
_
_
= (x
0
, x
1
, x
2
, . . .)
T
:
j=0
j[x
j
[
2
<
_
_
_
Then b is given by
b(x
0
, x
1
, x
2
, . . .)
T
= (x
1
,
2x
2
,
3x
3
, . . .)
T
and b
is given by
b
(x
0
, x
1
, x
2
, . . .) = (0, x
0
,
2x
1
,
3x
2
, . . .) .
The innite dimensional vectors
u
n
= (0, 0, . . . , 0, 1, 0, . . .)
T
where the 1 is at the n position (n = 0, 1, 2, . . .) form the standard basis in
H =
2
(N
0
). Is
= (1, 1/2, 1/3, . . . , 1/n, . . .)
an element of T(a)?
Problem 5. Given a function (signal) f(t) = f(t
1
, t
2
, . . . , t
n
) L
2
(R
n
)
of n real variables t = (t
1
, t
2
, . . . , t
n
). We dene the symplectic tomogram
associated with the square integrable function f
w(X, , ) =
n
k=1
1
2[
k
[
_
R
n
dt
1
dt
2
dt
n
f(t) exp
_
_
n
j=1
_
i
j
2
j
t
2
j

iX
j
j
t
j
_
_
_
2
where (
j
,= 0 for j = 1, 2, . . . , n)
X = (X
1
, X
2
, . . . , X
n
), = (
1
,
2
, . . . ,
n
), = (
1
,
2
, . . . ,
n
) .
(i) Prove the equality
_
R
n
w(X, , )dX =
_
R
n
[f(t)[
2
dt (1)
for the special case n = 1. The tomogram is the probabilty distribution
function of the random variable X. This probability distribution function
depends on 2n extra real parameters and .
(ii) The map of the function f(t) onto the tomogram w(X, , ) is invert-
ible. The square integrable function f(t) can be associated to the density
matrix
f
(t, t
) = f(t)f(t
) .
This density matrix can be mapped onto the Ville-Wigner function
W(q, p) =
_
R
n
f
_
q +
u
2
, q
u
2
_
e
ipu
du.
Show that this map is invertible.
(iii) How is the tomogram w(X, , ) related to the Ville-Wigner function?
(iv) Show that the Ville-Wigner function can be reconstructed from the
function w(X, , ).
(v) Show that the density matrix f(t)f
(t
) can be found from w(X, , ).

Problem 6. Starting with the set of polynomials 1, x, x
2
, . . . , x
n
, . . .
use the Gram-Schmidt procedure the scalar product (inner product)
f, g) =
_
1
1
(f(x)g(x) +f
(x)g
(x))dx
to nd the rst ve orthogonal polynomials, where f
denotes derivative.
Problem 7. Describe the one-dimensional scattering of a particle incident
on a Dirac delta function, i.e.
U(q) = U
0
(q)
where u
0
> 0. Find the transmission and reection coecient.
Problem 8. (i) Give the denition of the current density, transmission
coecient, and reection coecient.
(ii) Calculate the transmission and the reection coecients of a particle
having total energy E, at the potential barrier given by
V (x) = a(x), a > 0
1
2
k=
e
ikx
=
k=
(x 2k)
in the sense of generalized functions
Hint. Expand the 2 periodic function
f(x) =
1
2

x
2
into a Fourier series.
Problem 10. (i) Give the denition of a generalized function.
(ii) Calculate the rst and second derivative in the sense of generalized
function of
f(x) =
_
_
_
0 x < 0
4x(1 x) 0 x 1
0 x > 1
(iii) Calculate the Fourier transform of f(x) = 1 in the sense of generalized
functions.
Problem 11. Consider the generalized function
f(x) = [ cos(x)[ .
Find the derivative in the sense of generalized functions.
Problem 12. Consider the generalized function
f(x) :=
_
cos(x) for x [0, 2)
0 otherwise
Find the rst and second derivative of f in the sense of generalized func-
tions.
Problem 13. Find the derivative of f : R R
f(x) = [x[
in the sense of generalized functions.
Problem 14. Find the rst three derivatives of the function f : R R
dened by
f(x) = e
|x|
in the sense of generalized functions.
Problem 15. The Sobolev space of order m, denoted by H
m
(), is dened
to be the space consisting of those functions in the Hilbert space L
2
() that,
together with all their weak partial derivatives up to and including those
of order m, belong to the Hilbert space L
2
(), i.e.
H
m
() := u : D
u L
2
() for all such that [[ m .
We consider real-valued functions only, and make H
m
() an inner product
space by introducing the Sobolev inner product , )
H
m dened by
u, v)
H
m :=
_
||m
(D
u)(D
v)dx for u, v H
m
() .
This inner product generates the Sobolev norm | |
H
m dened by
|u|
2
H
m = u[u)
H
m =
_
||m
(D
u)
2
dx.
Thus H
0
() = L
2
(). We can write
u, v) =
||m
D
, D
v)
L2()
In other words the Sobolev inner product u, v)
H
m
()
is equal to the sum
of the L
2
() inner products of D
u and D
v over all such that [[ m.

(i) Consider the domain = (0, 2) and the function
u(x) =
_
x
2
0 < x 1
2x
2
2x + 1 1 < x < 2 .
Obviously u L
2
(). Find the Sobolev space to which u belongs.
(ii) Find the norm of u.
Problem 16. Let c > 0. Consider the Schrodinger equation

2
2m
d
2
dx
2
+c
(n)
(x) = E
where
(n)
(n = 0, 1, 2, . . .) denotes the n-th derivative of the delta function.
Derive the joining conditions on the wave function .
Problem 17. The Morlet wavelet consists of a plane wave modulated by
a Gaussian, i.e.
() =
1
1/4
e
i
e
2
/2
where is the dimensionless frequency. Show that if = 6 the admissibility
condition is satised.
Problem 18. Let
f
0
(x) = exp(x
2
/2) .
We dene the mother wavelets f
n
as
f
n
(x) =
d
dx
f
n1
(x), n = 1, 2, . . .
Show that the family of f
n
s obey the Hermite recursion relation
f
n
(x) = xf
n1
(x) (n 1)f
n2
(x), n = 2, 3, . . . .
Problem 19. Show that the 2-dimensional complex -function can be
written as ( C)
(2)
(z) =
1
2
_
C
d
2
exp(
z z
) =
1
2
_
C
d
2
exp(i(
z +z
)) .
(x x
) =
1
_
1 + 2
k=1
cos(kx) cos(kx
)
_
.
Problem 21. Let a > 0. Show that
m=
exp(i2m(x +q)/a) a
k=
(x +q ka) .
Miscellaneous
Problem 22. Let
i
t
=

H
be the Schrodinger equation, where
H =

2
2m
+U(r), :=

2
x
2
1
+

2
x
2
2
+

2
x
2
3
and r = (x
1
, x
2
, x
3
). Let
(r, t) :=

(r, t)(r, t)
Find j such that
divj +

t
= 0 .
Problem 23. A particle is enclosed in a rectangular box with impene-
trable walls, inside which in can move freely. The Hilbert space is
L
2
([0, a] [0, b] [0, c])
where a, b, c > 0. Find the eigenfunctions and eigenvalues. What can be
said about the degeneracy, if any, of the eigenfunctions?
Problem 24. Show that in one-dimensional problems the energy spec-
trum of the bound states is always non-degenerate. Hint. Suppose that
the opposite is true. Let u
1
, u
2
be two linearly independent eigenfunctions
with the same energy eigenvalue E, i.e.
d
2
u
1
dx
2
+
2m
2
(E V )u
1
= 0
d
2
u
2
dx
2
+
2m
2
(E V )u
2
= 0 .
Problem 25. Derive the Heisenberg uncertainty relation.
Problem 26. Give the standard postulates in quantum mechanics and
discuss the problematic.
Problem 27. Show that in one-dimensional problems the energy spec-
trum of the bound states is always non-degenerate.
Hint. Suppose that the opposite is true.
Let u
1
and u
2
be two linearly independent eigenfunctions with the same
energy eigenvalues E.
d
2
u
1
dx
2
+
2m
2
(E V )u
1
= 0
d
2
u
2
dx
2
+
2m
2
(E V )u
2
= 0
Problem 28. Let a > 0 and let f
a
: R R be given by
f
a
(x) =
_
x/a
2
+ 1/a for a x 0
x/a
2
+ 1/a for 0 x a
The function f
a
generates regular functional. Find the derivative of f
a
in
the sense of generalized functions.
Problem 29. Consider a one-dimensional lattice (chain) with lattice
constant a. Let k be the sum over the rst Brioullin zone we have
1
N
k1.BZ
F((k))
a
2
_
/a
/a
F((k))dk = G
where
(k) =
0
2
1
cos(ka) .
Using the identity
_

(E (k))F(E)dE F((k))
we can write
G =
a
2
_

F(E)
_
_
/a
/a
(E (k))dk
_
dE .
Calculate
g(E) =
_
/a
/a
(E (k))dk
where g(E) is called the density of states.
Problem 30. Let > 0. Consider the Schrodinger eigenvalue equation
_
d
2
dx
2
+ 2(x)
_
u(x, ) = E()u(x, )
with the boundary conditions u(1, ) = 0. Here is the coupling con-
stant and determines the penetrability of the potential barrier. Find the
eigenfunctions and the eigenvalues.
Problem 31. Show that in the sense of generalized functions
(x) =
1
2
lim
0
1
e
|x|/
(x) =
1
lim
sin
2
(x)
(x)
2
(x) =
1
4
lim
0
1
_
1 +
[x[
_
e
|x|/
.
Problem 32. Give two interpretations of the series of derivatives of
functions
f(k) = 2
n=0
c
n
(1)
n
(n)
(k) . (1)
H(x a) =
_

du
_

d
2
exp(iu( x)) .
z
_
1
z
_
= (z)
where
z

1
2
_

x
+i

y
_
.
Problem 35. Let a > 0. Show that
(x
2
a
2
) =
1
2a
((x a) +(x +a)) .
Problem 36. (i) Show that the Fourier transform in the sense of gener-
alized function of the Dirac comb
nZ
(x n)
is again a Dirac comb.
(ii) Find the Fourier transform in the sense of generalized functions of
1 +
2(x) .
Problem 37. (i) Consider the nonlinear dierential equation
3u
du
dx
= 2
du
dx
d
2
u
dx
2
+u
d
3
u
dx
3
.
Show that u(x) = e
|x|
is a solution in the sense of generalized function.
(ii) Consider the nonlinear partial dierential equation
u
t

3
u
x
2
t
+ 3u
u
x
= 2
u
x
2
u
x
2
+u
3
u
x
3
.
Show that u(x, t) = c exp([x ct[) (peakon) is a solution in the sense of
generalized functions.
Problem 38. Let f be a dierentiable function with a simple zero at x =
a such that f(x = a) = 0 and df(x = a)/dx ,= 0. Let g be a dierentiable
function with a simple zero at x = b ,= a such that g(x = b) = 0 and
dg(x = b)/dx ,= 0. Show that
(f(x)g(x)) =
1
[f
(a)g(a)[
(x a) +
1
[f(b)g
(b)[
(x b)
where

denotes dierentiation.
Problem 39. Consider the non-relativistic hydrogen atom, where a
0
is
the Bohr radius and a = a
0
/Z. The Schrodinger-Coulomb Green function
G(r
1
, r
2
; E) corresponding to the energy variable E is the solution of the
partial dierential equation
_

2
2m
2
1

2
amr
1
E
_
G(r
1
, r
2
; E) = (r
1
r
2
)
with the appropiate boundary conditions. Show that expanding G in terms
of spherical harmonics Y
m
G(r
1
, r
2
; E) =
=0
m=
g
(r
1
, r
2
; E)Y
m
(
1
,
1
)Y
m
(
2
,
2
)
we nd for the radial part g
of the Schrodinger-Coulomb Green function

_
1
r
2
1
d
dr
1
_
r
2
1
d
dr
1
_
( + 1)
r
2
1
+
2
ar
1
2
a
2
_
g
(r
1
, r
2
; ) =
2m
2
(r
1
r
2
)
r
1
r
2
where
2
a
2
:=
2
/(2mE).
Hint. Utilize the identity
(r
1
r
2
) =
(r
1
r
2
)
r
1
r
2
=0
m=
Y
m
(
1
,
1
)Y
m
(
2
,
2
)
Bibliography
Fuhrmann, P. A.
A Polynomial Approach to Linear Algebra
Springer-Verlag, New York (1996)
Golub, G. H. and Van Loan C. F.
Matrix Computations, Third Edition,
Johns Hopkins University Press (1996)
Grossman S. I.
Elementary Linear Algebra, Third Edition
Wadsworth Publishing, Belmont (1987)
Horn R. A. and Johnson C. R.
Topics in Matrix Analysis
Cambridge University Press (1999)
Johnson D. L.
Presentation of Groups
Cambridge University Press (1976)
Lang S.
Linear Algebra
Addison-Wesley, Reading (1968)
Miller W.
Symmetry Groups and Their Applications
Academic Press, New York (1972)
68
Bibliography 69
Schneider H. and Barker G. P.
Matrices and Linear Algebra,
Dover Publications, New York (1989)
Steeb W.-H.
Matrix Calculus and Kronecker Product with Applications and C++ Pro-
grams
World Scientic Publishing, Singapore (1997)
Steeb W.-H.
Continuous Symmetries, Lie Algebras, Dierential Equations and Com-
puter Algebra
Steeb W.-H.
Hilbert Spaces, Wavelets, Generalized Functions and Quantum Mechanics
Kluwer Academic Publishers, Dordrecht (1998)
Steeb W.-H.
Problems and Solutions in Theoretical and Mathematical Physics,
Third Edition, Volume I: Introductory Level
Steeb W.-H.
Problems and Solutions in Theoretical and Mathematical Physics,
Third Edition, Volume II: Advanced Level
Steeb W.-H., Hardy Y., Hardy A. and Stoop R.
Problems and Solutions in Scientic Computing with C++ and Java Sim-
ulations
Yosida K.
Functional Analysis, Fifth Edition
Springer Verlag (1978)
Wybourne B. G.
Classical Groups for Physicists
John Wiley, New York (1974)
Index
Admissibility condition, 49
Analysing wavelet, 49
Antilinear operator, 2
Bell basis, 8, 12
Bernstein polynomials, 33, 37
Cauchy sequence, 9
Chebyshev polynomials, 29
Comb, 2
Continuous wavelet transform, 49
Convolution theorem, 45
Dilations, 49
Dirac comb, 64
Fock space, 24, 36
Frame, 2
Fredholm alternative, 3
Gaussian measure, 36
Gram-Schmidt orthogonalization pro-
cess, 8
Gram-Schmidt technique, 15
Haar scaling function, 49
Hardy space, 5
Harper operator, 42
Hat function, 44
Inverse wavelet transform, 49
Kernel, 2
Legendre polynomials, 20, 26
Legrendre polynomials, 29
Lie group, 30
Morlet wavelet, 61
Null space, 2
Paley-Wiener space, 38
Peakon, 65
projection theorem, 4
Range, 3
Reproducing kernel, 36
Reproducing kernel Hilbert space, 38
Resolvent, 42
Schauder basis, 35
Shifted Legendre polynomials, 23
Sinc function, 38
Sobolev space, 60
Spectral theorem, 10
Spherical harmonics, 57
Symplectic tomogram, 58
Translations, 49
Triangle function, 35
Volterra integral equation, 23
70

Hilbert Space Problems

Caricato da

Informazioni sul documento

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Hilbert Space Problems

Caricato da

Copyright:

Formati disponibili

Problems and Solutions

y scalar product (inner product) in C

conjugate transpose of matrix A

conjugate transpose of matrix A

x = 0 have only the zero solution, and the linear

x = y have a unique solution x H for every y H;

x = 0 have nontrivial, nite-dimensional solution spaces

:= k H : k[u) = 0 for all u ( .

. Consider the Hilbert space R. Show

is a reproducing kernel Hilbert space. The Fourier transform of f L

y and the domain of

obey the Schrodinger equation

are dened as follows: They have a

) can be found from w(X, , ).

v over all such that [[ m.

of the Schrodinger-Coulomb Green function

Potrebbero piacerti anche