Functional Problems AnhLe Full

Functional Analysis Problems with Solutions
ANH QUANG LE, Ph.D.

September 14, 2013
Contents
Contents 1
1 Normed and Inner Product Spaces 3
2 Banach Spaces 15
3 Hilbert Spaces 27
3.1 Hilbert spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.2 Weak convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4 Linear Operators - Linear Functionals 45
4.1 Linear bounded operators . . . . . . . . . . . . . . . . . . . . . . . . 45
4.2 Linear Functionals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5 Fundamental Theorems 73
6 Linear Operators on Hilbert Spaces 87
7 Compact Operators 99
8 Bounded Operators on Banach Spaces and Their Spectra 113
9 Compact Operators and Their Spectra 131
10 Bounded Self Adjoint Operators and Their Spectra 143
1
www.MATHVN.com - Anh Quang Le, Ph.D
www.MATHVN.com
2 CONTENTS
Notations:
B(X, Y ): the space of all bounded (continuous) linear operators from X to Y .
Image (T) Ran(T): the image of a mapping T : X Y .
x
n
w
x: x
n
converges weakly to x.
X
: the space of all bounded (continuous) linear functionals on X.

F or K: the scalar eld, which is R or C.
Re, Im: the real and imaginary parts of a complex number.
www.MATHVN.com
Chapter 1
Normed and Inner Product Spaces
Problem 1.
Prove that any ball in a normed space X is convex.
Solution.
Let B(x
0
; r) be any ball of radius r > 0 centered at x
0
X, and x, y B(x
0
; r).
Then
|x x
0
| < r and |y x
0
| < r.
For every a [0, 1] we have
|ax + (1 a)y x
0
| = |(x x
0
)a + (1 a)(y x
0
)|
a|x x
0
| + (1 a)|y x
0
|
< ar + (1 a)r = r.
So ax + (1 a)y B(x
0
; r). .
Problem 2.
Consider the linear space C[0, 1] equipped with the norm
|f|
1
=
_
1
0
[f(x)[dx.
Prove that there is no inner product on C[0, 1] agreed with this norm.
3
www.MATHVN.com
4 CHAPTER 1. NORMED AND INNER PRODUCT SPACES
Solution.
We show that the norm |.|
1
does not satisfy the parallelogram law. Let
f(x) = 1 and g(x) = 2x.
Then
|f|
1
=
_
1
0
1.dx = 1, |g|
1
=
_
1
0
[2x[dx = 1,
while
|f g|
1
=
_
1
0
[1 2x[dx =
1
2
, |f +g|
1
=
_
1
0
[1 + 2x[dx = 2.
Thus,
|f g|
2
1
+|f +g|
2
1
=
17
4
,= 2(|f|
2
1
+|g|
2
1
) = 4.
Problem 3.
Consider the linear space C[0, 1] equipped with the norm
|f| = max
t[0,1]
[f(t)[.
Prove that there is no inner product on C[0, 1] agreed with this norm.
Solution.
We show that the parallelogram law with respect to the given norm does not hold
for two elements in C[0, 1].
Let f(t) = t, g(t) = 1 t, t [0, 1]. Then f, g C[0, 1] and
|f| = max
t[0,1]
t = 1, |g| = max
t[0,1]
(1 t) = 1,
and
|f +g| = max
t[0,1]
1 = 1, and |f g| = max
t[0,1]
[ 1 + 2t[ = 1.
Thus,
|f g|
2
1
+|f +g|
2
1
= 2 ,= 2(|f|
2
1
+|g|
2
1
) = 4.
Problem 4.
Prove that:
If the unit sphere of a normed space X contains a line segment [x, y] where x, y
X and x ,= y , then x and y are linearly independent and |x +y| = |x| +|y| .
www.MATHVN.com
5
Solution.
Suppose that the unit sphere contains a line segment [x, y] where x, y X and
x ,= y. Then
|ax + (1 a)y| = 1 for any a [0, 1].
Choose a = 1/2 then we get |
1
2
(x + y)| = 1, that is |x + y| = 2. Since x and y
belong to the unit sphere, we have |x| = |y| = 1. Hence
|x +y| = |x| +|y|.
Let us show that x, y are linearly independent. Assume y = x for some C. We
have
1 = |ax + (1 a)x| = [a + (1 a)[.
For a = 0 we get [[ = 1 and for a = 1/2 we get [1 + [ = 2. These imply that
= 1, and so x = y, which is a contradiction.
Problem 5.
Prove that two any norms in a nite dimensional space X are equivalent.
Solution.
Since equivalence of norms is an equivalence relation, it suces to show that an
arbitrary norm |.| on X is equivalent to the Euclidian norm |.|
2
. Let e
1
, ..., e
n
be
a basis for X. Every x X can be written uniquely as x =
n
k=1
c
k
e
k
. Therefore,
|x|
n
k=1
[e
k
[|e
k
|
_
n
k=1
[c
k
[
2
_
1/2
_
n
k=1
|e
k
|
2
_
1/2
A|x|
2
,
where A = (
n
k=1
[e
k
[
2
)
1/2
is a non-zero constant. This shows that the map x |x|
is continuous w.r.t. the Euclidian norm. Now consider S = x : |x|
2
= 1. This is
just the unit sphere in (X, |.|
2
), which is compact. The map
S R dened by x |x|
is continuous, so it attains a minimum m and a maximum M on S. Note that m > 0
because S ,= . Thus, for all x S, we have
m |x| M.
Now, for x X, x ,= 0,
x
|x|
2
S, so
m
|x|
|x|
2
M.
www.MATHVN.com
That is
m|x|
2
|x| M|x|
2
.
Hence, the two norms are equivalent.
Problem 6.
Let X be a normed space.
(a) Find all subspaces of X which are contained in some ball B(a; r) of X.
(b) Find all subspaces of X which contain some ball B(x
0
; ) of X.
Solution.
(a) Let Y be a subspace of X which is contained in some ball B(a; r) of X. Note
rst that the ball B(a; r) must contain the vector zero of X (and so of Y ); otherwise,
the question is impossible. For any number A > 0 and any x Y , we have Ax Y
since Y is a linear space. By hypothesis Y B(a; r), so we have Ax B(a; r). This
implies that |Ax| < r +|a|. Finally
|x| <
r +|a|
A
.
A > 0 being arbitrary, it follows that |x| = 0, so x = 0. Thus, there is only one
subspace of X, namely, Y = 0, which is contained in some ball B(a; r) of X.
(b) Let Z be a subspace of X which contain some ball B(x
0
; ) of X. Take any
x B(0; ). Then x + x
0
B(x
0
; ) and so x + x
0
Z since Z B(x
0
; ). Now,
since x
0
Z, x + x
0
Z and Z is a linear space, we must have x Z. Hence
B(0; ) Z.
Now for any nonzero x X, we have
x
2|x|
B(0; ) Z. Hence x Z. We can
conclude that Z = X. In other words, the only subspace of X which contains some
ball B(x
0
; ) of X is X itself.
Problem 7.
Prove that any nite dimensional normed space :
(a) is complete (a Banach space),
(b) is reexive.
Solution.
Let X be a nite dimensional normed space. Suppose dimX = d.
www.MATHVN.com
7
(a) By Problem 5, it suces to consider the Euclidian norm in X. Let e
1
, ..., e
d
be a basis for X. For x X there exist numbers c

1
, ..., c
d
such that
x =
d
k=1
c
k
e
k
and |x| =
_
d
k=1
[c
k
[
2
_
1/2
.
Let (x
(n)
) be a Cauchy sequence in X. If for each n, x
(n)
=
d
k=1
a
(n)
k
e
k
then
|x
(n)
x
(m)
| =
_
d
k=1
[a
(n)
k
a
(m)
k
[
2
_
1/2
0 as n, m .
Hence, for every k = 1, ..., d,
[a
(n)
k
a
(m)
k
[ 0 as n, m .
Therefore, each sequence of numbers (a
(n)
k
) is a Cauchy sequence, so
a
(n)
k
a
(0)
k
as n for every k = 1, 2, ..., d.
Let a =
d
k=1
a
(0)
k
e
k
then x
(n)
a X.
(b) Let f X
where X
is the space of all linear functionals on X. We have

f(x) = f
_
d
k=1
c
k
e
k
_
=
d
k=1
c
k
f(e
k
) =
d
k=1
c
k
k
,
where
k
= f(e
k
). Let us dene f
k
X
by the relation f
k
(x) = c
k
, k = 1, ..., d.
For any x X and f X
, we get
f(x) =
d
k=1
f
k
(x)
k
, i.e., f =
d
k=1
k
f
k
.
Hence, dimX
d.
Let
d
k=1
k
f
k
= 0. Then, for any x X,
d
k=1
k
f
k
(x) = 0, and by taking
x =
d
k=1

k
e
k
, we obtain f
k
(x) =
k
, and
d
k=1
k
f
k
(x) =
d
k=1
[
k
[
2
= 0.
Hence,
k
= 0 for all k = 1, ..., d and thus, dimX
= d. For the space X
we have
X
, so dimX
= n d and dim(X
= n. From the relation X (X

(X
we conclude that d n. Thus, n = d, and so X = (X
.
www.MATHVN.com
Problem 8. ( Reed-Simon II.4)
(a) Prove that the inner product in a normed space X can be recovered from the
polarization identity:
x, y) =
1
4
_
(|x +y|
2
|x y|
2
) i(|x +iy|
2
|x iy|
2
)
_
.
(b) Prove that a normed space is an inner product space if and only if the norm
satises the parallelogram law:
|x +y|
2
+|x y|
2
= 2(|x|
2
+|y|
2
).
Solution.
(a) For the real eld case, the polarization identity is
x, y) =
1
4
(|x +y|
2
|x y|
2
). ()
We use the symmetry of the inner product and compute the right hand side of ():
1
4
(|x +y|
2
|x y|
2
) =
1
4
_
x +y, x +y) x y, x y)
=
1
2
_
x, y) +y, x)
= x, y).
For the complex eld case, we again expand the right hand side, using the relation
we just established:
1
4
_
(|x +y|
2
|x y|
2
) i(|x +iy|
2
|x iy|
2
)
_
=
1
2
_
x, y) +y, x)
i
2
_
x, iy) +iy, x)
=
1
2
x, y) +
1
2
y, x)
i
2
2
x, y) +
i
2
2
y, x)
= x, y).
(b) If the norm comes from an inner product, then we have
|x +y|
2
+|x y|
2
= x +y, x +y) +x y, x y)
= 2x, x) + 2y, y) +x, y) +y, x) x, y) y, x)
= 2(|x|
2
+|y|
2
).
www.MATHVN.com
9
Now suppose that the norm satises the parallelogram law. Assume the eld is
C, and dene the inner product via the polarization identity from part (a). If
x, y.z X, we write
x +y = x +
y +z
2
+
y z
2
, x +z = x +
y +z
2

y z
2
,
and we have
x, y) +x, z) =
1
4
_
|x +y|
2
+|x y|
2
|x y|
2
|x z|
2
_
+
i
4
_
|x +iy|
2
+|x +iz|
2
|x iy|
2
|x iz|
2
_
=
1
2
_
_
_
_
_
x +
y +z
2
_
_
_
_
2
+
_
_
_
_
y +z
2
_
_
_
_
2
_
_
_
_
x
y +z
2
_
_
_
_
2
_
_
_
_
y z
2
_
_
_
_
2
_
i
2
_
_
_
_
_
x +i
y +z
2
_
_
_
_
2
+
_
_
_
_
i
y +z
2
_
_
_
_
2
_
_
_
_
x i
y +z
2
_
_
_
_
2
_
_
_
_
i
y z
2
_
_
_
_
2
_
=
1
2
_
_
_
_
_
x +
y +z
2
_
_
_
_
2
+
_
_
_
_
y +z
2
_
_
_
_
2
_
_
_
_
y z
2
_
_
_
_
2
_
_
_
_
x
y +z
2
_
_
_
_
2
_
i
2
_
_
_
_
_
x +i
y +z
2
_
_
_
_
2
+
_
_
_
_
i
y +z
2
_
_
_
_
2
_
_
_
_
i
y z
2
_
_
_
_
2
_
_
_
_
x i
y +z
2
_
_
_
_
2
_
=
1
4
_
|x +y +z|
2
+|x|
2
|x (y +z)|
2
|x|
2
_
i
4
_
|x +i(y +z)|
2
+|x|
2
|x i(y +z)|
2
|x|
2
_
= x, y +z).
This holds for all x, y, z X, so, in particular,
x, ny) = nx, y) for n N.
And it also satises
x, ry) = rx, y) for r Q.
Moreover, again by the polarization identity, we have
x, iy) =
1
4
_
|x +iy|
2
|x iy|
2
_
i
4
_
|x y|
2
|x +y|
2
_
= ix, y).
Combining these results we have
x, y) = x, y) for Q+iQ.
www.MATHVN.com
Now, if C, by the density of Q+iQ in C, there exists a sequence (
n
) in Q+iQ
converging to . It follows that
x, y) = x, y) for C.
Thus the ., .) is linear.
Since |i(x iy)| = |x iy|, we have
y, x) = x, y),
and
x, x) =
1
4
(|2x|
2
)
i
4
_
[1 +i[|x|
2
[1 i[
2
|x|
2
_
= |x|
2
.
So this shows that the norm is induced by ., .) and that it is also positive denite,
and thus it is an inner product.
Problem 9. (Least square approximation. Reed-Simon II.5)
Let X be an inner product space and let x
1
, ..., x
N
be an orthonormal set.
Prove that
_
_
_
_
_
x
N
n=1
c
n
x
n
_
_
_
_
_
is minimized by choosing c
n
= x
n
, x).
Solution.
For every x X, we write
x =
N
n=1
x
n
, x)x
n
+z, for some z X. ()
We observe that for all n = 1, ..., N,
x
n
, z) = x
n
, x)
N
k=1
x
n
, x)x
n
, x
k
)
= x
n
, x) x
n
, x) = 0.
Therefore zx
n
. Then due to () we can write
x
N
n=1
c
n
x
n
=
N
n=1
_
x
n
, x) c
n
_
x
n
. .
z
N
+z.
www.MATHVN.com
11
Since zz
N
, we have
_
_
_
_
_
x
N
n=1
c
n
x
n
_
_
_
_
_
2
= |z
N
|
2
+|z|
2
=
N
n=1
x
n
, x) c
n
2
+|z|
2
,
which attains its minimum if
c
n
= x
n
, x) for all n = 1, ..., N.
Review: Quotient normed space.
Let X be a vector space, and let M be a subspace of X. We dene an equivalence relation on
X by
x y if and only if x y M.
For x X, let [x] = x + M denote the equivalence class of x and X/M the set of all equivalence
classes. On X/M we dene operations:
[x] + [y] = [x +y]
[x] = [x], C.
Then X/M is a vector space.
If the subspace M is closed, then we can dene a norm on X/M by
|[x]| = inf
y[x]
|y| = inf
mM
|x +m| = inf
mM
|x z| = d(x, M).
What a ball in X/M looks like?
B([x
0
]; r) := [x] : |[x] [x
0
]| < r = x +M : |x x
0
+M| < r.
Suppose that M is closed in X. The canonical map (the natural projection) is dened by
: X X/M, (x) = [x] = x +M.
It can be shown that |(x)| |x|, x X, so is continuous.
Problem 10.
Let X be a normed space and M a closed subspace of X. Let : X X/M
be the canonical map. Show that the topology induced by the standard norm on
X/M is the usual quotient topology, i.e. that O X/M is open in X/M if and
only if
1
(O) is open in X.
www.MATHVN.com
Solution.
If O is open in X/M, then
1
(O) is open in X since is continuous.
Now suppose that O X/M and that
1
(O) is open in X. We show that O is
open in X/M. Consider an open ball B(0; r), r > 0 in X. Let x B(0; r). Then
|x| < r, and so
|[x]| |x| < r.
On the other hand, if |[x]| < r, then there is an y M such that |x + y| < r.
Hence x +y B(0; r), and so
[x] = (x +y)
_
B(0; r)
_
.
If [x
0
] O, then x
0

1
(O). Since
1
(O) is open in X, there is an r > 0 such
that
B(x
0
; r)
1
(O).
This implies that
O =
1
(O)
_
B(x
0
; r)
_
=
_
x
0
+B(0; r)
_
= x +M : |x x
0
+M| < r.
The last set is the open ball of radius r > 0 centered at [x
0
] O. Thus O is open
in X/M.
Problem 11.
Let X = C[0, 1], M = f C[0, 1] : f(0) = 0. Show that X/M = C.
Solution.
Given [f] X/M, let ([f]) = f(0). Then the map : X/M C is well-dened.
Indeed, if [f] = [g], then (f g)(0) = 0 so f(0) = g(0). It is clearly linear.
If f X, then g = f f(0) M, and so f g = f(0) is constant, which tells us
that
|[f]| = [f(0)[ = [([f])[,
so is an isometry (and thus injective and continuous). Finally, constants are in
X = C[0, 1], so is surjective and thus an isometric isomorphism.
Problem 12.
If 0 < p < 1 then
p
is a vector space but |x|
p
= (
n
[x
n
[
p
)
1/p
is not a norm for
p
.
www.MATHVN.com
13
Solution.
Recall that if x = (x
1
, x
2
, ...), y = (y
1
, y
2
, ...)
p
and C then
x +y = (x
1
+y
1
, x
2
+y
2
, ...) and x = (x
1
, x
2
, ...).
It is clear that x
p
. We show that x + y
p
. For t 0 it not hard to see that
(1 +t)
p
1 +t
p
, 0 < p < 1. This implies that
(a +b)
p
a
p
+b
p
, 0 < p < 1 and a, b 0.
Therefore,
|x +y|
p
p
|x|
p
p
+|y|
p
p
.
Since both |x|
p
p
and |y|
p
p
are bounded, |x +y|
p
p
is bounded. Hence x +y
p
.
To show |.|
p
is not a norm for
p
, let us take an example: If
x = (1, 0, ...) and y = (0, 1, 0, ...)
then |x|
p
= |y|
p
= 1 but |x +y|
p
= 2
1/p
> 2 since 1/p > 1. Therefore,
|x +y|
p
> |x|
p
+|y|
p
.
Problem 13.
Suppose that X is a linear space with inner product ., .). If x
n
x and y
n
y
as n , prove that
x
n
, y
n
) x, y) as n .
Solution.
Using the Cauchy-Schwarz and triangle inequalities, we have
[x
n
, y
n
) x, y)[ [x
n
x, y
n
)[ +[x, y
n
y)[
|x
n
x||y
n
| +|x||y
n
y|
|x
n
x|(|y
n
y| +|y|) +|x||y
n
y|
|x
n
x||y
n
y| +|x
n
x||y| +|x||y
n
y|.
Since |x
n
x| 0 and |y
n
y| 0 as n , we see that
x
n
, y
n
) x, y) as n .
www.MATHVN.com
Problem 14.
Prove that if M is a closed subspace and N is a nite dimensional subspace of a
normed space X, then M +N := m+n : m M, n N is closed.
Solution.
Assume dimN = 1, say N = Spanx. The case x M is trivial. Suppose x / M.
Consider the sequence z
k
:=
k
x + m
k
, where m
k
M,
k
C, and suppose
z
k
M + N y as k . We want to show y M + N. The sequence (
k
)
is bounded; otherwise, there exists a subsequence (
k
) such that 0 < [
k
[ as
k
t
. Then
z
k
and
m
k
0 as k
t
,
so x must be 0, which is in M. This is a contradiction. Consequently, (
k
) is bounded
and therefore it has a subsequence (
k
) which is converging to some C. Thus
m
k
= z
k

k
x y x as k
t
.
Hence, y x is in M since M is closed. Thus y M +N.
The solution now follows by induction.
www.MATHVN.com
Chapter 2
Banach Spaces
Problem 15.
Let X be a normed space. Prove that X is a Banach space if and only if the series
n=1
a
n
converges, where (a
n
) is any sequence in X satisfying
n=1
|a
n
| < .
Solution.
Suppose that X is complete. Let (a
n
) be a sequence in X such that
n=1
|a
n
| < .
Let S
n
=
n
i=1
a
i
be the partial sum. Then for m > n,
|S
m
S
n
| =
_
_
_
_
_
m
i=n+1
a
i
_
_
_
_
_
i=n+1
|a
i
|.
By hypothesis, the series
n=1
|a
n
| converges, so
m
i=n+1
|a
i
| 0 as n .
Therefore, (S
n
) is a Cauchy sequence in the Banach space X. Thus, (S
n
) converges,
that is, the series
n=1
a
n
converges.
Conversely, suppose
n=1
a
n
converges in X whenever
n=1
|a
n
| < . We show
that X is complete. Let (y
n
) be a Cauchy sequence in X. Then
n
1
N : |y
n
1
y
m
| <
1
2
whenever m > n
1
,
n
2
N : |y
n
2
y
m
| <
1
2
2
whenever m > n
2
> n
1
.
Continuing in this way, we see that there is a sequence (n
k
) strictly increasing such
that
|y
n
k
y
m
| <
1
2
k
whenever m > n
k
.
15
www.MATHVN.com
16 CHAPTER 2. BANACH SPACES
In particular, we have
|y
n
k+1
y
n
k
| <
1
2
k
for all k N.
Set x
k
= y
n
k+1
y
n
k
. Then
n
k=1
|x
k
| =
n
k=1
|y
n
k+1
y
n
k
| <
n
k=1
1
2
k
.
It follows that
k=1
|x
k
| < . By hypothesis, there is an x X such that
m
k=1
x
k
x as m . But we have
m
k=1
x
k
=
m
k=1
(y
n
k+1
y
n
k
)
= y
n
m+1
y
n
1
.
Hence y
n
m
x + y
n
1
in X as m . Thus, the sequence (y
n
) has a convergent
subsequence and so must itself converges.
Problem 16.
Let X be a Banach space. Prove that the closed unit ball B(0; 1) X is compact
if and only if X is nite dimensional.
Solution.
Suppose dimX = n. Then X is isomorphic to R
n
(with the standard topology).
The result then follows from the Heine-Borel theorem.
Suppose that X is not nite dimensional. We want to show that B(0; 1) is not
compact. To do this, we construct a sequence in B(0; 1) which have no convergent
subsequence.
We will use the following fact usually known as Rieszs Lemma: (See the proof below)
Let M be a closed subspace of a Banach space X. Given any r (0, 1), there exists an x X such
that
|x| = 1 and d(x, M) r.
Pick x
1
X such that |x
1
| = 1. Let S
1
:= Span x
1
. Then S
1
is closed. According
to Rieszs Lemma, there exists x
2
X such that
|x
2
| = 1 and d(x
2
, S
1
)
1
2
.
www.MATHVN.com
17
Now consider the subspace S
2
generated by x
1
, x
2
. Since X is innite dimensional,
S
2
is a proper closed subspace of X, and we can apply the Rieszs Lemma to nd
an x
3
X such that
|x
3
| = 1 and d(x
3
, S
2
)
1
2
.
If we continue to proceed this way, we will have a sequence (x
n
) and a sequence of
closed subspaces (S
n
) such that for all n N
|x
n
| = 1 and d(x
n+1
, S
n
)
1
2
.
It is clear that the sequence (x
n
) is in B(0; 1), and for m > n we have
|x
n
x
m
| d(x
m
, S
n
)
1
2
.
Therefore, no subsequence of (x
n
) can form a Cauchy sequence. Thus, B(0; 1) is not
compact.
Proof of Rieszs Lemma:
Take x
1
/ M. Put d = d(x
1
, M) = inf
mM
|m x
1
|. Then d > 0 since M is closed. For any
> 0, by denition of the inmum, there exists m
1
M such that
0 < |m
1
x
1
| < d +.
Set x =
x
1
m
1
x
1
m
1
. Then |x| = 1 and

|x m| =
1
|x
1
m
1
|
_
_
x
1
(m
1
+|x
1
m
1
|m)
. .
M
_
_
This implies that
d(x, M) = inf
mM
|x m| =
inf
mM
|x
1
m|
|x
1
m
1
|

d
d +
.
By choosing > 0 small,
d
d+
can be arbitrary close to 1.
Problem 17.
Let X be a Banach space and M a closed subspace of X. Prove that the quotient
space X/M is also a Banach space under the quotient norm.
Solution.
We use criterion established above (in problem 15). Suppose that ([x
n
]) is any
www.MATHVN.com
sequence in X/M such that
n=1
|[x
n
]| < . We show that
[x] X/M :
k
n=1
[x
n
] [x] as k .
For each n, |[x
n
]| = inf
zM
|x
n
+z|, and therefore there is z
n
M such that
|x
n
+z
n
| |[x
n
]| +
1
2
n
by denition of the inmum. Hence
n=1
|x
n
+z
n
|
n=1
|[x
n
]| +
1
2
n
< .
But (x
n
+z
n
) is a sequence in the Banach space X, and so
n=1
(x
n
+z
n
) = x for some x X.
Then we have
_
_
_
_
_
k
n=1
[x
n
] [x]
_
_
_
_
_
=
_
_
_
_
_
k
n=1
[x
n
x]
_
_
_
_
_
= inf
zE
_
_
_
_
_
k
n=1
(x
n
x +z)
_
_
_
_
_
_
_
_
_
_
k
n=1
_
(x
n
x) +z
n
_
_
_
_
_
_
=
_
_
_
_
_
k
n=1
(x
n
+z
n
) x
_
_
_
_
_
0 as k .
Hence
k
n=1
[x
n
] [x] as k .
www.MATHVN.com
19
SPACE
p
(Only properties concerning to norms and completeness will be considered. Other
properties such as duality will be discussed later.)
Problem 18.
Show that
p
, 1 p < equipped with the norm |.|
p
is a Banach space.
Solution.
Let x
(i)
= (x
(i)
1
, ..., x
(i)
k
, ...) for i = 1, 2, ... be a Cauchy sequence in
p
. Then
|x
(i)
x
(j)
|
p
0 as i, j .
Since |x
(i)
x
(j)
|
p
[x
(i)
k
x
(j)
k
[ for every k, it follows that
[x
(i)
k
x
(j)
k
[ 0 for every k as i, j .
This tells us that the sequence
_
x
(i)
k
_
is a Cauchy sequence in F, which is complete,
so that
_
x
(i)
k
_
converges to x
k
F as i for each k. Set x = (x
1
, ..., x
k
, ...). We
will show that
() |x
(i)
x|
p
0 as i, j and x
p
.
Given > 0, for any M N, there exists N N such that
_
M
k=1
[x
(i)
k
x
(j)
k
[
p
_
1
p
k=1
[x
(i)
k
x
(j)
k
[
p
_1
p
< if i, j > N.
Letting j , for i > N we get
()
_
M
k=1
[x
(i)
k
x
k
[
p
_
1
p
< .
By Minkowskis inequality,
_
M
k=1
[x
k
[
p
_
1
p
_
M
k=1
[x
(N)
k
x
k
[
p
_
1
p
+
_
M
k=1
[x
(N)
k
[
p
_
1
p
_
M
k=1
[x
(N)
k
x
k
[
p
_
1
p
+
_

k=1
[x
(N)
k
[
p
_1
p
< +
_

k=1
[x
(N)
k
[
p
_1
p
.
www.MATHVN.com
Letting M , since the last sum is nite, we see that
|x|
p
=
_

k=1
[x
k
[
p
_1
p
< .
This shows that x
p
. Finally letting M in (), for i > N we get
|x
(i)
x|
p
=
_

k=1
[x
(i)
k
x
k
[
p
_1
p
< .
This shows that x
(i)
x in
p
as required.
Problem 19.
(a) Show that
equipped with the norm |.|
is a Banach space.
(b) Let c
0
be the space of sequences converging to 0. Show that c
0
is a closed
subspace of
.
Solution.
(a) We need to show that
is complete. Assume that the sequence (x

(n)
) is Cauchy
in
. That is, for all > 0, there exists N N such that

(1) n, m N |x
(n)
x
(m)
|
< .
For a xed n, we write x
(n)
= (x
(n)
1
, x
(n)
2
, ...). Then for N = N() as above,
(2) [x
(n)
j
x
(m)
j
[ |x
(n)
x
(m)
|
< for all j.

So, for a xed j, the sequence (x
(n)
j
) is Cauchy, and therefore convergent in C.
Denote
x
j
:= lim
n
x
(n)
j
, and x := (x
1
, x
2
, ...).
We need to show x
and x
(n)
x as n . In (2), for a xed j, letting
n yields
[x
j
x
(m)
j
[ < for all m N.
Therefore
sup
j
[x
j
x
(m)
j
[ := |x x
(m)
|
for all m N.
That is x
(m)
x as m in
. Now for all j N, n N,

[x
j
[ [x
j
x
(n)
j
[ +[x
(n)
j
[ |x x
(n)
|
+|x
(n)
|
+|x
(n)
|
< .
www.MATHVN.com
21
This shows that |x|
< and so x
.
(b) Of course c
0

. Assume (x
(n)
) c
0
that converges in
to x. We have to
show that x c
0
. Let > 0 be arbitrary. Since x
(n)
x in
, we can choose
N N such that
|x
(N)
x|
<

2
.
Since x
(N)
= (x
(N)
1
, x
(N)
2
, ...) c
0
we have x
(N)
j
0 as j . Therefore, choose
J N such that
j J [x
(N)
j
[ <

2
.
Then, for j J,
[x
j
[ [x
j
x
(N)
j
[ +[x
(N)
j
[ |x x
(N)
|
+[x
(N)
j
[ <

2
+

2
= .
Therefore, x = (x
1
, x
2
, ...) c
0
.
Problem 20.
(a) Let c
00
be the space of sequences such that if x = (x
n
)
nN
c
00
then x
n
= 0
for all n n
0
, where n
0
is some integer number. Show that c
00
with the norm
|.|
is NOT a Banach space.

(b) What is the closure of c
00
in
?
Solution.
We observe that c
00
c
0

.
(a) Consider the sequence x
(n)
dened by
x
(n)
=
_
1,
1
2
,
1
3
, ...,
1
n
, 0, 0, ...
_
c
00
.
Then, for n, m N,
|x
(n)
x
(m)
|
=
_
1
m+1
if n m,
1
n+1
if n m.
In both cases we have for any N > 0
|x
(n)
x
(m)
|

1
N + 1
for n, m N.
So the sequence x
(n)
is a Cauchy sequence in c
00
. Evidently, it is also a Cauchy
sequence in
. Now consider the sequence

x =
_
1,
1
2
,
1
3
, ...,
1
n
,
1
n + 1
,
1
n + 2
, ...
_

.
www.MATHVN.com
We see that x
(n)
/ c
00
, and
lim
n
|x
(n)
x|
= lim
n
1
n + 1
= 0.
This tells us that there is a Cauchy sequence in c
00
which does not converge to
something in c
00
. Therefore, c
00
equipped with the |.|
norm is not a Banach

space.
(b) We claim that c
00
= c
0
(closure taken in
).
According to Problem 19, c
0
is closed, so we have c
00
c
0
. We show the inverse
inclusion. Take an arbitrary sequence x = (x
1
, x
2
, ...) c
0
. We build a sequence a
(n)
from x as follows:
a
(n)
= (x
1
, x
2
, .., x
n
, 0, 0....).
It is clear that a
(n)
c
00
. Now since the sequence x converges to 0, given any > 0,
there exists N N such that
[x
i
[ < for i N.
Then
|a
(n)
x|
= sup
iN
[x
i
[ .
hence a
(n)
converges to x (in
). Hence x c
00
. Thus, c
0
c
00
.
Problem 21.
Prove that:
(a) If 1 p < q < , then
p

q
and |x|
q
|x|
p
.
(b) If x
1p<
p
then |x|
p
|x|
as p .
Solution.
(a) Let x = (x
1
, x
2
, ...)
p
. Then, for n large enough, we have [x
n
[ < 1 and hence
[x
n
[
q
[x
n
[
p
since 1 p < q < . That implies x
q
. Now we want to show
_
[x
n
[
q
_
1/q
[x
n
[
p
_
1/p
.
Let a
n
= [x
n
[
p
and =
q
p
> 1. The above inequality is equivalent to
n

_
a
n
_
,
www.MATHVN.com
23
which follows by
n
(max a
n
)
1
a
n

_
a
n
_
1
a
n
=
_
a
n
_
.
(b) Let x = (x
1
, x
2
, ...)
p
0
for some p
0
. Clearly, |x|
p
max
n
[x
n
[ = |x|
for any
nite p. On the other hand,
|x|
p
=
_
n
[x
n
[
p
_
1/p
_
|x|
pp
0
n
[x
n
[
p
0
_
1/p
= |x|
pp
0
p
|x|
p
0
p
p
0
p
|x|
Problem 22.
Prove that:
(a) If 1 p < then
p
is separable.
(b)
is not separable.
Solution.
(a) First we show that E := x
p
: x
n
= 0, n N for some N is dense
in
p
. Indeed, if x
p
, x =
k=1
x
k
e
k
, where e
k
is the sequence such that the
k-component is 1 and the others are zero, then
_
_
_
_
_
x
n
k=1
x
k
e
k
_
_
_
_
_
p
=
_

k=n+1
[x
k
[
p
_
1/p
0 as n .
But
n
k=1
x
k
e
k
E, so E is dense in
p
. Now let A E consisting of elements
x = (x
1
, x
2
, ..., x
n
, 0, 0, ...) E such that x
k
= a
k
+ib
k
, a
k
, b
k
Q. Since Q is dense
in R, A is dense in E. Hence A is dense in
p
. Since A is countable,
p
is separable.
(b) We now show that it is not the case for
.
Let F := x
: k 1, x
k
= 0 or x
k
= 1. Then F is uncountable. Note
that for x F, |x|
= 1. Moreover, x, y F, x ,= y |x y|
= 1. Assume
that
is separable. Then there is a set A = a

1
, a
2
, ... dense in
. So, for all

x F, there exists k N such that |x a
k
|

1
3
. Let T be the family of closed
balls B(x;
1
3
), x F. If B ,= B
t
then B B
t
= . This allows us to construct an
injection f : T A which maps each B T with an element a B A. This is
impossible since T is uncountable and A is countable.

www.MATHVN.com
Problem 23. (The space C[0, 1])
Let C[0, 1] be the space of all continuous functions on [0, 1].
(a) Prove that if C[0, 1] is equipped with the uniform norm
|f| = max
x[0,1]
[f(x)[, f C[0, 1]
then C[0, 1] is a Banach space.
(b) Give an example to show that C[0, 1] equipped with the L
1
-norm
|f|
1
=
_
1
0
[f(x)[dx, f C[0, 1]
is not a Banach space.
Solution.
(a) Let (f
n
) be a Cauchy sequence in C[0, 1] with respect to the uniform norm. Then
for any > 0, there exists N N such that
|f
m
f
n
| < for m, n N.
Therefore
() [f
m
(x) f
n
(x)[ < for m, n N and x [0, 1].
This shows that for every x [0, 1], the sequence
_
f
n
(x)
_
is a Cauchy sequence of
numbers and therefore converges to a number which depends on x, say, f(x). In
(), x n and let m , we have
() [f(x) f
n
(x)[ < for n N and x [0, 1].
Thus the sequence (f
n
) converges uniformly to f on [0, 1] so that f is continuous on
[0, 1], that is, f C[0, 1]. From () we obtain
max
x[0,1]
[f(x) f
n
(x)[ = |f f
n
| for n N.
This shows that
lim
n
|f f
n
| = 0.
(b) For each n N, consider the function
f
n
(x) =
_
nx if 0 x <
1
n
1 if
1
n
x 1.
www.MATHVN.com
25
One can check that the sequence (f
n
) is a Cauchy sequence with respect to the
L
1
-norm, but it converges to the function
f(x) =
_
0 if x = 0
1 if 0 < x 1,
which is not continuous, that is, f / C[0, 1]. Thus the space C[0, 1] is not complete.
www.MATHVN.com
www.MATHVN.com
Chapter 3
Hilbert Spaces
3.1 Hilbert spaces
Problem 24.
Let M be a closed subspace of a Hilbert space H. Prove that:
(a) (M
= M.
(b) If codimM := dimH/M = 1 then dimM
= 1.
Solution.
(a) In general, if M is a subset of H then M (M
. Indeed,
M
:= x X : xM.
So we have
x M xM
x (M
.
Now suppose M is a closed subspace of H and x (M
. Since x H = MM
,
we have
x = u +v, u M, v M
.
Since M (M
we have
x = u +v, u (M
, v M
.
Since x u (M
and v M
and v = x u we obtain
v M
(M
,
27
www.MATHVN.com
28 CHAPTER 3. HILBERT SPACES
which implies v = 0. Hence, x = u M.
(b) Assume that there are two linearly independent vectors x, y M
. Recall that
M is the zero vector of the linear space X/M. Consider the cosets [x], [y]. Assume
[x] +[y] = M, for some scalars , . Then x+y M and, since x+y M
as well, we conclude that x + y = 0. Hence, = = 0 and therefore [x], [y] are

linearly independent. This contradicts the hypothesis codimM = 1.
Problem 25.
Let T : H
1
H
2
be an isometry of two Hilbert spaces H
1
and H
2
, i.e., |Tx| =
|x| for every x H
1
. Prove that
Tx, Ty) = x, y) for every x, y H
1
.
Solution.
By hypothesis, we have
|T(x +y)|
2
= |x +y|
2
for every x, y H
1
.
By opening up the norm using the inner product, we obtain
ReTx, Ty) = Rex, y).
Similarly, |T(x +iy)|
2
= |x +iy|
2
gives that
ImTx, Ty) = Imx, y).
Hence,
Tx, Ty) = x, y).
Problem 26.
Let C be a closed convex set in a Hilbert space H. Show that C contains a unique
element of minimal norm.
Solution.
Let = inf
zC
|z|.
www.MATHVN.com
3.1. HILBERT SPACES 29
If = 0, then, by denition of inmum, there is a sequence (z
j
) in C such that
|z
j
| 0 as j . Therefore, z
j
0 as j . Since C is closed, 0 C, and 0
is the unique element of minimal norm.
Suppose > 0. First we show that C contains an element of minimal norm.
Take (z
j
) in C such that |z
j
| as j . The convexity of C implies that
1
2
(z
j
+z
k
) C, so that
|z
j
+z
k
|
2
= 4.
1
4
|z
j
+z
k
|
2
4
2
.
Recall now the parallelogram law:
|x y|
2
+|x +y|
2
= 2
_
|x|
2
+|y|
2
_
.
Applying this we have:
|z
j
z
k
|
2
= 2
_
|z
j
|
2
+|z
k
|
2
_
|z
j
+z
k
|
2
2
_
|z
j
|
2
+|z
k
|
2
_
4
2
4
2
4
2
= 0 as j, k .
Thus the sequence (z
j
) is Cauchy, so converges to some z H. Since C is closed,
z C. The norm function is continuous, so |z| = . This shows that C contains
an element of minimal norm.
Assume that there are two elements a
1
, a
2
C such that |a
1
| = |a
2
| = . By the
above we have
|a
1
+a
2
|
2
4
2
.
By the parallelogram law we have
|a
1
a
2
|
2
= 2
_
|a
1
|
2
+|a
2
|
2
_
|a
1
+a
2
|
2
4
2
4
2
= 0.
Hence, a
1
= a
2
.
Problem 27.
Let 1 p < . Prove that
p
is a Hilbert space if and only if p = 2.
Solution.
Suppose p = 2. In the space
2
the inner product is dened by
x, y) =
i=1
x
i
y
i
for x = (x
i
), y = (y
i
)
2
.
www.MATHVN.com
This inner product gives rise to the norm
|x|
2
=
_
x, x) =
_

i=1
[x
i
[
2
_1
2
.
According to Problem 18, the normed space
2
is complete. So
2
is a Hilbert space.
Consider the general case where 1 p < . Assume that
p
with the corre-
sponding inner product is a Hilbert space. Consider two elements e
n
, e
m

p
with
m ,= n dened as follows:
e
n
= (0, .., 0
. .
n1
, 1, 0, ...),
e
m
= (0, .., 0
. .
m1
, 1, 0, ...).
Since
p
is a Hilbert space, by the parallelogram law we have
|e
n
+e
m
|
2
p
+|e
n
e
m
|
2
p
= 2
_
|e
n
|
2
p
+|e
m
|
2
p
_
,
That is
2
2/p
+ 2
2/p
= 2
2
.
The unique solution of this equation is p = 2. We conclude that
p
is a Hilbert space
if and only if p = 2.
Problem 28.
Consider the Hilbert space H = L
2
[1, 1] equipped with the usual scalar product:
x, y) =
_
1
1
x(t)y(t)dt, x, y H.
Let M = x H :
_
1
1
x(t)dt = 0.
(a) Show that M is closed in H. Find M
.
(b) Calculate the distance from y to M for y(t) = t
2
.
Solution.
(a) Let 1 H be the function 1(t) = 1, t [1, 1]. Dene the map T : H C
by
x 1, x).
www.MATHVN.com
Then T is linear. We show that T is bounded, so continuous.
[Tx[ =
_
1
1
1(t)x(t)dt

_
1
1
[1(t)[ [x(t)[dt
__
1
1
1dt
_
1
2
__
1
1
[x
2
(t)[
2
dt
_
1
2
=
2|x|
2
.
By denition, M = Ker T = T
1
(0). Since T is continuous, M is closed. Further-
more,
x M 1, x) = 0 M = (Span1)
.
Since M is closed, (see Problem 23 a).
M
= Span1.
(b) The distance from y H to M is the length of the projection vector of y on
M
. We have
d(y, M) =
[1, y)[
|1|
2
=
__
1
1
t
2
dt
___
1
1
1dt
_
1/2
=
2
3
.
Problem 29.
Consider the Hilbert space H = L
2
[1, 1] equipped with the usual scalar product:
f, g) =
_
1
1
f(t)g(t)dt, f, g H.
Let E = x H : f(t) = f(t), t [1, 1].
(a) Show that E is closed in H. Find E
.
(b) Calculate the distance from h to E for h(t) = e
t
.
Solution.
(a) Dene

f(t) := f(t). Dene the map
S : H H dened by Sf =

f.
Clearly, S is linear. S is bounded, so continuous. Indeed,
|Sf|
2
= |
f|
2
=
__
1
1
[
f(t)[
2
dt
_
1
2
=
__
1
1
[f(t)[
2
dt
_
1
2
= |f|
2
.
www.MATHVN.com
In fact, S is an isometry. It follows that I S is continuous. By denition, E =
Ker(I S), so E is closed.
By denition, E consists of all even functions, so E
is the linear subspace of all

odd functions. In fact, we have
f(t) =
1
2
[f(t) +f(t)]
. .
+
1
2
[f(t) f(t)]
. .
(t) +(t),
with is even, is odd, and
, ) =
_
1
1
(t)(t)dt = 0, so that .
(b) The distance from h H to E is the length of the projection vector of h on E
.
By the above expression,
Proj
E
(h) =
1
2
[h(t) h(t)] =
1
2
(e
t
e
t
).
Therefore,
_
dist(h, E)
_
2
=
_
_
_
_
1
2
(e
t
e
t
)
_
_
_
_
2
=
1
4
_
1
1
[e
t
e
t
[
2
dt
=
1
4
(e
2
e
2
4).
Thus,
dist(h, E) =
1
2
e
2
e
2
4.
Problem 30.
Let H be a Hilbert space and M be a closed subspace of H. Denoting by P :
H M the orthogonal projection of H onto M, prove that, for any x, y H,
Px, y) = x, Py).
(This is telling us that P is self-adjoint).
Solution.
We know that if M is a closed subspace of H, then
www.MATHVN.com
For all u H, there exist unique u
M
M and u
M
M
such that
u = u
M
+u
M
.
If P : H M the orthogonal projection of H onto M, then
Pu = u
M
.
Now for arbitrary x, y H, we have
x = x
M
+x
M
y = y
M
+y
M
Px = x
M
Py = y
M
.
With these, can have
Px, y) = x
M
, y
M
+y
M
) = x
M
, y
M
),
since x
M
, y
M
) = 0, and
x, Py) = x
M
+x
M
, y
M
) = x
M
, y
M
),
since x
M
, y
M
) = 0. Thus,
Px, y) = x, Py).
Problem 31.
Let H be a Hilbert space and A H a closed convex non-empty set. Prove that
P
A
: H H is non-expansive, i.e.,
|P
A
(x) P
A
(y)| |x y|, x, y H.
(P
A
is the orthogonal projection on A).
Solution.
We claim:
() Re
_
x P
A
(x), P
A
(x) a
_
0, x H, a A.
Let x
A
= P
A
(x). Then x can be decomposed uniquely as
x = x
A
+x
t
A
, x
A
A, x
t
A
A
.
www.MATHVN.com
We have
2Re
_
x P
A
(x), P
A
(x) a
_
= x
t
A
, x
A
a) +x
A
a, x
t
A
)
= x
t
A
, x
t
A
) +x
A
a, x
A
a)
= |x
t
A
|
2
+|x
A
a|
2
0.
Hence () is proved.
Replacing a with P
A
(y) in (), we obtain
(3.1) Re
_
x P
A
(x), P
A
(x) P
A
(y)
_
0.
Analogously,
Re
_
y P
A
(y), P
A
(y) P
A
(x)
_
0.
And therefore,
(3.2) Re
_
P
A
(y) y, P
A
(x) P
A
(y)
_
0.
Adding 3.1 and 3.2 we get
Re
_
x y [P
A
(x) P
A
(y)], P
A
(x) P
A
(y)
_
0, i.e.,
Re
_
x y, P
A
(x) P
A
(y)
_
|P
A
(x) P
A
(y)|
2
. (3.3)
Form the Cauchy-Schwarz inequality, we have
Re
_
x y, P
A
(x) P
A
(y)
_
[x y, P
A
(x) P
A
(y))[
|x y| |P
A
(x) P
A
(y)|
and from here, replacing in 3.3, we obtain that
|P
A
(x) P
A
(y)|
2
|x y| |P
A
(x) P
A
(y)|.
Thus,
|P
A
(x) P
A
(y)| |x y|.
Problem 32.
Let X be a Hibert space, and G
1
G
2
... G
n
... be a sequence of closed
linear subspaces of X. Let
G = Span
_
_
nN
G
n
_
.
www.MATHVN.com
(a) Prove that d(x, G) = lim
n
d(x, G
n
), x X.
(b) Prove that P
G
(x) = lim
n
P
G
n
(x), x X. Note: P
G
is the orthogonal
projection of X on G.
Solution.
(a) Let
A = Span
_
_
nN
G
n
_
.
Then we have
d(x, G) = d(x,

A) = d(x, A).
For any > 0 we have
d(x, G) + = d(x, A) + > d(x, A) = inf
aA
|x a|.
From this, we deduce that there is some a
A such that
d(x, G) + > |x a
|.
Since a
Span
_
nN
G
n
_
, we can nd
1
, ...,
k
K and x
1
, ..., x
k

nN
G
n
such
that
a
=
1
x
1
+... +
k
x
k
.
Then there are n
1
, ..., n
k
N such that x
1
G
n
1
, ..., x
k
G
n
k
. Let n = maxn
1
, ..., n
k
.
By hypothesis, the sequence (G
n
) is increasing , so we have G
n
1
, ..., G
n
k
G
n
, so
x
1
, ..., x
k
G
n
. And since G
n
is a linear space,
a
=
1
x
1
+... +
k
x
k
G
n
.
That is,
() > 0, n N : d(x, G) + > |x a
| d(x, G
n
).
From G
n
G it follows that
() d(x, G) d(x, G
n
), n N.
() and () imply that
d(x, G) d(x, G
n
) < d(x, G) +.
www.MATHVN.com
Hence,
d(x, G) = inf
nN
d(x, G
n
).
From G
n
G
n+1
, it follows that d(x, G
n+1
) d(x, G
n
). The sequence of real
numbers
_
d(x, G
n
)
_
, which is decreasing and bounded below, must converge to its
inmum. Thus
lim
n
d(x, G
n
) = d(x, G).
(b) Let a
n
= P
Gn
(x), a = P
G
(x). Then by part (a) we have
lim
n
|x a
n
| = lim
n
d(x, G
n
) = d(x, G).
Since G
n
G, n N, then (a
n
) G and so a
n
P
G
(x) in norm, that is,
lim
n
|a
n
a| = 0,
which gives that
lim
n
P
Gn
(x) = P
G
(x).
Problem 33.
Let H be a Hilbert space.
(a) Prove that for any two subspaces M, N of H we have
(M +N)
= M
.
(b) Prove that for any two closed subspaces E, F of H we have
(E F)
= E
+F
.
Solution.
(a) If x (M +N)
, then for every m M and n N we have

m+n, x) = 0
since m+n M +N. For n = 0 we have m, x) = 0. This holds for all m M, so
x M
. Similarly x N
. Thus x M
, and hence (M +N)
.
If x M
, then we have
m, x) = 0 and n, x) = 0, m M, n N.
www.MATHVN.com
Hence
m+n, x) = 0.
This means that x (M +N)
. Hence M
(M +N)
.
(b) From part (a) it follows that
M +N = (M
.
Setting E
in the place of M and F
in the place of N, we obtain

E
+F
= (E F)
.
Why in question (b) E and F must be closed?
Problem 34.
A system x
i
iN
in a normed space X is called a complete system if
Span
n
i=1
i
x
i
: n N,
i
F is dense on X.
If x
i
iN
is a complete system in a Hilbert space H and xx
i
for every i, show
that x = 0.
Solution.
Given x X, if xx
i
for every i, then xSpan
n
i=1
i
x
i
. Let D := Span
n
i=1
i
x
i
.
By denition, D = X. Then there exists a sequence (x
n
) in D converging to x and
xx
n
for every n. Hence
0 = x, x
n
) x, x) = |x|
2
as n .
Thus x = 0.
Problem 35.
Let H be a Hilbert space and
i
i=1
an orthonormal system in H. Show that
|
n
m
| =
2 for m ,= n.
Solution.
Using orthogonality and the fact that |
k
| = 1, k N, we get for n ,= m,
|
n
m
|
2
=
n
m
,
n
m
)
=
n
,
n
)
n
,
m
)
m
,
n
) +
m
,
m
)
= |
n
|
2
+|
m
|
2
= 2.
www.MATHVN.com
Problem 36.
Let e
i
iN
be an orthonormal set in a Hilbert space H. Prove that
i=1
[x, e
i
)y, e
i
)[ |x| |y|, x, y H.
Solution.
Using the Cauchy-Schwarz inequality for
2
and the Bessel inequality for H, we have
i=1
[x, e
i
)y, e
i
)[
_

i=1
[x, e
i
)[
2
_1
2
_

i=1
[x, e
i
)[
2
_1
2
_
|x|
2
_1
2
_
|y|
2
_1
2
= |x| |y|.
Problem 37.
Let H be a Hilbert space and A and B be any two subsets of H. Show that
(a) A
is a closed subspace of H.
(b) A (A
:= A
.
(c) A B B
.
(d) A
= A
= A
.
Solution.
(a) Let x, y A
and , F. Then for any a A,

x +y, a) = x, a) +y, a)
= x, a) +y, a) = 0.
Hence x +y A
, so A
is a subspace of H.
We show now that A
is closed. Suppose that the sequence (x

n
) in A
converges
to some x H. Since x
n
, a) = 0 for all a A, by continuity of the inner product,
we have
x, a) = lim
n
x
n
, a) = lim
n
x
n
, a) = 0.
So x A
, and A
is closed.
www.MATHVN.com
(b) For any x A, we have xA
. This implies that x (A
. Thus
A (A
:= A
.
(c) Take any x B
. Then xy for all y B. But A B, so xy for all y A.

Hence x A
.
(d) From (a) we have
A
= A
,
which is the second equality in (d). Now pick any xA, that is, x A
. If a A,
then there exists a sequence (a
n
) is A such that a
n
a. Since xA, we have xa
n
for all n. Hence
x, a) = x, lim
n
a
n
) = lim
n
x, a
n
) = 0.
Thus we have xA, so xA. Therefore,
A
. (i)
On the other hand, by (b) we have
A A =A
. (ii)
(i) and (ii) complete the proof.
Problem 38.
(a) Show that M :=
_
x = (x
n
)
2
: x
2n
= 0, n N
_
is a closed subspace of
the Hilbert space
2
.
(b) Find M
.
Solution.
(a) Take any x = (x
n
), y = (y
n
) M. It is clear that for any scalars , ,
(x +y)
2n
= x
2n
+y
2n
= 0.
That gives that x +y M. Hence M is a linear subspace of
2
.
Let us prove that it is closed. Take x M. There exists a sequence x
(k)
= (x
(k)
n
) M
converging to x as k . Since x
(k)
2n
= 0, we obtain
x
2n
= lim
k
x
(k)
2n
= 0,
www.MATHVN.com
that is, x M. Hence M is closed.
(b) Now
z M
z, x) = 0, x M
n=0
z
2n+1
x
2n+1
= 0 for all scalars x
2n+1
such that
n=0
[x
2n+1
[
2
<
z
2n+1
= 0, n = 0, 1, 2, ...
Therefore
M
= z = (z
n
)
2
: z
2n+1
= 0, n = 0, 1, 2, ....
Problem 39.
Let V be a subspace of a Hilbert space.
(a) Show that V
= V
.
(b) Show that V is dense in H if and only if V
= 0.
Solution.
(a) From Problem 37d we get (a).
(b) If V is dense, then V = H. Hence
V
= V
= H
= 0.
Conversely, suppose V
= 0. If V is not dense in H, that is, V H, pick

x H V . Let x
t
= P
V
x. Then
x x
t
V
= V
= 0.
Thus x = x
t
V . This is a contradiction. Thus V is dense in H.
3.2 Weak convergence
Problem 40.
Prove that in any nite dimensional vector space, strong convergence and weak
convergence are equivalent.
www.MATHVN.com
3.2. WEAK CONVERGENCE 41
Solution.
Consider rst the case that X = F
d
under the Euclidian norm |.|
2
. Suppose that
the sequence (x
n
) converges weakly to x in F
d
. Then for each standard basis vector
e
k
, k = 1, 1, .., d, we have
x
n
, e
k
) x, e
k
) as n .
That is, weak convergence implies componentwise convergence. But since there are
only nitely many components, this implies norm convergence, since
|x
n
x|
2
2
=
d
k=1
[x
n
, e
k
) x, e
k
)[
2
0 as n .
For the general case, choose any basis e
1
, e
2
, ..., e
d
for X, and then use the fact that
all norms on X are equivalent to dene an isomorphism between X and F
d
.
Problem 41.
Show that if the sequence (x
n
) in a normed space X is weakly convergent to
x
0
X, then
liminf
n
|x
n
| |x
0
|.
Solution.
If x
0
= 0 then |x
0
| = 0 and the statement is obviously true. Now assume |x
0
| ,= 0.
By a well known theorem
1
, there is some f X
such that
|f| = 1, f(x
0
) = |x
0
|.
Since (x
n
) converges weakly to x
0
and f is continuous, we have
lim
n
f(x
n
) = f(x
0
) = |x
0
|.
But
f(x
n
) [f(x
n
)[ |f| |x
n
| = |x
n
|.
1
(Kreyszig, p 223) Let X be a normed space and x
0
,= 0 be an element in X. Then there exists
a bounded linear functional f X
such that
|f| = 1, f(x
0
) = |x
0
|.
www.MATHVN.com
Hence,
liminf
n
|x
n
| lim
n
f(x
n
) = |x
0
|.
Note.
If X = H is a Hilbert space, using the denition of weak convergence we can have
dierent solution.
|x
0
|
2
= x
0
, x
0
) = lim
n
x
0
, x
n
).
Since x, x
n
) |x
0
| |x
n
|, so we have
|x
0
|
2
= lim
n
x, x
n
) |x
0
[ liminf
n
|x
n
|.
Thus
|x
0
| liminf
n
|x
n
|.
Problem 42.
Let X and Y be normed spaces, T B(X, Y ) and (x
n
) a sequence in X. Show
that if x
n
w
x, then Tx
n
w
Tx
Solution.
Recall: Denition of weak convergence in a normed space:
x
n
w
x f(x
n
) f(x), f X
.
We must show that
(Tx
n
) (Tx), Y
.
That is,
( T)x
n
( T)x, Y
.
But T X
, so our hypothesis x
n
w
x guarantees our desired conclusion.
Problem 43.
Let H be a Hilbert space and (x
n
) be a sequence in H. Suppose x
n
w
x. Show
that
lim
n
|x
n
x| = 0 |x| limsup
n
|x
n
| (1).
www.MATHVN.com
3.2. WEAK CONVERGENCE 43
Solution.
By Problem 41 we have
|x| liminf
n
|x
n
|.
So the right hand side of (1) is equivalent to |x| = lim
n
|x
n
|. Now note that
(i) |x
n
x|
2
= |x
n
|
2
2Rex
n
, x) +|x|
2
(ii)
|x
n
| |x|
|x
n
x|.
If lim
n
|x
n
x| = 0 then by (ii) we get
lim
n
|x
n
| = |x|.
If lim
n
|x
n
| = |x| then (i) give that
lim
n
|x
n
x| = 0. .
www.MATHVN.com
www.MATHVN.com
Chapter 4
Linear Operators - Linear
Functionals
4.1 Linear bounded operators
Problem 44.
Let (X, |.|
1
) and (Y, |.|
2
) be norm spaces, and T B(X, Y ). We dene |T| as
|T| := inf
xX
M : |Tx|
2
M|x|
1
.
(a) Show that
|T| = sup
|x|
1
=1
|Tx|
2
= sup
|x|
1
1
|Tx|
2
= sup
xX, x,=0
|Tx|
2
|x|
1
.
(b) Show that
|T| = sup
|x|
1
<1
|Tx|
2
.
Solution.
(a) We have
|Tx|
2
|x|
1
M, for all x ,= 0, x X.
45
www.MATHVN.com
46 CHAPTER 4. LINEAR OPERATORS - LINEAR FUNCTIONALS
By denition,
|T| := inf
xX
M : |Tx|
2
M|x|
1
= sup
xX, x,=0
|Tx|
2
|x|
1
(i).
Now, let y =
x
|x|
1
for x X, x ,= 0. Then y X and |y| = 1. By (i) we have
|T| = sup
xX, x,=0
|Tx|
2
|x|
1
= sup
|y|=1
_
_
_
_
T
_
|x|
1
y
|x|
1
__
_
_
_
2
= sup
|y|=1
|Ty|
2
= sup
|x|=1
|Tx|
2
(ii).
From (ii) it follows that
|T| = sup
|x|
1
=1
|Tx|
2
sup
|x|
1
1
|Tx|
2
sup
xX, |x|
1
1
|Tx|
2
|x|
1
sup
xX, x,=0
|Tx|
2
|x|
1
= |T|.
Thus,
|T| = sup
|x|
1
1
|Tx|
2
.
(b) Let
B := x X : |x|
1
1 and B
o
:= x X : |x|
1
< 1.
Since |T| = sup
xB
|Tx|
2
, there exists a sequence (x
n
) in B such that
|T| = lim
n
|Tx
n
|
2
.
Consider the sequence (y
n
) dened by y
n
= (1
1
2
n
)x
n
. It is clear that y
n
B
o
for
all n N. Moreover,
lim
n
|Ty
n
|
2
= lim
n
_
_
_
_
T
_
1
1
2
n
_
x
n
_
_
_
_
= lim
n
_
1
1
2
n
_
|Tx
n
|
2
= lim
n
(1
1
2
n
) lim
n
|Tx
n
|
2
= |T|.
Thus
sup
xB
o
|Tx|
2
|T|.
On the other hand,
sup
xB
o
|Tx|
2
sup
xB
|Tx|
2
= |T|.
The proof is complete.
www.MATHVN.com
4.1. LINEAR BOUNDED OPERATORS 47
Problem 46.
Let H =
2
be the well known Hilbert space. Consider the left shift dened by
L :
2
2
, x = (x
1
, x
2
, x
3
, ...) Lx = (x
2
, x
3
, ...).
(a) Show that L is a linear bounded operator. Find |L|.
(b) Dene the right shift and answer the same question as in part (a).
Solution.
We answer only the rst part.
It is easy to check the linearity of L.
For any x = (x
1
, x
2
, x
3
, ...)
2
, we have
|Lx|
2
= |(x
2
, x
3
, ...)|
2
=
n=2
[x
n
[
2
n=1
[x
n
[
2
= |x|
2
.
Hence
|Lx| |x|, x
2
.
This implies that T is bounded, and
|L| 1. ()
On the other hand, consider the sequence e = (0, 1, 0, 0, ...)
2
. We have
|e| = 1 and |Le| = |(1, 0, 0, ...)| = 1.
So (see Problem 39)
|L| = sup
|x|=1
|Lx| 1. ()
Combining () and () we obtain |L| = 1.
Problem 47.
Let X = C[0, 1] with the max-norm (the uniform norm). We dene the integral
operator
K : X X by Kf(x) =
_
x
0
f(y)dy.
Show that K is bounded. Find |K|.
www.MATHVN.com
Solution.
The operator K is bounded. Indeed,
|Kf| sup
x[0,1]
_
x
0
[f(y)[dy
_
1
0
[f(y)[dy |f|.
Hence |K| 1.
In fact, |K| = 1, since 1 X, K1 = x and so |K1| = |x| = 1.
**Problem 48.
Let X = L
2
[0, 1] with the norm |.|
2
. We dene the integral operator
A : X X by Af(x) =
_
x
0
f(y)dy.
Show that A is bounded. Find |A|.
Solution.
Warning: Its completely dierent from the previous problem!!!
First solution: Using Cauchy-Schwarz inequality, we get
|Af|
2
2
=
_
1
0
_
t
0
f(s)ds
2
dt =
_
1
0
_
t
0
_
cos
s
2
.
f(s)
_
cos
s
2
ds
2
dt
_
1
0
__
t
0
cos
s
2
ds
_
t
0
[f(s)[
2
cos
s
2
ds
_
dt
=
2
_
1
0
__
t
0
sin
t
2
.
[f(s)[
2
cos
s
2
ds
_
dt
=
2
_
t
0
__
1
0
sin
t
2
[f(s)[
2
cos
s
2
dt
_
ds
=
2
_
t
0
__
1
0
sin
t
2
dt
_
[f(s)[
2
cos
s
2
ds
=
_
2
_
2
_
t
0
[f(s)[
2
cos
s
2
ds.
Equality holds for f(s) = cos
s
2
. Thus
|A| =
2
.
www.MATHVN.com
Second solution: We nd the norm of T = A
A, then |A| =
_
|T|.
Since A is compact (we will see this somewhere later) and A
A is self adjoint, we
have that T is a compact, normal operator. Hence, its spectrum is countable with
0 as the only possible cluster point and |T| is equal to the spectral radius of T.
These two facts together imply that
|T| = max[[ : Tf = f for some f X,
We have that A
is given by
A
f(x) =
_
1
x
f(y)dy.
Hence
() Tf(x) =
_
1
x
_
y
0
f(z)dz dy.
By the Fundamental Theorem of Calculus, Tf is twice dierentiable for all f X,
so any eigenvector of T must satisfy the dierential equation
d
2
dx
2
f(x) =
d
2
dx
2
_
1
x
_
y
0
f(z)dz dy =
d
dx
_
_
x
0
f(z)dz
_
= f(x), C 0.
By a theorem from dierential equations, we know that there are only two linearly
independent solutions to the above dierential equation over C, namely, e
ix
and
e
ix
, where
2
= , , ,= 0. Hence the general solution to the above equation is
f(x) = e
ix
+e
ix
, , C.
www.MATHVN.com
When we apply T, we get
T(e
ix
+e
ix
) =
_
1
x
_
y
0
(e
iz
+e
iz
)dz dy
=
_
1
x
_
e
iz
i

e
iz
i
_
y
0
dy
=
_
1
x
__
e
iy
i

e
iy
i
_
_

i

i
__
dy
=
_
e
iy
2

e
iy
2
_
_

i

i
_
y
1
x
=
_
e
i
2

e
i
2
_
_

i

i
_
__
e
ix
2

e
ix
2
_
_

i

i
_
x
_
=
1
2
(e
ix
+e
ix
) +
_

i

i
_
x
_
e
i
2

e
i
2
_
_

i

i
_
.
Now since e
ix
+e
ix
is an eigenvector, we must have that
_

i

i
_
x
_
e
i
2

e
i
2
_
_

i

i
_
= 0, x [0, 1],
which implies = . Hence
e
ix
+e
ix
= 2cos x.
Moreover, we must have that Tf(1) = 0 for any eigenvector f by (), so 2cos = 0.
Since ,= 0, we must have that
=
(2n + 1)
2
, n Z.
Hence the eigenvalues are of the form
n
=
1
2
n
=
2
2
(2n + 1)
2
2
, n Z.
Thus
max
n
: n Z =
4
2
(take n = 0, 1).
www.MATHVN.com
We conclude that
|T| =
4
2
and |A| =
2
.
Problem 49.
Let a, b be real numbers such that a < b. Consider the Hilbert space L
2
[a, b] over
R and the operator T : L
2
[a, b] R dened by
Tf =
_
b
a
f(x)dx, f L
2
[a, b].
(a) Show that T is bounded. Compute |T|.
(b) According to the Rieszs Theorem, there exists a function g L
2
[a, b] such
that
Tf = f, g) for all f L
2
[a, b].
Find such a function g and verify that |g|
L
2 = |T|.
Solution.
(a) By Holders inequality, we have
[Tf[ =
_
b
a
f(x)dx
_
b
a
[1.f(x)[dx
__
b
a
1
2
dx
_
1/2
__
b
a
[f(x)[
2
_
1/2
=
b a |f|
L
2.
Hence, T is bounded, i.e., T B(L
2
, R) = (L
2
)
, the dual space of L

2
.
From the above we get
|T|
(L
2
)

b a.
Now consider the function h(x) =
1
ba
, x (a, b). It is obviously that h L
2
(a, b)
and
|h|
L
2 =
__
b
a
(b a)
1
_
1/2
= 1 and [Th[ =
_
b
a
1
b a
dx
b a.
Hence,
b a = [Th[ |T|
(L
2
)
|h|
L
2 = |T|
(L
2
)
.
www.MATHVN.com
Therefore,
|T|
(L
2
)
=
b a.
(b) By the Rieszs Theorem, there exists a function g L
2
[a, b] such that Tf =
f, g) for all f L
2
[a, b]. Here, functions are real-valued, so we can write
Tf = f, g)
_
b
a
f(x)dx =
_
b
a
f(x)g(x)dx.
It is evident that the above equation is satised for all f L
2
[a, b] if we choose
g(x) = 1 on [a, b]. By the uniqueness of this representation guaranteed by Rieszs
Theorem, we can denitely conclude that
g(x) = 1, x [a, b].
Also, we can verify that
|g|
L
2 =
__
b
a
1
2
dx
_
1/2
=
b a = |T|
(L
2
)
.
Problem 50.
Let X, Y be normed spaces and T B(X, Y ). Consider the following statement:
T is an isometry |T| = 1.
Do you agree with it? Why?
Solution.
The direct way () is correct. Indeed, suppose that T is an isometry; then
|T|
B(X,Y )
= sup
xX
|x|
X
=1
|Tx|
Y
= sup
xX
|x|
X
=1
|x|
X
(since |Tx|
Y
= |x|
X
)
= 1.
The other way () is false. The left shift on
2
is a counter-example (see Problem
40).
www.MATHVN.com
Problem 51.
Dene T : C[0, 1] C[0, 1] by
(Tx)(t) = t
_
t
0
x(s)ds.
(a) Prove that T is a bounded linear operator. Compute |T|.
(b) Prove that the inverse T
1
: Image(T) C[0, 1] exists but not bounded.
Solution.
(a) For all x
1
, x
2
C[0, 1] and all
1
,
2
R and all t [0, 1] we have
T(
1
x
1
+
2
x
2
)(t) = t
_
t
0
(
1
x
1
+
2
x
2
)(s)ds
=
1
t
_
t
0
x
1
(s)ds +
2
t
_
t
0
x
2
(s)ds
=
1
T(x
1
)(t) +
2
T(x
2
)(t)
= [
1
T(x
1
) +
2
T(x
2
)] (t).
Hence,
T(
1
x
1
+
2
x
2
) =
1
T(x
1
) +
2
T(x
2
).
This shows that T is linear.
For each x C[0, 1] we have
|Tx| = max
t[0,1]
t
_
t
0
x(s)ds
max
t[0,1]
[t[
_
t
0
[x(s)[ds
max
t[0,1]
t
_
t
0
|x|ds = max
t[0,1]
t
2
|x| = |x|.
Hence T is a bounded with |T| 1. Moreover, if x(t) = 1, t [0, 1] then |x| = 1
and (Tx)(t) = t
_
t
0
ds = t
2
, therefore, |Tx| = max
t[0,1]
t
2
= 1. Thus,
1 = |Tx| |T||x| and so |T| 1.
Hence we have proved that |T| 1 and |T| 1, so |T| = 1.
(b) Suppose that x C[0, 1] satises
Tx = 0, i.e., t
_
t
0
x(s)ds = 0, t [0, 1].
www.MATHVN.com
It follows that
_
t
0
x(s)ds = 0, t (0, 1]. By dierentiation w.r.t. t we get x(t) =
0, t (0, 1]. Since x(t) is continuous, we then also have x(0) = 0. Hence, x = 0.
We have thus proved
x C[0, 1], T(x) = 0 x = 0.
Hence, T
1
exists.
Now we show that T
1
is not bounded. Given any n N, we let x
n
(t) = t
n
. Then
x
n
C[0, 1] and |x
n
| = max
t[0,1]
[t
n
[ = 1.
We let
y
n
(t) = Tx
n
(t) = t
_
t
0
s
n
ds =
1
n + 1
t
n+2
.
Then
|y
n
| = max
t[0,1]
1
n + 1
t
n+2
=
1
n + 1
.
Also by construction, y
n
Image(T) and T
1
y
n
= x
n
; thus
|T
1
y
n
| = |x
n
| = 1.
This shows that T
1
cannot be bounded. (For if T
1
were bounded, then we would
have
|T
1
y
n
| |T
1
||y
n
|, i.e., 1 |T
1
|
1
n + 1
, n N.
This is impossible.
Problem 53.
Let 1 < p < and q be its exponent conjugate. For f L
p
(0, ), let
(Tf)(x) =
1
x
_
x
0
f(t)dt =
_
1
0
f(tx)dt.
Show that
(a) Tf is well-dened on (0, ), and that Tf is continuous with respect to x.
(b) Tf L
p
(0, ).
(c) T is a bounded linear operator from L
p
(0, ) to itself. Calculate its norm.
(Hint: Use f
n
= x
1/p
1xn
.)
www.MATHVN.com
Solution.
(a) For f L
p
(0, ), we need to show : x
_
x
0
f(t)dt is well-dened and
continuous w.r.t. x. By Holders inequality we have
_
x
0
[f(t)[dt [x[
1/q
|f|
p
< ,
which shows that is well-dened. Also, for x, y (0, ), we have
(x) (y)[ =
_
y
x
f(t)dt
_
y
x
[f(t)[dt [x y[
1/q
|f|
p
.
This shows that is continuous.
Since (Tf)(x) =
(x)
x
, Tf is also continuous on (0, ). Moreover we have
[(Tf)(x)[ x
1/p
|f|
p
0 as x .
So (Tf)(x) 0 as x .
(b) To prove this part (b) we will use the following theorem:
Given a -nite measure space (X, A, ). Let 1 < p < and q be its exponent
conjugate. If [[ < a.e on X and if
_
X
d exists in C for every L
q
(X),
then f L
p
(X).
Now, for any g L
q
(0, ), by Fubinis theorem we have
_
(0,)
[(Tf)(x)g(x)[dx
_
(0,)
_
1
0
[(Tf)(x)[[g(x)[dtdx
=
_
(0,)
_
1
0
1
t
[f(x)[[g(x/t)[dtdx
|f|
p
_
1
0
|g(./t)|
q
dt
t
.
But
|g(./t)|
q
=
__
(0,)
[g(x/t)[
q
dx
_
1/q
= t
1/q
|g|
q
.
Thus,
_
(0,)
[(Tf)(x)g(x)[dx |f|
p
|g|
q
_
1
0
t
1/q1
dt = q|f|
p
|g|
q
< .
Hence, Tf L
p
(0, ).
(c) By the last inequality we have
|Tf|
p
q|f|
p
.
www.MATHVN.com
This shows that T is an bounded operator on L
p
(0, ) with |T|
p,p
q. We would
like to establish the equality. Consider the function f
n
, n N dened by f
n
=
x
1/p
1xn
. We have
|f
n
|
p
p
=
_
n
1
_
x
1/p
_
p
dx = ln n.
For x > 1 we have
(Tf
n
)(x) =
1
x
_
minx,n
1
t
1/p
dt = q
minx, n
1/q
1
x
.
For 0 < < 1, since (1 )
p
1 p we have
_
(0,)
[(Tf
n
)(x)]
p
dx = q
p
__
n
1
(1 x
1/q
)
p
1
x
dx + (n
1/q
1)
p
_

n
x
p
dx
_
q
p
_
ln n
p
q
+
1
p 1
_
q
p
_
|f
n
|
p
p
p
q
+
1
p 1
_
.
This implies that
lim
n
|Tf
n
|
p
|f
n
|
p
q.
Hence, |T|
p,p
q. Finally, we get
|T|
p,p
= q.
Problem 54.
Let (c
j
)
j=1
be a sequence of complex numbers. Dene an operator D on
2
by
Dx = (c
1
x
1
, c
2
x
2
, ...) for x = (x
1
, x
2
, ...)
2
.
Prove that D is bounded if and only if (c
j
)
j=1
is bounded, and in this case |D| =
sup
j
[c
j
[.
Solution.
www.MATHVN.com
Suppose (c
j
)
j=1
is bounded. Let M = sup
j
[c
j
[ < . Then
|Dx| =
_

j=1
[c
j
x
j
[
2
_
1/2
=
_

j=1
[c
j
[
2
[x
j
[
2
_
1/2
j=1
M
2
[x
j
[
2
_
1/2
= M
_

j=1
[x
j
[
2
_
1/2
= M|x|.
Hence, D is bounded and |D| M.
Suppose D is bounded. We want to show that (c
j
)
j=1
is bounded. Consider the
vector e
j
= (0, 0, .., 0, 1, 0, ...) where the number 1 appears at the j-th coordinate.
Clearly |e
j
| = 1 and |De
j
| = [c
j
[ for all j = 1, 2, .. Since D is bounded,
[c
j
[ = |De
j
| |D| for any j = 1, 2, ...
Hence (c
j
)
j=1
is bounded and M = sup
j
[c
j
[ |D|. Finally
|D| = M.
Problem 55.
Prove that B(F, Y ) is not a Banach space if Y is not complete.
Hint: Take a Cauchy sequence (y
n
) in Y which does not converge and consider the sequence
of operators (B
n
) dened by :
B
n
:= y
n
; F.
Solution.
We follow the suggestion above. It is easy to see that
B
n
B(F, Y ) and |B
n
| = |y
n
|, n N.
Since (B
n
B
m
) = (y
n
y
m
), we have
|B
n
B
m
| = |y
n
y
m
|, n N.
Therefore (B
n
) is a Cauchy sequence in B(F, Y ). Suppose there exists B B(F, Y )
such that |B
n
B| 0 as n . Let y := B1 Y , where 1 is the unit element
in F. Then
|y
n
y| = |B
n
1 B1| |B
n
B| 0 as n ,
www.MATHVN.com
i.e., the sequence (y
n
) converges to y. This contradiction proves that (B
n
) cannot
be convergent. Hence, B(F, Y ) is not a Banach space.
Problem 56.
Let T
1
, T
2
, ... be the following bounded linear operators
1
:
T
1
(x
1
, x
2
, x
3
, ...) = (x
1
, x
1
, x
1
, x
1
, ...)
T
2
(x
1
, x
2
, x
3
, ...) = (x
1
, x
2
, x
2
, x
2
, ...)
T
3
(x
1
, x
2
, x
3
, ...) = (x
1
, x
2
, x
3
, x
3
, ...)...etc
Prove that the sequence (T
n
) is strongly operator convergent. Also prove that
(T
n
) is not uniformly operator convergent.
Solution.
Let T :
1
be the bounded linear operator given by

T(x
1
, x
2
, x
3
, ...) = (x
1
, x
2
, x
3
, ...).
T is obviously linear. It is bounded with norm|T| 1. Indeed, if x = (x
1
, x
2
, x
3
, ...)
1
then
k=1
[x
k
[ = |x|
1
< , so [x
k
[ |x|
1
, k = 1, 2, ...
Hence,
|x|
|x|
1
.
We claim that (T
n
) strongly operator converges to T.
For any x = (x
1
, x
2
, x
3
, ...)
1
,
|T
n
x Tx|
= |(0, ..0, x
n
x
n+1
, x
n
x
n+2
, ...)|
= sup
j1
[x
n
x
n+j
[.
Since x = (x
1
, x
2
, x
3
, ...)
1
,
k=1
[x
k
[ < , so we have lim
k
[x
k
[ = 0. Hence,
given any > 0, there is some K N such that
[x
k
[ , for all k K.
Then if n K we have
n +j K, j N.
www.MATHVN.com
Hence,
[x
n
x
n+j
[ [x
n
[ +[x
n+j
[ 2.
Thus, for n K,
|T
n
x Tx|
= sup
j1
[x
n
x
n+j
[ 2.
In other words,
lim
n
|T
n
x Tx|
= 0.
This is true for every x
1
. Hence, the sequence (T
n
) is strongly operator conver-
gent to T.
It follows from this that if (T
n
) would be uniformly operator convergent, then the
limit must be equal to T. We show that this is not true. Now we have
(T
n
T)(x) = (0, .., 0, x
n
x
n+1
, x
n
x
n+2
, ...).
In particular, if x = (0, .., 0, 1, 0, ..) with 1 at the n-th position, then
(T
n
T)(x) = (0, .., 0, 1, 1, ...).
Here |x|
1
= 1 and |(0, .., 0, 1, 1, ...)|
= 1. Hence,
|T
n
T| 1, for all n.
Thus, (T
n
) is not uniformly operator convergent.
Problem 57.
Let H
1
and H
2
be two Hilbert spaces. Let a
1
, ..., a
n
be an orthonormal system of
H
1
and b
1
, ...,
n
be an orthonormal system of H
2
, and
1
, ...,
n
K. Consider
the operator
U : H
1
H
2
dened by U(x) =
i=1
i
b
i
x, a
i
).
Calculate |U|.
Solution.
www.MATHVN.com
From the Pythagoras theorem and the Bessel inequality we have
|U(x)|
2
=
n
i=1
[
i
[
2
|b
i
|
2
[x, a
i
)[
2
=
n
i=1
[
i
[
2
[x, a
i
)[
2
M
2
n
i=1
[x, a
i
)[
2
M
2
|x|
2
,
where M = max
1in
[
i
[. Hence, we have that
|U(x)| M|x|, x H.
Therefore, |U| max
1in
[
i
[.
On the other hand, we have
|U(a
i
)| |U| |a
i
| = |U|,
U(a
i
) =
i
b
i
, i 1, ..., n.
It follows that [
i
[ |U|, i 1, ..., n. This implies max
1in
[
i
[ |U|. Thus,
|U| = max
1in
[
i
[.
Problem 58.
(a) Let X be a Hilbert space. Let a, b H be two non-zero orthogonal elements.
Consider the operator
U : H H, U(x) = ax, b) +bx, a).
Calculate |U|.
(b) Consider the operator T : L
2
[0, ] L
2
[0, ] dened by
(Tf)(x) = sin x
_

0
f(t) cos tdt + cos x
_

0
f(t) sin tdt.
Calculate |T|.
Solution.
(a) Note that ax, b) and bx, a) are two orthogonal vectors. Using the Pythagoras
theorem we have
|U(x)|
2
= |a|
2
[x, b)[
2
+|b|
2
[x, a)[
2
.
www.MATHVN.com
From the Bessel inequality we deduce that
|U(x)|
2
= |a|
2
|b|
2
_
_
x,
b
|b|
_
2
+
_
x,
a
|a|
_
2
_
|a|
2
|b|
2
|x|
2
.
Hence, |U| |a| |b|.
But
U(a) = aa, b) +ba, a) = |a|
2
b.
Therefore,
|U(a)| = |a|
2
|b| |U| |a|.
Hence, |U| |a| |b|. Thus,
|U| = |a| |b|.
(b) Let H = L
2
[0, ]. Then H, with the usual inner product, is a Hilbert space.
The two vectors a = sin x and b = cos x are orthogonal. Indeed,
a, b) = sin x, cos x) =
_

0
sin x cos xdx = 0.
By (a) we get
|T| = | sin x| | cos x|.
But
| sin x|
2
=
_

0
(sin x)
2
dx =

2
; | cos x|
2
=
_

0
(cos x)
2
dx =

2
.
Thus,
|T| =

2
.
Problem 59.
(a) Let H be a Hilbert space and e
1
, e
2
H an orthonormal system. Let
A =
_
a b
c d
_
be scalar square matrix. Consider the operators U, V : H H dened by
U(x) = ax, e
1
)e
1
+bx, e
2
)e
2
,
V (x) = cx, e
1
)e
1
+dx, e
2
)e
2
.
www.MATHVN.com
Prove that
|U +V |
2
+|U V |
2
= 2(|U|
2
+|V |
2
)
if and only if
(max[a+c[,[b+d[)
2
+(max[ac[,[bd[)
2
=2
(
max[a[,[b[
2
+max[c[,[d[
2
)
.
(b) Prove that if dimH 2 then B(H) is not a Hilbert space
Solution.
(a) Using Problem 57 we have
|U| = max[a[, [b[; |V | = max[c[, [d[.
|U +V | = max[a +c[, [b +d[; |U V | = max[a c[, [b d[.
From here we obtain the statement.
(b) Since dimH 2 we can nd x, y H, two linearly independent vectors. Using
the Gram-Schmidt procedure on x, y we can construct an orthonormal system
e
1
, e
2
H. Now construct U, V using the matrix
A =
_
2 1
0 2
_
.
That is U, V : H H dened by
U(x) = 2x, e
1
)e
1
+x, e
2
)e
2
, V (x) = 2x, e
2
)e
2
.
If B(H) were a Hilbert space, then by the parallelogram law we must have
|U +V |
2
+|U V |
2
= 2(|U|
2
+|V |
2
),
i,e., by (a)
(max2, 3[)
2
+ (max2, 1)
2
= 2
_
max2, 1
2
+ max0, 2
2
_
.
This is not true.
www.MATHVN.com
4.2. LINEAR FUNCTIONALS 63
4.2 Linear Functionals
Review:
Denition 1 Let X be a linear space. A map p : X R is called sublinear if it is subadditive
and positive homogeneous, i.e.,
p(x +y) p(x) +p(y), x, y X.
p(x) = p(x), x X, 0.
Denition 2 Let X be a linear space. A map p : X K is called seminorm if it is subadditive
and absolutely homogeneous, i.e.,
p(x +y) p(x) +p(y), x, y X.
p(x) = [[p(x), x X, K.
The Hahn-Banach theorems
Theorem 1 (The case K = R). Let X be a real linear space, G X a linear subspace, p : X R
a sublinear functional and f : G R a linear functional such that f(x) p(x), x G. Then
there is a linear functional

f : X R which extends f, i.e.,
f(x) = f(x), x G and

f(x) p(x), x X.
Theorem 2 (The case K = R or C). Let X be a linear space, G X a linear subspace,
p : X R a seminorm and f : G K a linear functional such that [f(x)[ p(x), x G. Then
there is a linear functional

f : X K which extends f such that
[
f(x)[ p(x), x X.
Theorem 3 (The normed space case). Let X be a linear space, G X a linear subspace and
f : G K a linear and continuous functional. Then there is

f : X K, a linear and continuous
functional which extends f such that |
f| = |f|. Such an

f is called a Hahn-Banach extension for
f.
Theorem 4 Let X be a normed space, x
0
X and G X a linear subspace such that =
d(x
0
, G) > 0. Then there is f : X K, a linear and continuous functional, such that
f = 0 on G, f(x
0
) = 1 and |f| =
1
.
www.MATHVN.com
Three classic problems.
Problem 60.
The dual space of
1
is
, that is, (
1
)
.
Solution.
Let (e
k
) be the standard basis for
1
, where e
k
=
kj
. Every x
1
has unique
representation
x =
k=1
k
e
k
.
Norms on
1
and on
are respectively:
|x|
1
=
k=1
[
k
[, |x|
= sup
kN
[
k
[.
Take f (
1
)
, that is, f B(
1
, F). For x
1
,
f(x) = f
_

k=1
k
e
k
_
=
k=1
k
where
k
= f(e
k
).
Then for all k N,
[
k
[ = [f(e
k
)[ |f| |e
k
| = |f|.
So,
() sup
kN
[
k
[ |f|.
On the other hand, for every b = (
k
)
we can obtain a corresponding bounded

linear functional g on
1
. In fact, we may dene g on
1
by
g(x) =
k=1
k
where x = (
k
)
1
.
Then g :
1
F is linear, and the boundedness follows from
[g(x)[
k=1
[
k
k
[ sup
kN
[
k
[
k=1
[
k
[ = sup
kN
[
k
[ |x|
1
.
www.MATHVN.com
Hence g (
1
)
.
We nally show that the norm of f is the norm on
. We have
[f(x)[ =
k=1
sup
kN
[
k
[
k=1
[
k
[ = |x| sup
kN
[
k
[.
Taking the supremum over x such that |x| = 1, we see that
|f| sup
kN
[
k
[ = |(
k
)|
.
From() and this, we obtain
|f| = |(
k
)|
.
Thus there is an isometric isomorphism between (
1
)
and
, so that we can write

(
1
)
.
Problem 61.
The dual space of
p
is
q
, that is, (
p
)
=
q
, here 1 < p < and
1
p
+
1
q
= 1.
Solution.
The basis for
p
is (e
k
), where e
k
=
kj
as in the previous problem. Every x
p
has unique representation
x =
k=1
k
e
k
.
Take f (
p
)
, that is, f B(
p
, F). For x
p
,
(1) f(x) =
k=1
k
where
k
= f(e
k
).
Let q be the conjugate of p. Consider x
n
= (
(n)
k
) with
(n)
k
=
_
[
k
[
q
k
if 1 k n and
k
,= 0
0 if k > n or
k
= 0.
By substituting this into (1) we obtain
f(x
n
) =
k=1
(n)
k

k
=
n
k=1
[
k
[
q
.
www.MATHVN.com
We also have
f(x
n
) |f| |x
n
| = |f|
_
n
k=1
[
(n)
k
[
p
_1
p
= |f|
_
n
k=1
[
k
[
(q1)p
_1
p
= |f|
_
n
k=1
[
k
[
q
_1
p
.
Together,
f(x
n
) =
n
k=1
[
k
[
q
|f|
_
n
k=1
[
k
[
q
_1
p
.
Dividing the last factor and using 1
1
p
=
1
q
, we get
_
n
k=1
[
k
[
q
_
1
1
p
=
_
n
k=1
[
k
[
q
_1
q
|f|.
Since n is arbitrary, letting n , we obtain
(2)
_

k=1
[
k
[
q
_1
q
|f|.
Conversely, for any b = (
k
)
q
we can get a corresponding bounded linear func-
tional g on
p
. In fact, we may dene g on
p
by setting
g(x) =
k=1
k
where x = (
k
)
p
.
Then g is linear. The boundedness follows from the Holder inequality. Hence g
(
p
)
. We nally prove that the norm of f is the norm of (

k
) in
q
. From (1) and
the Holder inequality we have
[f(x)[
k=1
k=1
[
k
[
p
_1
p
_

k=1
[
k
[
q
_1
q
= |x|
_

k=1
[
k
[
q
_1
q
.
www.MATHVN.com
Hence
|f|
_

k=1
[
k
[
q
_1
q
.
From (2) we obtain
|f| =
_

k=1
[
k
[
q
_1
q
.
This can be written
|f| = |(
k
)|
q
where
k
= f(e
k
).
Thus there is an isometric isomorphism between (
p
)
and
q
.
Problem 62.
The dual space of c
0
is
1
, that is, (c
0
)
=
1
.
Solution.
Recall that c
0

, and if x = (
n
) c
0
the norm of x in c
0
is |x| = sup
nN
[
n
[.
Let x = (
n
) c
0
and (e
k
), where e
k
=
kj
, be the basis for
as in preceding
examples. Then x has unique representation
x =
k=1
k
e
k
, and lim
k
k
= 0.
Consider any f (c
0
)
. Since f is linear,
f(x) =
k=1
k
where
k
= f(e
k
).
For a given N N, take a special sequence x
N
0
= (
(0)
k
) c
0
where
(0)
k
=
_

k
[
k
[
if 1 k N and
k
,= 0
0 if k > N or
k
= 0.
Note that |x
N
0
| = 1, and we have
[f(x
N
0
)[ =
N
k=1
[f(e
k
)[ |f| |x
N
0
| = |f| < .
www.MATHVN.com
Since N is arbitrary,
k=1
[f(e
k
)[ =
k=1
[
k
[ = |f| < .
This shows that (
k
)
1
and
(i) |(
k
)| |f|.
On the other hand,
[f(x)[ =
k=1
sup
kN
[
k
[
k=1
[
k
[ = |x| |(
k
)| < .
It follows that
(ii) sup
|x|=1
[f(x)[ = |f| |(
k
)|.
From (i) and (ii) we obtain |f| = |(
k
)|.
Now given (
k
)
1
, we want to construct a linear bounded functional g on c
0
. Let
x = (
k
) c
0
. Consider the function g : c
0
F dened by
g(x) =
k=1
k
.
It is clear that g is linear. Its boundedness follows from
[g(x)[ sup
kN
[
k
[
k=1
[
k
[ = |x|
k=1
[
k
[ < .
Hence g (c
0
)
. Thus there is an isometric isomorphism between (c

0
)
and
1
, so
we can write (c
0
)
=
1
.

Problem 63.
Let X be a normed space and f, g are nonzero linear functionals on X. Show
that
ker(f) = ker(g) f = cg for some nonzero scalar c.
www.MATHVN.com
Solution.
The reverse way () is trivial.
We show the direct way. Suppose ker(f) = ker(g). Since f ,= 0, there exists some
x
0
X such that f(x
0
) ,= 0, and by rescaling, we can assume that f(x
0
) = 1. Since
x
0
/ ker(f) = ker(g), we have g(x
0
) ,= 0. Given any y X, we have
f
_
y f(y)x
0
_
= f(y) f(y)f(x
0
) = 0.
Therefore, y f(y)x
0
ker(f) = ker(g). Hence,
g(y) f(y)g(x
0
) = g
_
y f(y)x
0
_
= 0.
This implies that
g(y) = g(x
0
)f(y), y X.
Hence g = cf with c = g(x
0
) ,= 0.
Problem 64.
Let X be a normed space and f are nonzero linear functional on X. Show that
f is continuous if and only if ker(f) is closed.
Solution.
Suppose that f is continuous. Since ker(f) = f
1
(0), so ker(f) is closed.
Conversely, suppose that ker(f) is closed. Pick an x
0
X such that f(x
0
) = 1.
Assume that f is not continuous, that is, f is not bounded. Then there exists a
sequence (x
n
) in X such that
|x
n
| = 1 and f(x
n
) n, n N.
Dene y
n
= x
0
xn
f(x
n
)
. Then
y
n
ker(f) for all n and lim
n
y
n
= x
0
.
By hypothesis, ker(f) is closed, so x
0
ker(f), that is, f(x
0
) = 0. This contradicts
our assumption that f(x
0
) = 1. Thus f is continuous.
Problem 65
Let Z be a subspace of a normed space X, and y X. Let d = d(y, Z). Prove
that there exists X
such that || 1, (y) = d and (z) = 0 for all

z Z.
www.MATHVN.com
Solution.
If y Z, then d = 0 and so the zero functional works, so we may assume that y / Z.
Consider the subspace Y = Cy + Z X. Since y / Z, so for every x Y , there is
a unique C and z Z such that x = y +z. Dene
(x) = d.
We observe that (z) = 0 since
x = z y = 0
= 0 (otherwise y = 0 so y Z).
Also we have (y) = d since y = 1.y + 0. It is clearly that is linear in x and
|x| = [[|y +
1
z| d = [(x)[.
Thus, is continuous, that is, Y
and || 1. By the Hahn-Banach theorem

we may extend to an element X
with the same norm.

Problem 66
Let X be a normed space and (x
n
) be a sequence in X. Set V := Span x
1
, x
2
, ....
Let W be the set of all continuous f X
such that f(x

n
) = 0, n N. Prove
that
V = x X : f(x) = 0, f W.
Solution.
Let Y := x X : f(x) = 0, f W.
We show that V Y . Take any x
0
V . There exists a sequence (u
k
) in V such
that u
k
x
0
as k . Since V := Span x
1
, x
2
, ..., for each k N, there are
scalars c
(k)
1
, ..., c
(k)
n
k
such that u
k
=
n
k
n=1
c
(k)
n
x
n
. By linearity of f and by denition
of W, we have
f(u
k
) =
n
k
n=1
c
(k)
n
f(x
n
) = 0, f W.
By continuity of f, we have
f(x
0
) = 0, f W.
www.MATHVN.com
Hence x
0
Y , and so V Y .
Now given any x
0
Y , we want to show x
0
V . Assume that x
0
/ V . Then
d(x
0
, V ) = d > 0. Using the result in Problem 65, there is an F X
such that
F(x
0
) = d and F(x) = 0, x V .
It follows that
F(x
0
) = d and F(x
n
) = 0, n N.
Hence F W but F(x
0
) ,= 0. It means that x
0
/ Y : a contradiction. Thus x
0
V ,
and so Y V .
The proof is complete.
Problem 67.
Let X be a normed space and x
1
, ..., x
n
X a linearly independent system.
Prove that for any
1
, ...,
n
K there exists x
such that
x
(x
i
) =
i
, i 1, ..., n.
Solution.
Let Y = Spanx
1
, ..., x
n
. Since the system x
1
, ..., x
n
X a linearly independent,
it follows that x
1
, ..., x
n
is an algebraic basis for Y . Dene f : Y K linear with
f(x
i
) =
i
, i 1, ..., n. Since Y is of nite dimension, f is continuous. By the
Hahn-Banach theorem there is an extension x
for f. Then
x
(x
i
) = f(x
i
) =
i
, i 1, ..., n.
Problem 68.
Let X be a normed space and Y X a linear subspace. For x
0
X Y we
dene
f : SpanY, x
0
K, f(y +x
0
) = , y Y, K.
(a) Prove that f is well dened and linear.
(b) Prove that f is continuous if and only if x
0
/ Y .
(c) Prove that Y =
ker x
: x
, Y ker x
.
www.MATHVN.com
Solution.
(a) Let us rst observe that
SpanY, x
0
= y +x
0
: y Y, K.
To show that f is well dened, we show that
(y +x
0
= y
t
+
t
x
0
) (y = y
t
, =
t
).
Assume that ,=
t
. Then
x
0
=
y y
t

t
Y
since y, y
t
Y and Y is a linear subspace. This is a contradiction. Thus, =
t
implies that y = y
t
. The linearity of f is obvious.
(b) Suppose that f is continuous. We will prove that x
0
/ Y . Assume, for a
contradiction, that x
0
Y . Then there is a sequence (y
n
) Y such that y
n
x
0
.
Since y
n
, x
0
Y , y
n
x
0
and f is continuous, f(y
n
) f(x
0
). By denition of
f, f(y
n
) = 0, n N. It follows that f(x
0
) = 0, which is false, since by denition
of f, we have f(x
0
) = 1.
Suppose now that that x
0
/ Y , i.e., x
0
(Y )
c
, which is open. Then
> 0 : B(x
0
; ) (Y )
c
.
The inclusion is equivalent to Y B(x
0
; ) = . Let x = y + x
0
SpanY, x
0
. If
= 0, then
() [f(x)[ = 0
|x|
.
If ,= 0, then
y
Y so
y
/ B(x
0
; ), i.e.,
_
_
x
0
_
_
. Therefore,
() |x| = [[ |
y
x
0
| [[ = [f(x)[.
Both cases (*) and (**) show that f is continuous.
(c) If x
and Y ker x
then Y ker x
(since ker x
is closed) and therefore

Y
ker x
: x
, Y ker x
.
Conversely, let x
0

ker x
: x
, Y ker x
, and assume that x

0
/ Y .
By (b), there is an f : SpanY, x
0
K linear and continuous such that f[
Y
=
0, f(x
0
) = 1. Let x
be a Hahn-Banach extension for f given by the Hahn-

Banach theorem. Then x
= f on SpanY, x
0
. In particular, x
= 0 on Y . Hence
Y ker x
and x
(x) = f(x
0
) = 1, that is x / ker x
and so
x
0
/
ker x
: x
, Y ker x
,
which is a contradiction.
www.MATHVN.com
Chapter 5
Fundamental Theorems
Review some main points:
1. The Baire category theorem:
Denition 3 Let X be a metric space and E be a subset of X.
(a) E is called a set of the rst category if E is a countable union of nowhere dense (non-dense)
sets in X.
(b) If E is not a set of the rst category, then E is called a set of the second category.
Theorem 5 (Baire category theorem - form I)
If X is a complete metric space, then X is a set of the second category.
Let X be a complete metric space. The Baire category theorem tells us that if X =

n=1
A
n
, then
some of the set A
n
must have non-empty interior.
Theorem 6 (Baire category theorem - form II)
Let X be a complete metric space. If (U
n
)
nN
is a sequence of open dense subsets of X, then
nN
U
n
is dense in X.
It can be shown that these two forms of Baire category theorem are equivalent (try it!).
2. The principle of uniform boundedness (P.U.B.)
Theorem 7 Let X be a Banach space and Y be a normed space. Let T be a family of bounded
linear operators from X to Y . Suppose that for each x X, |Tx| : T T is bounded, then
|T| : T T is bounded.
One of the P.U.B. consequences is:
Theorem 8 (Banach-Steinhaus theorem)
Let X be a Banach space and Y be a normed space and (T
n
) be a sequence in B(X, Y ). Suppose
for every x X, T
n
x Tx as n . Then the family |T
n
|, n N is bounded, i.e.,
sup
nN
|T
n
| < , and T is linear and bounded, i.e., T B(X, Y ).
73
www.MATHVN.com
74 CHAPTER 5. FUNDAMENTAL THEOREMS
3. The open mapping theorem
Theorem 9 Let X and Y be Banach spaces and T B(X, Y ). If T is onto (surjective), then T
is an open mapping, (that is, if U is open in X, then T(U) is open in Y ).
Consequence:
Theorem 10 (The inverse mapping theorem)
Let X and Y be Banach spaces. If T B(X, Y ) is bijective, then T
1
exists and T
1
B(Y, X).
5. The closed graph theorem
Theorem 11 Let X and Y be Banach spaces and T be a linear map from X to Y . If we dene
the graph of T by
(T) := (x, y) X Y : y = Tx,
then T is bounded (continuous) if and only if (T) is closed in X Y .
Problem 69.
(a) If X is a normed space, prove that any proper closed linear subspace of X is
a nowhere dense set.
(b) Prove that c
0
is a nowhere dense set of c.
Solution.
(a) Let A be a proper closed linear subspace of X. If A is not nowhere dense, then
A =

A ,= since A is closed. Therefore, A contain an open ball B(a; ). Take
x B(0; ) then we have
|x| < a +x B(a; ) A.
Since a, a + x A and A is a linear space, x must be in A. Therefore B(0; ) A.
Now take any y X with y ,= 0, then we have
2|y|
y B(0; ).
This implies that

2|y|
y A, and so y A. Thus X = A. This is a contradiction.
(b) We know that c
0
is a proper closed linear subspace of c. So c
0
is nowhere dense
in c.
www.MATHVN.com
75
Problem 70.
Show that L
2
[0, 1] is a subset of the rst category in L
1
[0, 1].
Solution.
For short, we write L
1
:= L
1
[0, 1] and L
2
:= L
2
[0, 1]. We know that L
2
L
1
. For
each n N, let
A
n
= f : |f|
2
n.
Every function in L
2
will be in some A
n
, Thus
L
2
=
_
nN
A
n
.
We rst prove that A
n
is closed. Fix n. Let (f
k
) be a sequence in A
n
such that
|f
k
f|
1
0. We show that f A
n
. Since |f
k
f|
1
0, (f
k
) converges
in measure to f. Hence there exists a subsequence (f
k
j
) converging to f almost
everywhere. By denition of A
n
we have
_
1
0
f
2
k
j
(t)dt n
2
for j = 1, 2, ...
Applying Fatous lemma, we obtain
_
1
0
f(t)dt n
2
.
Thus f A
n
, and A
n
is closed.
Now we show that A
n
is nowhere dense, i.e.
_
A
n
_
0
= (A
n
)
0
= . To do this, it
suces to show that for any open ball B
(f) in L
1
, there exists a point g B
(f)
but g / A
n
. Take
h(t) =

4
1
t
.
Then h / L
2
and
_
1
0
h(t)dt =

2
,
so that
h L
1
and |h|
1
=

2
.
Let g = f +h. Then g / L
2
so g / A
n
, and
|f g|
1
= |h|
1
=

2
.
www.MATHVN.com
Hence g B
(f). Thus A
n
is nowhere dense, and L
2
is of the rst category in
L
1
.
Problem 71.
Let X ,= 0 be a normed space, and A X which is not nowhere dense. We
denote by A
t
the derivative set of A, i.e., the set of all accumulation points of
A. Prove that
x X and > 0 such that B(x; ) A
t
.
(Note:

A = A A
t
).
Solution.
Since

A ,= ,
x X and > 0 such that B(x; )

A.
We will prove B(x; ) A
t
. Let y B(x; ). If y / A
t
, then there is r > 0 such that
B(y; r) (A y) = .
Let = minr,
1
2
( |y x|) > 0. We claim B(y; ) B(x; ). Indeed, if
z B(y; ) then
|z x| |z y| +|y x|
< +|y x|
1
2
( |y x|) +|y x|
=
1
2
( +|y x|) < .
Hence z B(x; ).
Obviously, B(y; ) B(y; r). So
B(y; ) (A y) B(y; r) (A y) = .
That is B(y; ) A y. Since y B(x; )

A, so y

A, from whence
B(y; ) A ,= , and therefore
B(y; ) A = y.
www.MATHVN.com
77
Since B(y; )

A B(y; ) A and B(y; ) B(x; )

A, i.e., B(y; )

A =
B(y; ). Hence, B(y; ) = y, which is false, since X ,= 0. (Indeed, a X, a ,=
0. Take b =
a
|a|
X. Then |b| = 1 and
y +

2
b B(y; ) = y,
which implies that b = 0 : a contradiction).
Problem 72.
Let X is a Banach space and A X a dense set. Can we nd a function
f : X R such that, for every x A, we have lim
tx
[f(t)[ = ?
Solution.
Suppose that such a function exists. Since f takes nite values, for every x X,
there is k N such that [f(x)[ k, i.e.,
X =
_
kN
A
k
where A
k
= x X : [f(x)[ k.
Since X is a Banach space, X is of the second Baire category. So there is k N
such that

A
k
,= . Using Problem 74, it follows that there is x X and > 0 such
that B(x; ) A
t
k
. Since A is dense, we have
,= B(x; ) A A
t
k
A.
It follows that there is an element a A which is in A
t
k
. By denition of accumula-
tion points, there is a sequence (x
n
) A
k
with x
n
,= a, n N such that x
n
a.
Since a A, by hypothesis we have lim
ta
[f(t)[ = . Hence, lim
n
[f(x
n
)[ = .
This is not possible since (x
n
) A
k
implies that [f(x
n
)[ k, k N.
Problem 73.
Show that Q is a subset of the rst category of R.
Solution.
Since Q is countable, we have
Q =
_
nN
x
n
and (x
n
)
0
= (x
n
)
0
= .
www.MATHVN.com
Problem 74.
Show that the set of piecewise linear functions on R is of the rst category.
Solution.
Let P denote the set of piecewise linear functions on R. Let P
n
be the set of piecewise
linear functions having n intervals of linearity. We have
P =
_
n=1
P
n
.
We need to prove that P
n
is nowhere dense for every n. Fix an arbitrary n. Let
f P
n
. Consider the ball B
r
(f), r > 0. Let
g(t) =
r
2
sin(24nt) +f(t).
Then
[f(t) g(t)[ =
r
2
[ sin(24nt)[.
Hence d(f, g) =
r
2
, and so g B
r
(f). We claim that the ball B
r
2
(g) contains no
element from P
n
. Pick h B
r
(f) P
n
and suppose d(g, h) <
r
2
. Then
() d(g, h) = sup [g(t) h(t)[ = sup
r
2
sin(24nt) +f(t) h(t)
<
r
2
.
Observing that the term
r
2
sin(24nt) oscillates between
r
2
and
r
2
4n times on
[0, 1]. Thus the term f(t) h(t) must also oscillate between negative and positive
values 4n times for (*) to hold. But this is impossible since the term f(t) h(t) is
a piecewise linear function with at most 2n intervals of linearity. So, the open ball
B
r
2
(g) contains no element from P
n
. Since n is arbitrary, we see that P
n
is nowhere
dense, and hence P is of the rst category.
Problem 75. (Inverse mapping theorem)
Let X and Y be Banach spaces and T B(X, Y ). Suppose T is bijective. Show
that there exist real numbers a, b > 0 such that
a|x| |Tx| b|x|, x X.
www.MATHVN.com
79
Solution.
Since T is linear, bijective and bounded, T
1
exists, is linear and bounded by the
inverse mapping theorem. Let |T
1
| =
1
a
and |T| = b. Note that T ,= 0, a, b > 0.
Now since T is bounded,
(i) |Tx| |T| |x| = b|x|, x X.
Also, since T
1
is bounded,
(ii) |x| = |T
1
(Tx)| |T
1
| |Tx| =
1
a
|Tx|, x X.
(i) and (ii) imply that
a|x| |Tx| b|x|, x X.
Problem 76. Let X = C
1
[0, 1] be the space of continuously dierentiable func-
tions on [0, 1] and Y = C[0, 1] . The norm on C[0, 1] and C
1
[0, 1] is the sup-
norm. Consider the map
T : C
1
[0, 1] C[0, 1] dened by Tx(t) =
dx(t)
dt
, t [0, 1].
Show that the the graph of T is closed but T is not bounded. Does this contradict
the closed graph theorem ?
Solution.
It is clear that T is linear.
We show that (T) is closed. Suppose x
n
x in X = C
1
[0, 1] and Tx
n
y in
Y = C[0, 1]. We must show that y = Tx. For any t [0, 1], we have
_
t
0
y(s)ds =
_
t
0
lim
n
dx
n
ds
ds, y C[0, 1]
= lim
n
_
t
0
dx
n
ds
ds (uniform convergence)
= lim
n
_
x
n
(t) x
n
(0)
_
= x(t) x(0).
Thus, with y C[0, 1], we have
x(t) = x(0) +
_
t
0
y(s)ds, t [0, 1].
www.MATHVN.com
Hence
x C
1
[0, 1] and
dx
ds
= y on [0, 1].
That is Tx = y, and so (x, y) (T), and (T) is closed.
We show that T is not bounded. Take f
n
(t) = t
n
, n N, t [0, 1]. Then
f
n
C
1
[0, 1] and Tf
n
= nf
n1
for n > 1. But we have
|f
n
| = 1 and |Tf
n
| = n,
which shows that T is unbounded. The reason? Is C
1
[0, 1] with the sup-norm a
Banach space?
Problem 77. (P.U.B.)
Let H be a separable Hilbert space with an orthonormal basis e
n
n=1
. Let x
n
be a sequence in H. Prove that the following two statements are equivalent:

(1) lim
n
x, x
n
) = 0, x H.
(2) lim
n
e
m
, x
n
) = 0, m N and |x
n
| is bounded.
Solution.
(1) (2)
Assume that (1) is true . Then the rst part of (2) is automatically true. We
have only to show that |x
n
| is bounded. Consider, for each n, the functional
f
n
(x) = x, x
n
). This is a bounded functional by Schwarz inequality, and, because
f
n
(x) 0 for each x, we have that the set |f
n
(x)| : n N is bounded. The
principle of uniform boundedness then gives us that the set |f
n
| is bounded. But
|f
n
| = |x
n
|, so |x
n
| is bounded.
(2) (1)
Assume that (2) holds. Let B be the bound of |x
n
| and let x X. We write
x =
n
e
n
, x)e
n
. For every > 0, let K be such that
m>K
[e
m
, x)[
2
<

B
.
We know that, for m xed, e
m
, x
n
) 0 as n . So we may nd N such that if
n > N then [e
m
, x
n
)[ <

K sup[er,x)[
(the denominator is nite because the sequence
www.MATHVN.com
81
e
r
, x) is an
2
-sequence). Then, if n > N, we have
[x, y)[ =
m=1
e
m
, x)e
m
, x
n
)
m=1
[e
r
, x)[
[e
m
, x)[ +
m>K
e
m
, x)e
m
, x
n
)
+
_
m>K
[e
m
, x)[
2
_
1/2
+
_
m>K
[e
m
, x
n
)[
2
_
1/2
+

B
|x
n
| 2.
Problem 78. (Closed graph theorem)
Let H be a Hilbert space and T : H H a linear operator which is symmetric,
i.e.,
Tx, y) = x, Ty), x, y H.
Prove that T is continuous.
Solution.
Since H is a complete space and T is a linear operator, it is sucient to prove that
T is a closed graph operator. Let (x
n
) be a sequence in H such that x
n
x H
and Tx
n
y H. We will show that y = Tx .
By hypothesis, we have
T(x
n
x), y) = x
n
x, Ty), y H.
Hence
T(x
n
x), y) 0, y H.
That is
T(x
n
x)
w
0, or, equivalently Tx
n
w
Tx.
But Tx
n
y, so Tx
n
w
y. Since the limit is unique, we have Tx = y. Thus, T is
a closed graph operator.
Problem 79. (Banach-Steinhaus theorem)
Let a = (a
1
, a
2
, ...) = (a
n
)
nN
be a sequence of scalars such that the sequence
(a
n
x
n
)
nN
c
0
for all sequences x = (x
n
)
nN
c
0
. Prove that a
.
www.MATHVN.com
Solution.
For every n N, consider the operator
U
n
: c
0
c
0
dened by U
n
(x) = U
n
(x
1
, x
2
, ...) = (a
1
x
1
, a
2
x
2
, ..., a
n
x
n
, 0, ...).
Then U
n
is linear, and
|U
n
(x
1
, x
2
, ...)| = |(a
1
x
1
, a
2
x
2
, ..., a
n
x
n
, 0, ...)|
= max([a
1
x
1
[, ..., [a
n
x
n
[)
|x| max([a
1
[, ..., [a
n
[), x c
0
.
So U
n
is continuous, and |U
n
| max([a
1
[, ..., [a
n
[).
For 1 k n, we have
|U
n
| |U
n
(e
k
)| = |(0, ..., 0, a
k
, 0, ...)| = [a
k
[,
where e
k
= (0, .., 0
. .
k1
, 1, 0, ...) c
0
. Therefore max([a
1
[, ..., [a
n
[) |U
n
|. Hence,
|U
n
| = max([a
1
[, ..., [a
n
[), n N.
Now consider the operator
U : c
0
c
0
dened by U(x) = (a
1
x
1
, a
1
x
2
, ...) = (a
n
x
n
)
nN
.
We have that U
n
(x) U(x), x c
0
because
|U
n
(x) U(x)| = sup
kn+1
[a
k
x
k
[ 0.
Now from the Banach-Steinhaus theorem we get that sup
nN
|U
n
| < , i.e., sup
nN
[a
n
[ <
. In other words, a
.
Problem 80. (Similar problem)
Let a = (a
1
, a
2
, ...) = (a
n
)
nN
be a sequence of scalars such that the sequence (a
n
x
n
)
nN
c
0
for
all sequences x = (x
n
)
nN

. Prove that a c
0
.
Problem 81. (P.U.B.)
Let x = (x
1
, x
2
, ...) = (x
i
)
i=1
be a sequence of scalars such that the series
i=1
x
i
y
i
is convergent for all y = (y
1
, y
2
, ...) c
0
. Prove that x
1
.
www.MATHVN.com
83
Solution.
For every n N, we dene the linear operator
T
n
: c
0
C, T
n
(y) =
n
i=1
x
i
y
i
.
Then we have
[T
n
(y)[
n
i=1
[x
i
y
i
[
_
n
i=1
[x
i
[
_
|y|
.
This shows that T
n
is bounded with
|T
n
|
n
i=1
[x
i
[.
By hypothesis,
i=1
x
i
y
i
< , the sequence
_
T
n
(y)
_
nN
converges for every y c
o
,
and c
0
is a Banach space, the principle of uniform boundedness implies that
M > 0 : |T
n
| M, n N.
Now let x
(n)
=
_
x
(n)
1
, x
(n)
2
, ..., x
(n)
n
, 0, 0, ...
_
be a truncated version of x (so that x
(n)
1
). Let
T
n
(y) =
n
i=1
x
(n)
i
y
i
, y = (y
1
, y
2
, ...) c
0
.
Dene y
(n)
=
_
y
(n)
1
, y
(n)
2
, ...,
_
by
y
(n)
k
=
_
_
_
x
(n)
k
[x
(n)
k
[
if x
(n)
k
,= 0,
0 if x
(n)
k
= 0.
Then
[T
n
(y
(n)
)[ =
n
k=1
[x
(n)
k
[ = |x
(n)
|
1
= |x
(n)
|
1
|y
(n)
|
.
Hence
|T
n
| |x
(n)
|
1
,
which in turn implies that
|x
(n)
|
1
M, n N.
www.MATHVN.com
But it is clear from the denition of x
(n)
that
_
|x
(n)
|
1
_
is an increasing sequence of
real numbers. Being bounded above by M, it must converge. Hence
i=1
[x
i
[ < ,
and so x
1
.
Problem 82. (Very similar problem)
Let c = (c
1
, c
2
, ...) = (c
i
)
i=1
be a sequence of scalars such that the series

i=1
c
i
a
i
is convergent
for all a = (a
1
, a
2
, ...)
1
. Prove that x
.
Let X, Y and Z be Banach spaces. Suppose that T : X Y is linear, that
J : Y Z is linear, bounded and injective, and that JT J T : X Z is
bounded.
Show that T is also bounded.
Solution.
We will show that the graph (T) is closed. Then by the closed graph theorem, this
implies that T is bounded (continuous).
Let
_
(x
n
, y
n
)
_
nN
be a convergent sequence in X Y , that is,
x
n
x in X, y
n
y in Y and Tx
n
= y
n
.
Since J and JT are continuous,
_
y
n
y Jy
n
Jy
and
_
x
n
x JTx
n
JTx
.
Since Tx
n
= y
n
, we have Jy
n
JTx. Since the limit is unique, this gives that
Jy = JTx. But by hypothesis J is injective, so we have
Jy = JTx y = Tx.
This shows that (x, y) (T), and (T) is closed.
www.MATHVN.com
85
Let X be a Banach space and E, F two closed subspaces of X such that X =
E F. Consider the projections on E and on F dened by
P
E
: X E, P
E
(u) = x,
P
F
: X F, P
F
(u) = y, where u = x +y, x E, y F.
Use the closed graph theorem to show that P
E
B(X, E) and P
F
B(X, F).
Solution.
The linearity of these two maps are easy to check. Let us prove that they are
bounded by using the closed graph theorem. Denote the graph of P
E
by
E
. We
can write
E
= (x, y) X E : x y F.
Let (x
n
, y
n
)
E
for every n N. Suppose (x
n
, y
n
) (x, y) as n . Since
x
n
y
n
F for every n N, and F is a closed subspace of X, lim
n
(x
n
y
n
) =
x y F. It follows that (x, y)
E
. Thus,
E
is closed, and so P
E
is bounded.
The proof for P
F
is the same.
Problem 85. (Inverse mapping theorem)
Let (X, |.|
1
) and (X, |.|
2
) be Banach spaces. Suppose that
C 0 : |x|
2
C|x|
1
, x X.
Show that the two norms |.|
1
and |.|
2
are equivalent.
Solution.
Consider the identity map
id : (X, |.|
1
) (X, |.|
2
), id(x) = x.
It is clear that the identity map is linear and bijective. It is continuous since by
hypothesis we have
() |id(x)|
2
= |x|
2
C|x|
1
, x X.
By the inverse mapping theorem, the inverse map id
1
exists and continuous. That
is
() C
t
0 : |x|
1
C
t
|x|
2
, x X.
www.MATHVN.com
() and () together imply that |.|
1
and |.|
2
are equivalent.
www.MATHVN.com
Chapter 6
Linear Operators on Hilbert
Spaces
Review.
1. Denition and elementary properties
Let T : H K be a linear operator between Hilbert spaces H and K.
The following statements are equivalent:
1. T is continuous at 0,
2. T is continuous,
3. T is bounded on H.
An isomorphism between H and K is a linear surjection U : H K such that
Ux, Uy) = x, y), x, y H.
An isomorphism is an isometry and so preserves completeness, but an isometry need not to be an
isomorphism.
Proposition 1 Two Hilbert spaces are isomorphic if and only if they have the same dimension.
2. Adjoint of an Operator
Let A B(H). Then A
is called the adjoint operator of A if

Ax, x) = x, A
x), x H.
Proposition 2 If A, B B(H) and F, then
(a) (A+B)
= A
+B
.
(b) (AB)
= B
.
(c) A
:= (A
= A.
(d) If A is invertible in B(H) and A
1
is its inverse, then A
is invertible and (A
)
1
= (A
1
)
.
87
www.MATHVN.com
88 CHAPTER 6. LINEAR OPERATORS ON HILBERT SPACES
Proposition 3 It A B(H) then
|A| = |A
| = |A
A|
1/2
.
3. Self-adjoint, normal, unitary operators
Denition 4 If A B(H), then
(a) A is Hermitian or self-adjoint if A
= A.
(b) A is normal if AA
= A
A.
(c) A is unitary if it is a surjective isometry.
Proposition 4 Let A B(H). The following statements are equivalent.
(a) A
A = AA
= I.
(b) A is unitary.
(c) A is a normal isometry.
4. Positive operators
Denition 5 Let H be a Hilbert space. An operator A B(X) is called positive if
Ax, x) 0, x X.
We write A 0. If A, B B(X) and AB 0 then we write B A.
Proposition 5 If A B(X) then A
A 0. In addition, if A 0, then
1. A is self adjoint,
2. there exists a unique B B(X) such that B 0 and B
2
= A. Furthermore, B is also self
adjoint and commutes with every bounded operator which commutes with A. We write B =
A.
We dene [A[ =
A.
5. Projection, Orthogonal projection
Denition 6
If P B(H) and P
2
= P, then P is called a projection.
If P B(H), P = P
2
and P
= P, then P is called an orthogonal projection.

Proposition 6
If P : H H is a projection then H = Image P ker P.
If H = M N, where M, N are subspaces of H, then there is a projection P : H H with
Image P = M and ker P = N.

www.MATHVN.com
89
Problem 86. Let P be an orthogonal projection dened on a Hilbert space H.
Show that |P| = 1.
Solution.
If x H and Px ,= 0, then the use of the Cauchy-Schwarz inequality implies that
|Px| =
Px, Px)
|Px|
=
x, P
2
x)
|Px|
(since P
= P)
=
x, Px)
|Px|

|x| |Px|
|Px|
|x|.
Therefore |P| 1.
Now, if P ,= 0, then there is an x
0
H such that
Px
0
,= 0 and |P(Px
0
)| = |Px
0
|.
This implies that |P| 1. Thus |P| = 1.
Problem 87. Given a function : [0, 1] C, consider the operator
P : L
2
[0, 1] L
2
[0, 1] dened by Pf(x) = (x)f(x).
Find necessary and sucient conditions on the function for P to be an or-
thogonal projection.
Solution.
First, in order for P to be a well-dened operator acting on L
2
[0, 1], the function
f needs to be in L
2
[0, 1] for all f L
2
[0, 1]. In particular f is measurable, and
taking f 1, it follows that is a measurable function on [0, 1].
Secondly, P is an orthogonal projection if and only if P
= P and P
2
= P. The
last equality is equivalent to
2
(x)f(x) = (x)f(x), f L
2
[0, 1]. Again by taking
f 1, we have a
2
(x) = a(x) for almost every x [0, 1]. Thus
a(x) = 0 or a(x) = 1 for almost all x [0, 1].
www.MATHVN.com
In particular takes real values. Then
Pf, g) =
_
1
0
Pf(x)g(x)dx
=
_
1
0
(x)f(x)g(x)dx
=
_
1
0
f(x)(x)g(x)dx
=
_
1
0
f(x)Pg(x)dx
= f, Pg),
which proves that P is self-adjoint. Since 0 (x) 1 for a.e. on [0, 1], we have
that
|Pf|
L
2 =
__
1
0
(x)
2
[f(x)[
2
dx
_
1
2
__
1
0
[f(x)[
2
dx
_
1
2
= |f|
L
2.
Thus, P is bounded.
In conclusion, the necessary and sucient conditions for P to be an orthogonal
projection is a measurable satisfying (x) = 0 or (x) = 1 for almost all x
[0, 1].
Problem 88. Consider the right-shift on the Hilbert space
2
:
S :
2
2
, S(
1
,
2
, ...) = (0,
1
,
2
, ...).
Dene its adjoint operator.
Solution.
For (
n
) = (
1
,
2
, ...) and (
n
) = (
1
,
2
, ...) in
2
,
(
n
), (
n
)
_
=
(
n
), S(
n
)
_
= (
1
,
2
, ...), (0,
1
,
2
, ...))
=
2
1
+
3
2
+...
= (
2
,
3
, ...), (
1
,
2
, ...)).
Thus
S
(
1
,
2
, ...) = (
2
,
3
, ...).
Hence, the adjoint of the right-shift is the left-shift.
www.MATHVN.com
91
Problem 89. Let A :
2
2
be dened by
Ax = A(x
1
, x
2
, x
3
, ...) = (0, 0, x
3
, x
4
, ...).
Prove A is linear, continuous, self-adjoint and positive. Find
A.
Solution.
With similar argument as the previous problem, we can show that A is linear. We
have
|Ax|
2
=
k=3
[x
k
[
2
k=1
[x
k
[
2
= |x|
2
, x = (x
1
, x
2
, x
3
, ...)
2
.
Therefore,
|Ax| |x|, x
2
.
This shows that A is continuous. Also, for all x, y
2
,
Ax, y) =
k=3
x
k
y
k
= x, Ay).
Hence, A is self-adjoint. And,
Ax, x) =
k=3
[x
k
[
2
0, x
2
.
So A is positive. Then there exists the square root
A :
2
2
. We have
A
2
x = A(Ax) = A(0, 0, x
3
, x
4
, ...) = (0, 0, x
3
, x
4
, ...), x
2
.
It follows that A
2
= A. Hence,
A = A.
Problem 90.(Multiplication operator)
Let (X, , ) be a -nite measure space. Consider the Hilbert space H =
L
2
(X, , ) =: L
2
(). If L
(), dene
M
: L
2
() L
2
() by M
f = f.
(a) Show that
M
B(H) and |M
| = ||
.
Here ||
is the -essential supremum norm.

(b) Show that M
= M
.
(c) Show that M
is normal. When M
is self adjoint? unitary?

www.MATHVN.com
Solution.
(a) The linearity of the operator M
is evident. We show that M
is bounded and
calculate its norm.
By denition, ||
is the inmum of all c > 0 such that [(x)[ c a.e. [], and so
[(x)[ ||
a.e. []. Thus we can assume that is a bounded measurable and

[(x)[ ||
for all x. If f L
2
(), then
_
[f[
2
d ||
2
_
[f[
2
d.
That is,
M
B
_
L
2
()
_
and |M
| ||
().
If > 0, the -niteness of the measure space implies that
such that 0 < () < and [(x)[ ||
, x .
If we take f =
1
()
, then we have
f L
2
() and |f|
2
= 1.
So
|M
|
2
|f|
2
=
1
()
_
[[
2
d (||
)
2
.
Letting 0, we get that
|M
| ||
().
(*) and (**) give that |M
| = ||
.
(b) For f, g L
2
(), we have
f, M
g) =
_
f
_
M
g
_
d
=
_
f(g)d
=
_
(
f) gd
= M
f, g).
This shows that
M
= M
.
(c) Every multiplication operator M
is normal. Indeed,
M
= M
= M
= M
.
www.MATHVN.com
93
M
is self-adjoint if and only if =

, that is, is real-valued.
M
is unitary if and only if [[ = 1 a.e. [].

Problem 91. Let H be a Hilbert space and A B(H). Show that
(a) Image A = (ker A
.
(b) ker A = (Image A
.
Solution.
(a) Take any x Image A. Then there is a y H such that x = Ay. For any
z ker A
, we have
x, z) = Ay, z) = y, A
z) = y, 0) = 0.
hence x (ker A
. This proves that Image A (ker A
. Since (ker A
is
closed, it follows that
(i) Image A (ker A
.
On the other hand, if x (Image A)
, then for all y H, we have

0 = Ay, x) = y, A
x).
Therefore A
x = 0, that is, x ker A
. This prove that (Image A)
ker A
.
Taking orthogonal complements both sides, we obtain
(ii) (ker A
Image A Image A.
From (i) and (ii) it follows that
Image A = (ker A
.
(b) Replacing A by A
in (a), we get
(ker A)
= Image A
.
Taking orthogonal complements both sides and using a result in Problem 39, we
obtain
ker A = (Image A
= (Image A
.
www.MATHVN.com
Problem 92.(Integral operator)
Let (X, , ) be a measure space. Let k : X X F be an -measurable
function for which there are constants c
1
and c
2
such that
_
X
[k(x, y)[d(y) c
1
a.e.[],
_
X
[k(x, y)[d(x) c
2
a.e.[].
Consider the operator K : L
2
() L
2
() dened by
(Kf)(x) =
_
k(x, y)f(y)d(y).
The function k is called the kernel of the operator K.
(a) Show that K is a bounded linear operator and |K|
c
1
c
2
.
(b) Show that K
is the integral operator with kernel k
(x, y) = k(x, y).

Solution.
(a) Linearity of K comes from linearity of the integral
_
. It suces to show that K
is bounded. Actually it must be shown rst that Kf L
2
(), but this will follow
from the argument that demonstrates the boundedness of K. If f L
2
(),
[Kf(x)[
_
[k(x, y)[ [f(y)[d(y)
=
_
[k(x, y)[
1/2
[k(x, y)[
1/2
[f(y)[d(y)
__
[k(x, y)[d(y)
_
1/2
__
[k(x, y)[ [f(y)[
2
d(y)
_
1/2

c
1
__
[k(x, y)[ [f(y)[
2
d(y)
_
1/2
.
Hence
_
[Kf(x)[
2
d(y) c
1
_ _
[k(x, y)[ [f(y)[
2
d(y)d(x)
= c
1
_
[f(y)[
2
_
[k(x, y)[d(x)d(y)
c
1
c
2
|f|
2
.
www.MATHVN.com
95
Now this shows that the formula used to dene Kf is nite a.e. [], and so
Kf L
2
() and |Kf|
2
c
1
c
2
|f|
2
.
(b) By denition,
Kf, g) =
_
k(x, y)f(y)g(y)d(y)
=
_
f(y)k(x, y)g(y)d(y)
= f, K
g), where K
g(y) =
_
k(x, y)g(y)d(y).
Hence, the kernel of K
is k
(x, y) = k(x, y).

Problem 93. Let H = H H where H be a Hilbert space. Let A B(H) and
B be the operator dened on H by
B =
_
0 iA
iA
0
_
Prove that |A| = |B| and that B is self-adjoint.
Solution.
For any x =
_
x
1
x
2
_
, y =
_
y
1
y
2
_
with x
1
, x
2
, y
1
, y
2
H we have
Bx, y) =
__
0 iA
iA
0
__
x
1
x
2
_
,
_
y
1
y
2
__
= iAx
2
, y
1
) +iA
x
1
, y
2
)
= x
2
, iA
y
1
) +x
1
, iAy
2
)
=
__
x
1
x
2
_
,
_
0 iA
iA
0
__
y
1
y
2
__
= x, By).
Moreover,
|Bx|
2
= |iAx
2
|
2
+| iA
x
1
|
2
(max|A|, |A
|)
2
|x|
2
= |A|
2
|x|
2
.
www.MATHVN.com
Hence, |B| |A|. Conversely, one can take x =
_
0
x
2
_
and obtain
|B x| = |Ax
2
| |B|| x| = |B||x
2
|.
Therefore, |A| |B|. Finally, we obtain |A| |B|.
Remark:
Note that norm on H H is
|(a, b)| = |a| +|b|, a, b H.
Problem 94. Let T be a self-adjoint operator on a Hilbert space H. Show that
its norm is given by
|T| = sup
|x|=1
[Tx, x)[.
Solution.
For |x| = 1 we have
[Tx, x)[ |Tx||x| = |Tx| |T|.
Therefore
(i) sup
|x|=1
[Tx, x)[ |T|.
In order to establish the inverse inequality, we consider the case:
z H, |z| = 1, Tz ,= 0 and u =
1
Tz where =
_
|Tz|.
If we denote by := sup
|x|=1
[Tx, x)[, then we have
|Tz|
2
=
T(z), u
_
=
1
4
_
T(z +u), z +u
_
T(z u), z u
_

4
_
|z +u|
2
+|z u|
2
=

2
_
|z|
2
+|u|
2
=

2
_
||
2
+|Tz|
2
= |Tz|.
www.MATHVN.com
97
This implies that, for any z H with |z| = 1, we have |Tz| , and hence
(ii) |T| = sup
|x|=1
[Tx, x)[.
(i) and (ii) completes the proof.
Problem 95
Let H be a Hilbert space and A a positive self-adjoint operator on H. Prove that
the following assertions are equivalent:
(i) A(H) is dense in H.
(ii) Ker A = 0.
(iii) A is positive denite, i.e., Ax, x) > 0, x H 0.
Solution.
(i) (ii)
Suppose Ax = 0. Then, for any y H,
Ax, y) = x, Ay) = 0 ( since A is self-adjoint)
xA(H)
xA(H) = H ( since A(H) is dense)
x = 0.
(ii) (iii)
Since A is positive,
A = B exists. It is also a self-adjoint operator on H. To show

A is positive denite, we show Ax, x) = 0 x = 0. Now
0 = Ax, x) = B
2
x, x) = B(Bx), x) = Bx, Bx) = |Bx|
2
.
This implies that Bx = 0. Therefore,
Ax = B(Bx) = 0.
Since Ker A = 0, we have x = 0.
(iii) (i)
Assume that A(H) is not dense in H. Then there is x H0 such that xA(H).
In particular, xAx, i.e., Ax, x) = 0. But A is positive denite, so x = 0, a
contradiction.
www.MATHVN.com
Problem 96
Let H be a Hilbert space. If A, B : H H are self-adjoint operators with
0 A B and B is compact, prove that A is compact.
Solution.
Let (x
n
) be a sequence in the closed unit ball B
H
. Since B is compact, there is a
subsequence (x
n
k
) such that (Bx
n
k
) converges. From the Cauchy-Schwarz inequality
we have
Bx, x) |Bx| |x|, x H.
It follows that
Bx
n
k
Bx
m
k
, x
m
k
x
n
k
) |Bx
m
k
Bx
n
k
| |x
m
k
x
n
k
|
|Bx
m
k
Bx
n
k
| (|x
m
k
| +|x
n
k
|)
2|Bx
m
k
Bx
n
k
|.
From 0 A B we get
A(x
n
k
x
m
k
), x
m
k
x
n
k
) B(x
n
k
x
m
k
), x
m
k
x
n
k
)
2|Bx
m
k
Bx
n
k
|.
On the other hand, we get
|
A(x)|
2
=
A(x),
A(x))
=
A
2
(x), x)
= Ax, x)
|Ax| |x|, x H.
Then
|
Ax
m
k

Ax
n
k
|
2
|A(x
m
k
x
n
k
)| |x
m
k
x
n
k
|
|A(x
m
k
x
n
k
)|(|x
m
k
| +|x
n
k
|)
2|A(x
m
k
x
n
k
)|.
Therefore,
|
Ax
m
k

Ax
n
k
|
2
A(x
m
k
x
n
k
), x
m
k
x
n
k
).
Hence
|
Ax
m
k

Ax
n
k
|
2
2|Bx
m
k
Bx
n
k
|.
From this we see that the sequence (
Ax
n
k
) is Cauchy, hence converges. The
operator
A is compact. And so is the operator A =
A
2
.
www.MATHVN.com
Chapter 7
Compact Operators
In this chapter we study general properties of compact operators on Banach and Hilbert spaces.
Spectral properties of these operators will be discussed later.
Denition 7 Let X and Y be Banach spaces. An operator T B(X, Y ) is called compact op-
erator if the image of every bounded set in X has compact closure in Y (relatively compact set).
Equivalently, T B(X, Y ) is compact if and only if for every bounded sequence (x
n
) in X, (Tx
n
)
has a convergent subsequence in Y .
The set of all compact operators is denoted by B
0
(X, Y ).
Proposition 7 Let X and Y be Banach spaces. Then B
0
(X, Y ) is a closed subspace of B(X, Y ).
That is, if (T
n
) is a sequence of compact operators and T B(X, Y ) such that |T
n
T| 0, then
T B
0
(X, Y ).
Denition 8 (Finite rank operators)
An operator T : X Y has nite rank if Image T := T(X) is nite dimensional.
Proposition 8 Let X and Y be Banach spaces. Every nite rank operator from X to Y is
compact.
Proposition 9 Let X and Y be Banach spaces, and T : X Y be a compact operator. If x
n
w
x
then Tx
n
Tx.
Proposition 10 Let H and K be Hilbert spaces. Then T is compact if and only if T
is compact.
Proposition 11 Let H and K be Hilbert spaces and T B(H, K). Then T is compact if and only
if for any sequence (x
n
) H converging weakly to x, the sequence (Tx
n
) converges (strongly) to
Tx in K.
99
www.MATHVN.com
100 CHAPTER 7. COMPACT OPERATORS
Problem 97. Let X be a Banach space. Prove that if T B(X) is arbitrary
and A B
0
(X), then AT and TA are compact operators. (This is called the two
sides ideal property for compact operators).
Solution.
Suppose (x
n
) is a sequence in H such that |x
n
| 1 for every n N. Since T is
continuous,
|Tx
n
| |T| |x
n
| |T|, n N.
If we set y
n
=
Tx
n
|T|
, and then we have |y
n
| 1 for every n N. Since A is compact,
the sequence (Ay
n
) has a convergent subsequence. Now we have
|T|Ay
n
=
|T|ATx
n
|T|
= ATx
n
, n N.
It follows that the sequence (ATx
n
) also has a convergent subsequence. Thus AT is
compact. The similar argument for TA
Problem 98.
(a) Let X be a Banach space. Show that the identity I : X X is compact if
and only if X has nite dimensional.
(b) Let X, Y be Banach spaces and A B(X, Y ). Suppose that A has the
property:
c > 0 : |Ax| c|x|, x X,
Find condition(s) for X so that A can be a compact operator.
Solution.
(a) See Problem 16.
(b) First we note that A is injective. Indeed,
Ax = 0 cx = 0 x = 0.
Let Z = A(X), then U : X Z dened by U(x) = A(x) is bijective. Let us
consider U
1
: Z X. Clearly U
1
is linear. we claim:
() |U
1
(y)|
1
c
|y|, y Z
www.MATHVN.com
101
Proof: Since y A(X), there is an x X such that A(x) = U(x) = y. This implies
that x = U
1
(y). By our hypothesis,
|Ax| = |U(x)| = |y| c|x|.
Thus,
|y| c|U
1
(y)|.
Hence, (*) is proved. It follows that U
1
is linear and continuous. If U is a compact
operator, then by the ideal property for the compact operators (Problem 97), it
follows that I = U
1
U : X X is compact, which means that X is nite dimen-
sional. Conversely, if X is nite dimensional, then every A B(X, Y ) is compact
(in a nite dimensional normed space, a set is compact if and only if it is closed and
bounded). Hence A is compact if and only if X is nite dimensional.
Problem 99.
Let H and K be Hilbert spaces and A B(H, K). Show that A is compact if
and only if A
A is compact.
Solution.
Suppose A B(H, K) is compact. Let (x
n
) be a sequence in X converging weakly
to 0. We have
|A
Ax
n
| |A
| |Ax
n
|.
Since A is compact, Ax
n
0 (strongly) in Y . Thus A
Ax
n
0, and so A
A is
compact.
Reciprocally, suppose A
A is compact. For any sequence (x

n
) such that x
n
w
0,
we have
|Ax
n
|
2
= Ax
n
, Ax
n
) = x
n
, A
Ax
n
) |A
Ax
n
| |x
n
|.
Since |x| is uniformly bounded and A
A is compact, A
Ax
n
0. Therefore,
Ax
n
0, and hence A is compact.
Problem 100.
Let X be c
0
or
p
, 1 p . Consider the operator
U : X X, U(x) = U(x
1
, x
2
, ...) = (0, x
1
, 0, x
3
, 0, x
5
, ...).
Prove that U is not compact but U
2
is compact.
www.MATHVN.com
Solution.
We rst note that c
0
and
p
, 1 p (with appropriate norms) are Banach
spaces (see Problems 18,19). We have
U
2
(x) = U(U(x)) = U(0, x
1
, 0, x
3
, 0, x
5
, ...) = (0, 0, ...).
Thus, U
2
= 0, therefore U
2
is compact.
On the other hand, if
e
n
= (0, ..., 0
. .
n
, 1, 0, ...) X
then
U(e
2n1
) = e
2n
, n N.
Now, we have explicitly
e
2n
= (0, ..., 0
. .
2n
, 1, 0, ...), e
2(n+k)
= (0, ..., 0
. .
2n
, 0, ..., 0
. .
2k
, 1, 0, ...),
so that
e
2(n+k)
e
2n
= (0, ..., 0
. .
2n
, 1, 0, ..., 0
. .
2k
, 1, 0, ...).
For X = c
0
or X =
we have
|e
2(n+k)
e
2n
|
= 1.
For X =
p
, 1 p < we have
|e
2(n+k)
e
2n
|
p
= 2
1/p
.
It follows that, in both cases, the sequence
_
U(e
2n1
)
_
cannot have any convergent
subsequence. Thus, U is not a compact operator.
Problem 101
Let 1 p < , and = (
n
)
nN
K with sup
nN
[
n
[ < . We dene the
multiplication operator
M
:
p
p
, M
(x) = (
1
x
1
,
2
x
2
, ...), x = (x
1
, x
2
, ...)
p
.
Prove that:
(a) M
is continuous and |M
| = sup
nN
[
n
[.
(b) M
is a compact operator if and only if c

0
.
www.MATHVN.com
103
Solution.
(a) We have
[
n
x
n
[
p
||
p
[x
n
[
p
, n N.
Since the series
n=1
[x
n
[
p
converges and ||
< by hypothesis, the series
n=1
[
n
x
n
[
p
converges. Moreover,
|M
(x)| =
_

n=1
[
n
x
n
[
p
_
1/p
||
n=1
[x
n
[
p
_
1/p
= ||
|x|, x
p
.
This shows that M
is continuous and |M
| ||
. Also, for any n N,

[
n
[ = [M
(e
n
)[ |M
| |e
n
| = |M
|.
Here e
n
= (0, ..., 0
. .
n1
, 1, 0, ...)
p
. Therefore, ||
= sup
nN
[
n
[ |M
|. Thus,
|M
| = sup
nN
[
n
[.
(b) Suppose M
is a compact operator. i.e., M
(B
p) is relatively compact (B
p the
closed unit ball in
p
). Then
> 0, n
N :
k=n
[M
(x)[
p

p
, x B
p.
Let n n
. Then for e
n
B
p we have
k=n
[M
(e
n
)[
p

p
,
that is
n n
[
n
[ < .
So
n
0, that is, c
0
.
Conversely, if
n
0, then
> 0, n
N : n n
[
n
[ < .
Let x = (x
1
, x
2
, ...) B
p, then
k=n
[M
(x)[
p
=
k=n
[
k
x
k
[
p

p
k=n
[x
k
[
p

p
k=1
[x
k
[
p

p
.
www.MATHVN.com
Therefore M
(B
p) is relatively compact.
Problem 102
Consider the linear operator dened by
T :
2
2
, x = (
1
,
2
,
3
, ...) Tx =
_
1
1
,

2
2
,

3
3
, ...
_
.
Show that T is compact.
Solution.
It is clear that T is linear. To show that it is compact, we will show that it is the
norm limit of a sequence of compact operators. Let
T
n
:
2
2
, T
n
x =
_
1
1
,

2
2
, ...,

n
n
, 0, 0, ...
_
.
Then T
n
is linear, bounded, and of nite rank so compact. Furthermore,
|(T T
n
)x|
2
=
i=n+1
1
i
2
[
i
[
2
1
(n + 1)
2
i=1
[
i
[
2
=
|x|
2
(n + 1)
2
.
Taking the supremum over all x of norm 1, we see that
|T T
n
|
1
n + 1
.
Hence, T
n
T in norm. Thus, T is compact.
Problem 103
Let (c
j
)
j=1
be a sequence of complex numbers. Dene an operator D on
2
by
Dx = (c
1
x
1
, c
2
x
2
, ...), x = (x
1
, x
2
, ...)
2
.
Prove that D is compact if and only if lim
j
c
j
= 0.
www.MATHVN.com
105
Solution.
We note that D is linear. To show that it is compact, we will show that it is the
norm limit of a sequence of compact operators. Suppose lim
j
c
j
= 0. Dene D
n
by
D
n
= (c
1
x
1
, ..., c
n
x
n
, 0, 0, ...).
We obtain that
(D D
n
) = (0, ..., 0.c
n+1
x
n+1
, c
n+2
x
n+2
, ...)
and moreover,
|D D
n
| = sup
jn+1
[c
j
[ 0 as n .
Since each D
n
has nite rank and hence is compact, the operator D is compact.
Assume that (c
j
) does not converge to zero as j . Then, for a given > 0,
there exists a subsequence (c
j
k
) such that [c
j
k
[ . Consider the sequence of vectors
(e
j
) of the standard basis. We have |e
j
k
| = 1 and for any indices m, k we have
|De
jm
De
j
k
|
2
= |c
jm
e
jm
c
j
k
e
j
k
|
2
= [c
jm
[
2
+[c
j
k
[
2
2
2
> 0.
We conclude that the sequence (De
j
k
) does not contain a convergent subsequence
and thus the operator D is not compact.
Trick used in problems 102 and 103 is called cut o method: From the sequence x = (x
1
, x
2
, ...)
we get the sequence (x
1
, ..., x
n
, 0, 0, ...) by cutting o the tail of x.
Problem 104
Let g C[0, 1] be a xed function. Consider the operator A B(C[0, 1]) dened
by
(Au)(s) := g(s)u(s),
i.e., the operator of multiplication by g. Is this operator compact?
Solution.
Note rst that C[0, 1], equipped with the sup-norm, is a Banach space.
It is clear that if g 0 then A is compact. Let us prove that if g is not identically
zero then A is not compact. Indeed, since g is not identically zero, there exists a
subinterval [a, b] [0, 1] such that
m := min
s[a,b]
[g(s)[ > 0.
www.MATHVN.com
Consider the sequence (u
n
):
u
n
C[0, 1]; u
n
(s) := sin
_
2
n
s a
b a

_
; s [0, 1], n N.
It is clear that (u
n
) is a bounded sequence. On the other hand, (Au
n
) does not have
Cauchy subsequences. Indeed, take arbitrary k, n N with k > n. Let
s
n
:= a +
1
2
n+1
(b a)
Then s
n
[a, b] and
|Au
k
Au
n
| = max
s[a,b]
[g(s)(u
k
(s) u
n
(s))[
m max
s[a,b]
[u
k
(s) u
n
(s)[
m[u
k
(s
n
) u
n
(s
n
)[
= m[ sin(2
kn1
) sin(/2)[
= m[0 1[ = m > 0.
Hence (Au
n
) cannot have any convergent subsequence, A is not compact.
Problem 105
Given k L
2
([0, 1] [0, 1]), dene the operator A : L
2
([0, 1]) L
2
([0, 1]) by
(Af)(x) =
_
1
0
k(x, y)f(y)dy.
(a) Show that A is bounded.
(b) Under what condition on k, the operator A is self-adjoint.
(c) Show that A is compact.
Solution.
(Look at Problem 92! They are dierent!)
(a) We estimate |A| to see A is bounded.
|Af|
2
=
_
1
0
_
1
0
k(x, y)f(y)dy
2
dx
_
1
0
__
1
0
[k(x, y)[
2
dy
_
dx.
_
1
0
[f(y)[
2
dy (Cauchy-Schwarz)
|f|
2
.
_
1
0
_
1
0
[k(x, y)[
2
dydx.
www.MATHVN.com
107
Since k L
2
([0, 1] [0, 1]),
_
1
0
_
1
0
[k(x, y)[
2
dydx < ; hence,
|A|
__
1
0
_
1
0
[k(x, y)[
2
dydx
_
1/2
< .
Thus A is bounded.
(b) We have
(Af)(x) =
_
1
0
k(x, y)f(y)dy.
(A
g)(x) =
_
1
0
k(x, y)g(y)dy.
Therefore,
Af, g) =
_
1
0
_
1
0
k(x, y)f(y)dy g(x) dx
=
_
1
0
f(y)
_
1
0
k(x, y) g(x) dxdy
= f, A
g).
Hence, A is self-adjoint if
k(x, y) = k(y, x).
(c) Let (u
j
)
j=1
be an orthonormal basis in L
2
[0, 1]. Then
k(x, y) =
j=1
k
j
(y)u
j
(x), where k
j
(y) =
_
1
0
k(x, y)u
j
(x) dx,
for almost all y. Due to the Parseval identity, we have, for almost all y
_
1
0
[k(x, y)[
2
dx =
j=1
[k
j
(y)[
2
,
and
(1)
_
1
0
_
1
0
[k
j
(y)[
2
dxdy =
j=1
_
1
0
[k
j
(y)[
2
dy.
We now dene the following operator of rank N
k
N
f(x) =
_
1
0
k
N
(x, y)f(y)dy,
www.MATHVN.com
where k
N
(x, y) =
N
j=1
k
j
(y)u
j
(x). By Cauchy-Schwarz inequality we obtain
|(A k
N
)f|
2
=
_
1
0
_
1
0
(k(x, y) k
N
(x, y))f(y)dy
2
dx
__
1
0
_
1
0
[k(x, y) k
N
(x, y)[
2
dxdy
___
1
0
[f(y)[
2
dy
_
.
|f|
2
__
1
0
_
1
0
[k(x, y) k
N
(x, y)[
2
dxdy
_
.
Thus by using that the right hand side in (1) is absolutely convergent, we nd
|(A k
N
)|
2
_
1
0
_
1
0
[k(x, y) k
N
(x, y)[
2
dxdy
=
_
1
0
_
1
0
[k(x, y)[
2
dxdy
_
1
0
_
1
0
k(x, y)
N
j=1
k
j
(y)u
j
(x) dxdy
_
1
0
_
1
0
k(x, y)
N
j=1
k
j
(y)u
j
(x)dxdy +
N
j=1
_
1
0
[k
j
(y)[
2
dy
=
_
1
0
_
1
0
[k(x, y)[
2
dxdy
N
j=1
[k
j
(y)[
2
dy 0 as N .
Problem 106
Part I
Consider the operator
U : C[0, 1] C[0, 1] dened by (Uf)(x) =
_
x
0
e
t
f(t)dt, x [0, 1],
and the sequence of operators
U
n
: C[0, 1] C[0, 1] dened by (U
n
f)(x) =
_
x
0
_
n
k=0
t
k
k!
_
f(t)dt, x [0, 1].
Prove that lim
n
|U
n
U| = 0.
Part II
1. Let M be a set of C
1
-functions f on [0, 1]. Prove that M is relatively compact
in C[0, 1] if f satises following conditions
[f(0)[ k
1
and
_
1
0
[f
t
(x)[
2
dx k
2
www.MATHVN.com
109
where k
1
, k
2
are positive constants. (Hint: Use Arzela-Ascoli theorem).
2. Show that the operator U in Part I is compact.
Solution.
Part I
From Calculus we know that if C[0, 1] then x
_
x
0
(t)dt is continuous. Hence,
U, U
n
take their values in C[0, 1]. Using the Taylor expansion e
t
=
n=0
t
n
n!
we
obtain
(Uf U
n
f)(x) =
_
x
0
_

k=n+1
t
k
k!
_
f(t)dt, x [0, 1].
Then
[(Uf U
n
f)(x)[
_
x
0
_

k=n+1
t
k
k!
_
[f(t)[dt
|f|
_
1
0
_

k=n+1
t
k
k!
_
dt, x [0, 1].
Thus,
|U
n
U| |f|
_
1
0
_

k=n+1
t
k
k!
_
dt, f C[0, 1],
_
1
0
_

k=n+1
t
k
k!
_
dt.
But if u
n
(t) =
k=n+1
t
k
k!
then u
n
: [0, 1] R and
[u
n
(t)[
k=n+1
1
k!
0,
that is, u
n
0 uniformly on [0, 1]. Hence,
_
1
0
u
n
(t)dt 0 as n . Thus,
lim
n
|U
n
U| = 0.
www.MATHVN.com
Part II
1. For all x, y [0, 1] with x < y, and all f M, we have
[f(x) f(y)[ =
_
y
x
[f
t
(x)[
dt
__
y
x
[f
t
(x)[
2
dt
_
1/2
__
y
x
1dt
_
1/2

y x
__
y
x
[f
t
(x)[
2
dt
_
1/2
_
k
2
y x.
This shows that M is qui-continuous.
Now for all x M, and all f M, we have
[f(x)[ [f(x) f(0)[ +[f(0)[
k
1
+
_
k
2
.
So M is uniformly bounded. Thus, by Arzela-Ascoli Theorem
1
, M is relatively
compact in C[0, 1].
2. For |f| 1, let g(x) =
_
1
0
e
tx
f(t)dt, x [0, 1]. Using the dierentiation
theorem for the Riemann integral with parameter, we have
g
t
(x) =
_
1
0
x
(e
tx
f(t))dt =
_
1
0
te
tx
f(t)dt,
[g
t
(x)[
_
1
0
te
tx
[f(t)[dt
_
1
0
te
tx
dt e, x [0, 1].
It follows that the set of all functions g is uniformly Lipschitz. We also have
[g(0)[ =
_
1
0
f(t)dt
_
1
0
[f(t)[dt |f| 1, g.
From previous question, it follows that A = Uf : |f| 1 is relatively compact.
Thus the operator U is compact.
1
Arzela-Ascoli Theorem: Let (X, d) be a compact metric space and A C(X). Then the
following assertions are equivalent:
1. A is relatively compact.
2. A is uniformly bounded and equi-continuous,
3. Any sequence (f
n
) A contains a uniformly convergent subsequence.
www.MATHVN.com
111
Problem 107
(a) Let X be an innite dimensional Banach space, and A be a compact operator
on X. Prove that there is y X such that the equation A(x) = y has no solution,
i.e., A is not surjective.
(b) Let 1 p < , and the operator
U :
p
p
, U(x) =
_
x
1
,
x
2
2
,
x
3
3
, ...
_
, x = (x
1
, x
2
, ...)
p
.
Find an element y
p
for which the equation U(x) = y has no solution.
(c) Consider the operator
A : c
00
c
00
dened by A(x) = A(x
1
, x
2
, ...) =
_
x
1
,
x
2
2
,
x
3
3
, ...
_
.
Prove that A is compact and bijective. On c
00
we have the
- norm.
Solution.
(a) Let us suppose, for a contradiction, that A is surjective. From the open mapping
theorem it follows that A is an open operator, in particular A(B(0; 1)) X is an
open set, i.e.,
> 0 : B(0; ) = B(0; 1) A(B(0; 1)),
where B(0; 1) = x X : |x| < 1. Since A is compact it follows that A(B(0; 1))
is relatively compact. Hence, B(0; ) is relatively compact, from whence B(0; 1) is
relatively compact, therefore compact. But if B(0; 1) is compact, then, by Problem
16, X must be nite dimensional, which is a contradiction. Thus, A is not surjective.
(b) Choose y =
_
1,
1
2
,
1
3
, ...
_

p
. If x = (x
1
, x
2
, ...)
p
has the property that
U(x) = y, then
1
n
x
n
=
1
n
, n N,
that is,
x
n
=
1
n
1
, n N.
www.MATHVN.com
Since x = (x
1
, x
2
, ...)
p
, the generalized harmonic series
n=1
1
n
(1)p
converges
2
,
therefore, ( 1)p > 1 > 1 +
1
p
. From here, it follows that for
y =
_
1,
1
2
1
p+1
,
1
3
1
p+1
, ...
_
,
the equation U(x) = y has no solution.
(c) For n N, consider
A
n
: c
00
c
00
, A
n
(x) = A
n
(x
1
, x
2
, ...) =
_
x
1
,
x
2
2
, ..,
x
n
n
, 0, ...
_
.
Then
|A A
n
| =
1
n + 1
, n N.
Therefore A
n
A in norm. Since every A
n
is a nite rank operator, so A
n
is
compact. The sequence of compact operators (A
n
) converges to A in norm, so A
must be compact. The fact that A is bijective is obvious from the expression which
denes A.
2
The generalized harmonic series

n=1
1
n
p
converges if and only if
p > 1 >
1
p
.
www.MATHVN.com
Chapter 8
Bounded Operators on Banach
Spaces and Their Spectra
Review:
1. Denitions
Let X be a Banach space and T B(X), C.
Resolvent and spectrum of T:
Set T
= T I. The set (T) of all such that T
has an inverse R
(T) = (T I)
1
is called
the resolvent of T. The set (T) = C (T) is called the spectrum of T.
Eigenvalues and eigenvectors of T:
An x ,= 0 which satises Tx = x for some is called an eigenvector of T. The corresponding
is an eigenvalue of T. It is evident that (T).
2. Basic properties
Theorem 12 (Spectrum)
If T is a bounded linear operator on a Banach space X, then its spectrum (T) is compact and lies
in the disk given by
[[ |T|.
Theorem 13 (Resolvent equation)
Let T B(X, X) where X is a Banach space. Then
1. The resolvent R
(T) satises the following equation called the resolvent equation

R
= ( )R
for , (T).
2. R
commutes with any S B(X, X) which commutes with T.

3. We have
R
= R
for , (T).
113
www.MATHVN.com
114CHAPTER 8. BOUNDEDOPERATORS ONBANACHSPACES ANDTHEIR SPECTRA
3. Classication of spectrum
C is called a regular point of A B(X) i (AI)
1
exists and is bounded.
If is not a regular point, it is called a spectrum point. All such points form the spectrum
(A).
Every C with [[ > |A| is a regular point.
Classication of spectrum:
1. The point spectrum:
p
(A) is the set of eigenvalues of A.
2. The continuous spectrum:
c
(A) i (A)
p
(A) and Image(AI) is dense
in X.
3. The residual spectrum:
r
(A) = (A) (
p
(A)
c
(A)). If
r
(A) then
Image(AI) ,= X and ker(AI) = 0
.
4. Spectral radius
Denition 9 Let T B(X, X) where X is a Banach space. The spectral radius r
(T) of T is the
radius of the smallest closed disk centered at the orgin and containing (T).
r
(T) := sup
(T)
[[.
Formula for spectral radius:
It can be shown that
r
(T) = lim
n
n
_
|T
n
|.
5. Spectral mapping theorem
Theorem 14 (Spectral theorem for polynomials)
Let T B(X, X) where X is a Banach space, and
p() =
n
n
+
n1
n1
+... +
0
,
n
,= 0.
Then
(p(T)) = p((T)),
that is, the spectrum of the operator
p(T) =
n
T
n
+
n1
T
n1
+... +
0
I
consists precisely of all those values which the polynomial p assumes on the spectrum (T) of T.

www.MATHVN.com
115
Problem 108
Let X be a Banach space. Suppose that A B(X) is an invertible operator.
Show that
(A
1
) =
1
: ,= 0, (A).
Solution.
For ,= 0, we can write
A
1
1
I = (I A)
1
A
1
.
From this equality we conclude that A
1

1
I is invertible if and only if A I
is invertible. Hence, we have
(A
1
) =
1
: (A).
Problem 109
Let X be a Banach space, let A B(X) and C. Assume that there exists a
sequence (x
n
) in X such that
|x
n
| = 1, n N and Ax
n
x
n
0 as n .
Prove that (A).
Solution.
Assume that AI is invertible. Then, there exists a number c > 0 such that (see
problem 75)
|(A I)x| c|x|, x X.
Replace x by x
n
with |x
n
| = 1 for all n, we have
|Ax
n
x
n
| = |(A I)x
n
| c|x
n
| = c.
This contradicts the condition in the statement of the problem.
www.MATHVN.com
Problem 110
Let (a
n
) and (b
n
) be complex sequences such that
[a
n1
[ > [a
n
[
(n)
0 and [b
n1
[ > [b
n
[
(n)
0.
Consider the operator T :
2
2
dened by
Tx = (a
1
x
1
, a
2
x
2
+b
1
x
1
, a
3
x
3
+b
2
x
2
, ...), x = (x
j
)
2
.
(a) Show that T is compact.
(b) Find all eigenvalues and eigenvectors of T.
Solution.
(a) Let the sequence of operators T
n
:
2

2
, n = 1, 2, ... be dened by for any
x
2
:
(T
n
x)
j
=
_
(Tx)
j
if 1 j n
0 if j > n.
All the T
n
s are operators of nite rank and hence compact. Moreover, we have
|T
n
x Tx| ([a
n+1
[ +[b
n
[)|x|,
which implies that
|T
n
T| 0 as n .
But a uniform limit of a sequence of compact operators is compact. Hence, T is
compact.
(b) Suppose C is an eigenvalue of T and that x ,= 0 is the corresponding
eigenvector. Then
0 = Tx x =
_
(a
1
)x
1
, (a
2
)x
2
+b
1
x
1
, ..., (a
n
)x
n
+b
n1
x
n1
, ...
_
.
If coincides with none of the a
n
s, then x = 0: impossible. So it is necessary that
= a
n
for some n. In this case, x
n
can be chosen arbitrary, and x
1
= ... = x
n1
= 0,
and for k = 1, 2, ... we have 0 = (a
n+k
)x
n+k
+b
n+k1
x
n+k1
. If we choose x
n
= 1
then we get
x
n+k
=
b
n+k1
a
n+k
b
n+k2
a
n+k1
...
b
n
a
n+1
, k = 1, 2, ...
Thus, for any n, = a
n
is a simple eigenvalue. The corresponding eigenvector is
x = (0, ..0, x
n
, x
n+1
, ...) dened as above.
www.MATHVN.com
117
Problem 111
Let X be a Banach space and let A B(X) such that A
n
= 0 for some n N
(A is nilpotent). Find (A).
Solution.
The spectral mapping theorem implies
n
: (A) = (A
n
) = (0) = 0.
Therefore,
(A)
n
= 0 = 0.
Thus, (A) = 0.
Problem 112
Let P B(X) be a projection, i.e., a linear operator on X such that P
2
= P.
Construct the resolvent R(P; ) of P
Solution.
If P = 0, then obviously
P I = I; (P) = 0; R(P; ) := (P I)
1
=
1
I.
If P = I, then
P I = (1 )I; (P) = 1; R(P; ) := (P I)
1
= (1 )
1
I.
Suppose P is non- trivial, i.e., P ,= 0, I. Take any ,= 0, 1. Then using the equalities
P
2
= P; Q
2
= Q and QP = PQ where Q = I P, we obtain
_
(1 )
1
P
1
Q
_
(P I) =
_
(1 )
1
P
1
Q
_
((1 )P Q)
= P +Q = I.
Similarly, we have
(P I)
_
(1 )
1
P
1
Q
_
= I.
Thus,
R(P; ) = (1 )
1
P
1
Q =
1
_
(1 )
1
P I
_
.
www.MATHVN.com
Problem 113
Let C
R
be the space of all continuous and bounded functions x(t) on R with norm
|x| = sup
R
[x(t)[. On the space C
R
we dene the operator A by
(Ax)(t) = x(t +c),
where c R is a constant. Prove that
(A) = C : [[ = 1.
Solution.
Notice that
|Ax| = sup
tR
[x(t +c)[ = sup
R
[x()[ = |x|.
It follows that |A| = 1 and therefore all of the point of C : [[ > 1 are
regular points of A. The operator A is invertible since the operator dened by
(A
1
x)(t) = x(t c)
is bounded and is the inverse of A. Next |A
1
| = 1 and hence all of the point of
C : [[ > 1 are regular points of A
1
. From Problem 65 we deduce that all
of the points C : [[ < 1 are regular points of A.
Consider [[ = 1. This means that
= e
i
, 0 2.
Set a =
i
c
and x
a
(t) = e
at
. We obtain x
a
C
R
and
(Ax
a
)(t) = e
a(t+c)
= e
at
e
ac
= x
a
.
This means that is an eigenvalue of A. Thus,
(A) =
p
(A) = C : [[ = 1.
Problem 114
(a) Let H be a Hilbert space and a, b H. Consider the rank-one operator
U : H H, U(x) = x, a)b, x H.
www.MATHVN.com
119
Calculate |U| and the spectral radius r(U). Show that
r(U) = |U| a, b are linearly independent.
(b) Let X be a Banach space, x
and y X. Consider the rank-one

operator
V : X X, V (x) = x
(x)y, x X.
Calculate |V | and the spectral radius r(V ). Show that
r(V ) = |V | [x
(y)[ = |x
| |y|.
Solution.
(a) We have
|U(x)| = [x, a)[ |b| |a| |b| |x|, x H.
Hence, |U| |a| |b|. We also have
|U| |a| |U(a)| = [a, a)[ |b| = |a|
2
|b|.
Therefore, |U| |a| |b|. Thus,
|U| = |a| |b|.
For x H, let y = U(x). Then
U
2
(x) = U(y) = y, a)b, where y, a) = Ux, a) = x, a)a, b).
Denoting = a, b), we have
U
2
(x) = x, a)b = U(x), x H.
Therefore
U
2
= U.
From here, by induction, we get
U
n
=
n1
U, n N.
Now
r(U) = lim
n
(|U|
n
)
1
n
= lim
n
_
[[
n1
n
|U|
1
n
_
= [[ = [a, b)[.
www.MATHVN.com
The last assertion is a consequence of the fact proved above and the fact that in
the Cauchy-Schwarz inequality we have equality if and only if a, b are linearly
independent.
r(U) = |U| |a| |b| = [a, b)[ a, b are linearly independent.
(b) We have
|V (x)| = [x
(x)[ |y| |x
| |y| |x|, x X.
Therefore, |V | |x
| |y|.
If y = 0 then |V | = 0. Suppose y ,= 0. For x X, we have
[x
(x)[ |y| = |V (x)| |V | |x|.

Then
[x
(x)[
|V |
|y|
|x| |x
|
|V |
|y|
|V | |x
| |y|.
Thus,
|V | = |x
| |y|.
To calculate r(V ) we use the same procedure as in (a). For x H, let z = V (x).
Then
V
2
(x) = V (z) = x
(z)y
z = V (x) = x
(x)y,
x
(z) = x
(x)x
(y).
Therefore,
V
2
(x) = x
(y)x
(x)y = x
(x)y, = x
(y).
Thus, V
2
= V . And from here, by induction, we get
V
n
=
n1
V, n N.
And as above we obtain
r(V ) = [[ = [x
(y)[.
Problem 115
Let k C([0, 1] [0, 1]) be a given function. Consider the operator
B B(C[0, 1]) dened by (Bu)(s) =
_
s
0
k(s, t)u(t)dt.
Find (B) and r(B).
www.MATHVN.com
121
Solution.
Let us prove by induction that
() [(B
n
u)(s)[
M
n
n!
s
n
|u|
, s [0, 1], n 0, 1, 2, ...

where
M := max
(s,t)[0,1]
2
[k(s, t)[.
For n = 0, then B
0
= I, and (*) is trivial. Suppose (*) holds for n = k. Then for
n = k + 1 we have
(B
k+1
u)(s)
_
s
0
k(s, t)(B
k
u)(t)dt
_
s
0
[k(s, t)[
(B
k
u)(t)
dt
M
_
s
0
(B
k
u)(t)
dt
M
_
s
0
M
k
k!
t
k
|u|
dt
=
M
k+1
k!
|u|
_
s
0
t
k
dt
=
M
k+1
(k + 1)!
s
k+1
|u|
, s [0, 1].
Hence, (*) is proved by induction.
It follows from (*) that
|B
n
u|

M
n
n!
|u|
, u C[0, 1],
i.e.,
|B
n
|
M
n
n!
, n 0, 1, 2, ....
Therefore
r(B) = lim
n
|B
n
|
1/n
lim
n
M
(n!)
1/n
= 0.
Since r(B) = 0, (B) cannot contain nonzero elements. Taking into account that
(B) is nonempty, we conclude that (B) = 0.
www.MATHVN.com
Problem 116
Let X be a Banach space and A, B B(X). Suppose AB = BA. Prove that
r(A +B) r(A) +r(B),
where r(T) is the spectral radius of an operator T B(X).
Solution.
Recall
r(T) := sup[[ : (T) and r(T) = lim
n
|T
n
|
1/n
.
Take an arbitrary > 0. The spectral radius formula implies that
|A
n
|
_
r(A) +
_
n
, |B
n
|
_
r(B) +
_
n
for suciently large n N. Therefore there exists a constant M 1 such that
|A
n
| M
_
r(A) +
_
n
, |B
n
| M
_
r(B) +
_
n
, n N.
Since AB = BA, we have
(A +B)
n
=
n
k=0
n!
k!(n k)!
A
nk
B
k
.
Hence
|(A +B)
n
|
n
k=0
n!
k!(n k)!
|A
nk
| |B
k
|
M
2
n
k=0
n!
k!(n k)!
_
r(A) + )
nk
(r(B) +
_
k
= M
2
_
r(A) +r(B) + 2
_
n
, n N.
Consequently,
r(A +B) = lim
n
|(A +B)
n
|
1/n
r(A) +r(B) + 2, > 0.
This implies that
r(A +B) r(A) +r(B).
www.MATHVN.com
123
Problem 117
Let K C be an arbitrary non-empty compact set. Construct an operator
B B(
p
), 1 p , such that (B) = K.
Solution.
Let
k
kN
be a dense subset of K. (Recall that every metric compact space is
separable, that is, it contains a countable dense subset.) Consider the operator
B :
p
p
dened by
Bx = (
1
x
1
,
2
x
2
, ...), x = (x
1
, x
2
, ...)
p
.
Then B is a bounded operator and
k
s are its eigenvalues. Consequently,
kN
(B).
Since (B) is closed and
k
kN
is dense in K,
K (B).
On the other hand, let C K. Then d := inf
kN
[
k
[ > 0 and B I has a
bounded inverse (B I)
1
:
p
p
dened by
(B I)
1
x =
_
1
x
1
, ...,
1
x
k
, ...
_
, x = (x
1
, x
2
, ...)
p
.
Hence, / (B). Therefore, (B) K. Finally,
(B) = K.
Problem 118
Let g C[0, 1] be a xed function and A B(C[0, 1]) be dened by
(Af)(t) = g(t)f(t), t [0, 1].
Find (A) and construct eectively the resolvent R(A; ). Find the eigenvalues
and eigenvectors of A.
www.MATHVN.com
Solution.
Let C, / g([0, 1]) := g(t); t [0, 1]. Then, since g C[0, 1],
1
g
C[0, 1]
and A I has an inverse
R(A; ) = (A I)
1
B(C[0, 1])
dened by
R(A; )f(t) = (g(t) )
1
f(t), t [0, 1].
Hence, (A) g[0, 1].
Suppose now g([0, 1]), i.e., = g(t
0
) for some t
0
[0, 1]. Then
(A I)f(t
0
) = (g(t
0
) )f(t
0
) = 0,
i.e., Image(A I) consists of functions vanishing at t
0
.
Consequently, Image(A I) ,= C[0, 1] and A I is not invertible. Therefore,
g([0, 1]) (A). Finally,
(A) = g([0, 1]).
Take an arbitrary g([0, 1]). Let g
1
() := [0, 1] : g() = . The equation
Af = f, i.e., (g(t) )f(t) = 0 is equivalent to f(t) = 0, t [0, 1] g
1
() . If
g
1
() contains an interval of positive length, then it is easy to see that the set
f C[0, 1] 0 : f(t) = 0, t [0, 1] g
1
()
is non-empty and coincides with the set of all eigenvectors corresponding to the
eigenvalues . If g
1
() does not contain an interval of positive length, then [0, 1]
g
1
() is dense in [0, 1] and f(t) = 0, t [0, 1] g
1
() implies by continuity that
f 0. In this case is not an eigenvalue.
Problem 119
Let X be a Banach space and A, B B(X). Show that for any (A) (B),
R(B; ) R(A; ) = R(B; )(A B)R(A; ).
www.MATHVN.com
125
Solution.
R(B; )(A B)R(A; ) = R(B; )
_
(A I) (B I)
R(A; )
=
_
R(B; )(A I) R(B; )(B I)
R(A; )
= R(B; )(A I)R(A; ) R(B; )(B I)R(A; )
= R(B; ) R(A; ).
Problem 120
Let k C([0, 1] [0, 1]) be given. Consider the operator B B(C[0, 1]) dened
by
(Bu)(s) =
_
s
0
k(s, t)u(t)dt.
Find the spectral radius of B. What is the spectrum of B. ( Hint: Prove by
induction that
[(B
n
u)(s)[
M
n
n!
s
n
|u|
, n N,
for some constant M > 0).
Solution.
Let us rst prove by induction that
[(B
n
u)(s)[
M
n
n!
s
n
|u|
, s [0, 1], n N, (1)

where
M := max
(s,t)[0,1]
2
[k(s, t)[.
(1) is true for n = 1. Indeed,
[(Bu)(s)[ M
_
s
0
[u(t)[dt M|u|
_
s
0
dt =
M
1
1!
s
1
|u|
.
www.MATHVN.com
Suppose (1) holds for n = k. Then for n = k + 1 we have
[(B
k+1
u)(s)[ =
_
s
0
k(s, t)(B
k
u)(t)dt
_
s
0
[k(s, t)[[(B
k
u)(t)[dt
M
_
s
0
[(B
k
u)(t)[dt
M
_
s
0
M
k
k!
t
k
|u|
dt
=
M
k+1
k!
|u|
_
s
0
t
k
dt
=
M
k+1
(k + 1)!
s
k+1
|u|
, s [0, 1].
Hence, (1) is true for n = k + 1. Thus (1) is proved by induction.
It follows from (1) that
|B
n
u|

M
n
n!
|u|
, u C[0, 1].
It follows that
|B
n
|
M
n
n!
, n N.
Therefore,
r(B) = lim
n
|B
n
|
1/n
lim
n
M
(n!)
1/n
= 0.
Since r(B) = 0, (B) cannot contain non-zero elements. Taking into account that
(B) is not empty, we conclude that (B) = 0.
Problem 121
Determine the spectra of the left and the right shift operators on
2
:
R(x
1
, x
2
, x
3
, ....) = (0, x
1
, x
2
, ...),
L(x
1
, x
2
, x
3
, ....) = (x
2
, x
3
, x
4
, ....).
Classify them into point, continuous, and residual spectrum.
Solution.
We have shown that |R| = |L| = 1 (problem 46). It follows that
C : [[ > 1 (R) and C : [[ > 1 (L).
www.MATHVN.com
127
Now I prove the following four claims, and then I will state the conclusion.
Claim(1): R I is injective (one-to-one) for all C such that [[ 1.
Proof.
If = 0, then
Rx = 0 x
i
= 0, i N.
Hence x = 0, and so R I is injective.
Suppose 0 < [[ 1. Then
(R I)x = 0 (0, x
1
, x
2
, ...) = (x
1
, x
2
, x
3
, ....).
Since ,= 0, this implies that x
1
= x
2
= ... = 0. Hence x = 0, and R I is
injective.
Claim(2): R I is not surjective (onto) for all C such that [[ 1.
Proof.
Note that if = 0, then e
1
= (1, 0, ...) / Image(R I) = Image R.
Suppose 0 < [[ 1. Then
(R I)x = e
1
x
n
=
1
n
, n N
|x|
2
=
n=1
[x
n
[
2
=
n=1
_
1
[[
2
_
n
.
The above series cannot be convergent because 0 < [[ 1 implies that
1
[[
2
1.
And hence e
1
/ Image(R I). Therefore, R I is not surjective.
Claim(3): L I injective (one-to-one) for all C such that [[ = 1.
Proof.
Suppose it was not injective. There would be some nonzero x
2
such that
(L I)x = 0. Then
(x
2
, x
3
, x
4
, ....) = (x
1
, x
2
, x
3
, ....).
Hence
x
n
=
n1
x
1
, n N.
Since x ,= 0, x
1
,= 0. Since [[ = 1, we have
|x| =
n=0
[x
1
[
2
[[
2n
=
n=0
[x
1
[
2
.
This sum cannot be nite since x
1
,= 0, but this is impossible. So x must be zero,
and hence L I injective.
www.MATHVN.com
Claim(4): L I is not injective (one-to-one) for all C such that [[ < 1.
Proof.
By a similar argument as above, we see that any nonzero x that satises the equation
(L I)x = 0 is of the form
x = (x
1
, x
1
,
2
x
1
,
3
x
1
, ....).
Choose x
1
= 1, then
|x| =
n=0
([[
2
)
n
.
The series is convergent since [[ < 1, so x
2
is nonzero and satises the equation
(L I)x = 0. Thus L I is not injective.
Conclusion:
Claims (1) and (2) show that
(R) = C : [[ 1.
Recall that Image(RI) is dense if and only if ker(RI)
= ker(L
I) = 0.
Since [
[ = [[, claims (3) and (4) show that Image(R I) is dense if and only if
[[ = 1. Therefore
c
(R) = C : [[ = 1 and
r
(R) = C : [[ < 1.
Also from the above results we get
p
(R) = .
Note that claim (4) shows us that
p
(L) = C : [[ < 1.
Since (L) is a closed set, we get
(L) = C : [[ 1.
(As from the above we know that [[ > 1 implies that (L).)
For [[ = 1 we know that ker(L I)
= ker(R

I) = 0 by claim (1). Hence
Image(L I) is dense. Therefore
c
(L) = C : [[ = 1 and
r
(L) = .

www.MATHVN.com
129
Alternate solution.
Consider R, L :
2
2
dened by
Rx = (0, x
1
, x
2
, ...); Lx = (x
2
, x
3
, ...); x = (x
1
, x
2
, x
3
, ...)
2
.
It is clear that |R| = |L| = 1. So, every C with [[ > 1 is a regular point for
both of the operators R and L. Concerning the eigenvalues of these operators, we
obtain the following:
Lx = x (x ,= 0) x
2
= x
1
; x
3
= x
2
; ...
x = (1, ,
2
,
3
, ...)x
1
.
Such a vector belongs to
2
i [[ < 1. Hence,
p
(L) = C : [[ < 1.
From the above result we also have dimker(L I) = 1.
For R we have
Rx = x (x ,= 0) 0 = x
1
; x
1
= x
2
; x
2
= x
3
; ...
x
1
= x
2
= x
3
= ... = 0...
x = 0 : a contradiction.
Hence,
p
(R) = .
Next, since L
= R and R
= L, we obtain
(1) Image(R I)
= ker(L
I),
(2) Image(L I)
= ker(R
I).
For [[ < 1 the relation (1) yields
codimImage(R I) = dimker(L
I) = 1.
Hence, C : [[ < 1
r
(R). Since the spectrum of an operator is closed, we
conclude that
C : [[ = 1 (L) and C : [[ = 1 (R).
Moreover, for [[ = 1, form (1) and (2) we have
Image(R I) = Image(L I) =
2
.
www.MATHVN.com
Hence,
c
(R) =
c
(L) = C : [[ = 1.
Conclusion:
(R) = (L) = C : [[ 1
p
(R) =
p
(L) = C : [[ < 1
r
(L) = ,
r
(R) = C : [[ < 1
c
(L) =
c
(R) = C : [[ = 1.
www.MATHVN.com
Chapter 9
Compact Operators and Their
Spectra
As bounded linear operators, compact operators share spectral properties of bounded linear oper-
ators. Besides, compact operators have some more particular spectral properties.
Let T B(X) be a compact operator on a Banach space X. Suppose dimX = .
1. 0 (T). Every spectral value ,= 0 is an eigenvalue.
2. For ,= 0, dimker(T
) dimker(T I) < .
3. For ,= 0, the range of T
T I is closed.
4. The set of eigenvalues of T, namely
p
(T), is at most countable. The value = 0 is the only
possible point of accumulation of that set.
Problem 122
Show that
(a) dimker(T
n
) < n N,
(b) 0 = ker(T
0
) ker(T
1
) ker(T
2
) ...
Solution.
Since T
is linear, T
0 = 0. By induction we get
T
n
x = 0 T
n+1
x = 0, n N,
131
www.MATHVN.com
132 CHAPTER 9. COMPACT OPERATORS AND THEIR SPECTRA
and so (b) follows.
We now prove (a). By the binomial formula,
T
n
= (T I)
n
=
n
k=0
_
n
k
_
T
k
()
nk
= ()
n
I +T
n
k=1
_
n
k
_
T
k1
()
nk
. .
S
.
This can be written
T
n
= W I, where = ()
n
.
Note that T is compact and S is bounded, so W = TS = ST is compact. The
property 2 above gives that dimker(T
n
) < .
Problem 123
Show that
0 (T).
Solution.
If 0 / (T) then T is invertible, and we have TT
1
= I. But T and T
1
are compact,
so I is compact. This requires that the dimension of X is nite (problems 16, 98):
a contradiction. Thus 0 (T).
Problem 124
Let T : X X be a compact operator on a normed space X and let ,= 0. Then
there exists a smallest integer r (depending on ) such that from n = r on, the
kernels ker(T
n
) are equal, and if r > 0, the inclusions

ker(T
0
) ker(T
1
) ... ker(T
r
)
are all proper (strict).
www.MATHVN.com
133
Solution.
For simplicity, we let N
n
:= ker(T
n
).
We know that N
m
N
m+1
(Problem 122). Suppose that N
m
= N
m+1
for no m.
Then N
n
is a proper subspace of N
n+1
for every n. Since these kernels are closed,
by Riesz lemma, there is a sequence (y
n
) in N
n
such that
|y
n
| = 1 and |y
n
x|
1
2
x N
n1
.
We show that
() |Ty
n
Ty
m
|
1
2
[[ for m < n,
so that the sequence (Ty
n
) has no convergence subsequences. This contradicts the
compactness of T.
From T
= T I we have T = T
+I and
Ty
n
Ty
m
= y
n
x where x = T
y
m
+y
m
T
y
n
.
Let m < n. We show that x N
n1
. Since m n1, we clearly have y
m
N
m

N
n1
. Also y
m
N
m
implies
0 = T
m
y
m
= T
m1
(T
y
m
),
that is, T
y
m
N
m1
N
n1
. Similarly, y
n
N
n
implies T
y
n
N
n1
. Together,
x N
n1
. Also x =
1
x N
n1
. Hence
|y
n
x| = [[ |y
n
x|
1
2
[[.
Thus we have (). Therefore, we must have N
m
= N
m+1
for some m.
We now prove that
() N
m
= N
m+1
=N
n
= N
n+1
for all n > m.
Assume that this does not hold. Then N
n
n+1
for some
n > m. We consider an x N
n+1
N
n
. By denition,
T
n+1
x = 0 but T
n
x ,= 0.
Set z = T
nm
x. Then
T
m+1
z = T
n+1
x = 0 but T
m
z = T
n
x ,= 0.
Hence
z N
m+1
but z / N
m
.
So N
m
m+1
. This contradicts (). The rst statement is
proved, where r is the smallest n such that N
n
= N
n+1
. Consequently, if r > 0, the
inclusions in the theorem are strict.
www.MATHVN.com
Problem 125
Let T : X X be a compact operator on a Banach space X and let ,= 0.
Then there exists a smallest integer q (depending on ) such that from n = q
on, the ranges T
n
(X) are equal, and if q > 0, the inclusions

T
0
(X) T
1
(X) ... T
q
(X)
are all proper (strict).
Solution.
For simplicity, we let R
n
:= T
n
(X). Suppose that R

s
= R
s+1
for no s. Then R
n+1
is
a proper subspace of R
n
for every n. Since these ranges are closed, by Riesz lemma,
there is a sequence (x
n
) in R
n
such that
|x
n
| = 1 and |x
n
x|
1
2
x R
n+1
.
Let m < n. Since T = T
+I, we can write

(i) Tx
m
Tx
n
= x
m
(T
x
m
+T
x
n
+x
n
).
On the right hand side, x
m
R
m
, x
m
R
m
, so that T
x
m
R
m+1
. Since n > m,
also T
x
n
+x
n
R
n
R
n+1
. Hence (i) is of the form
Tx
m
Tx
n
= (x
m
x) with x R
m+1
.
Consequently,
|Tx
m
Tx
n
| = [[ |x
m
x|
1
2
[[.
This contradicts the fact that (Tx
n
) has a convergent subsequence since (x
n
) is
bounded and T is compact. Thus, R
s
= R
s+1
for some s. Let q be the smallest s
such that R
s
= R
s+1
. Then, if q > 0, the inclusions in the theorem are proper. Fur-
thermore, R
q+1
= R
q
means that T
maps R
q
onto itself. Hence repeated application
of T
gives R
n+1
= R
n
for every n > q.
Problem 126
Let A be an invertible operator, and let K be a compact operator in a Banach
space. Prove that
(a) dim(ker(A +K)) < .
(b) codim(Image(A +K)) < .
www.MATHVN.com
135
Solution.
(a) Since A is invertible, we can write
A +K = A(I +A
1
K).
The operator A
1
K is compact (see note above), so dim(ker(I +A
1
K)) < . This
implies that
A invertible ker(AB) = ker(B), B B(X).
Indeed,
x ker(B) ABx = A0 = 0
x ker(AB).
And
x ker(AB) ABx = 0
Bx = 0
x ker(B).
It follows that
dim(ker(A +K)) = dim(A(I +A
1
K))
= dim(I +A
1
K) < .
(b) One can write
A +K = (I +KA
1
)A.
The operator KA
1
is compact, so codim(I + KA
1
) < . This implies that if A
is invertible, and then Image(BA) = Image(B). Indeed,
x Image(B) x = By
x = (BA)A
1
y Image(BA).
And
x Image(BA) x = BAy
x = B(Ay) Image(B).
Thus, we obtain
codim(Image(A +K)) = codim(Image((I +KA
1
)A))
= codim(Image(I +KA
1
)) < .
www.MATHVN.com
Problem 127
2
([0, 1]) L
2
([0, 1]) dened by
(Kf)(t) =
_
1
0
k(t, s)f(s)ds
where k(s, t) = mint, s : t, s [0, 1].
(a) Prove that K is a compact self-adjoint operator.
(b) Find the spectrum (K) and the norm |K|.
Solution.
(a) Since k(t, s) = k(s, t) and k(t, s) is a continuous function, the operator is self-
adjoint and compact (see Problem 105).
(b) Since K is compact and self-adjoint, the spectrum of K consists of zero and real
eigenvalues
1
. Assume that y = Ky. This means that
(1) y(t) =
_
t
0
mint, sy(s)ds +
_
1
t
mint, sy(s)ds
=
_
t
0
sy(s)ds +
_
1
t
ty(s)ds
=
_
t
0
sy(s)ds +t
_
1
t
y(s)ds.
Taking the derivative twice, we obtain
(2) y
t
(t) = ty(t) +
_
1
t
y(s)ds ty(t) =
_
1
t
y(s)ds
y
tt
(t) = y(t).
Clearly, ,= 0; otherwise, y = 0 so ker(K) = 0. We have the dierential equation
(3) y
tt
+y = 0 with b.v.c. y
t
(0) = y(0) = 0
because of (1). Let us prove that > 0, which means that the operator K is positive.
Multiplying (3) by y and integrating we obtain
_
1
0
y
tt
(t) y(t)dt +|y|
2
= 0.
1
If T B(H) is self adjoint, all its eigenvalues are real. (We will see this in the next chapter.)
www.MATHVN.com
137
Integrating by parts we obtain
_
y
t
y[
1
0
_
1
0
[y
t
[
2
dt
_
+|y|
2
= 0.
The b.v.c. yield
_
1
0
[y
t
[
2
dt +|y|
2
= 0.
Hence > 0.
The solution of the dierential equation is
y = C
1
cos
t
+C
2
sin
t
.
From the b.v.c. it follows that C
1
= 0 and
C
2
cos
1
= 0. Therefore, the eigenvalues

of K are
k
=
4
2
(2k 1)
2
, k = 1, 2, ...
Since K is seft-adjoint, we obtain
|K| = max
kN
[
k
[ = [
1
[ =
4
2
.
Problem 128(Similar problem)
2
([0, 1]) L
2
([0, 1]) dened by
(Kf)(t) =
_
1
0
k(t, s)f(s)ds
where k(s, t) = maxt, s : t, s [0, 1].
(a) Prove that K is a compact self-adjoint operator.
(b) Find the spectrum (K) and the norm |K|.
(c) Is K a positive operator?
Problem 129
Let S be the operator dened on C[0, 1] by
(Sf)(x) =
_
x
0
f(y)dy.
(a) Compute the spectrum of S.
(b) Show that S is compact.
www.MATHVN.com
Solution.
(a) First we show that S is continuous. For f, g C[0, 1] we have
|Sf Sg| = |S(f g)| =
_
x
0
(f(y) g(y))dy
_
x
0
[f(y) g(y)[dy
sup
x[0,1]
_
x
0
[f(y) g(y)[dy
[[0, 1][.|f g| = |f g|.
Hence, S is Lipschitz continuous with constant 1. Thus, |S| 1. Next, we show
that |S
n
|
1/n
0 as n . Observe that
[(Sf)(x)[
_
x
0
[f(t)[dt |f|x = |f|
x
1
1!
, x [0, 1].
By induction
[(S
n+1
f)(x)[
_
x
0
[S
n
f(t)[dt
|f|
_
x
0
t
n
n!
dt
= |f|
x
n+1
(n + 1)!
, x [0, 1], n = 1, 2, ...
Thus,
|S
n
f| |f|
1
n!
, n = 1, 2, ...
So
|S
n
|
1/n
_
1
n!
_
1/n
0 as n .
Recall that the spectral radius of S is given by
r(S) = lim
n
|S
n
|
1/n
.
This implies that r(S) = 0. Thus (S) = 0.
(b) Suppose F C[0, 1] is a bounded subset. Put
|F| := sup
fF
|f|.
www.MATHVN.com
139
Then SF is equi-continuous by the Fundamental Theorem of the Calculus:
f F, [(Sf)
t
(x)[ |f| |F|.
SF is bounded since:
f F, [(Sf)(x)[ |f| |F|.
By Ascoli-Arzela theorem SF is relatively compact. Thus S is compact.
Problem 130
Set
(Tf)(x) =
_
1x
0
f(y)dy, f C[0, 1], x [0, 1].
(a) Prove that T is a linear bounded and compact operator on C[0, 1].
(b) Calculate (T) and the eigenvalues of T.
Solution.
(a)
Linearity of T: trivial.
Boundedness of T:
[(Tf)(x)[ =
_
1x
0
f(y)dy
|f|,
where |f| = max
x[0,1]
[f(x)[. Hence,
|T| 1.
Compactness of T: let (f
n
)
n=1
be a bounded sequence in C[0, 1]. Hence,
|f
n
| M n N
for some M 0. By Arzela-Ascoli theorem, it suces to show that A :=
Tf
n
; n N is bounded and equicontinuous. We have
|Tf
n
| |T||f
n
| |f
n
| M n N.
www.MATHVN.com
Given any > 0, without lost of generality, we can assume x < y, then
[(Tf
n
)(x) (Tf
n
)(y)[ =
_
1x
1y
f(y)dy
M[x y[.
Thus,
[(Tf
n
)(x) (Tf
n
)(y)[ < n N provided [x y[ <

M
.
(b) First we see that = 0 is an eigenvalue. Assume that ,= 0 is an eigenvalue,
i.e.,
g(x) = (Tg)(x) =
_
1x
0
g(y)dy, x [0, 1]
for some 0 ,= g C[0, 1]. This implies that
Tg C
1
[0, 1] and g
t
(x) = g(1 x), x [0, 1].
Moreover, we have g(1) = 0. But g C
1
[0, 1] implies that g C
2
[0, 1]. By
dierentiating once more we get
g
tt
(x) =
g(x)
, x [0, 1] and g(1) = g

t
(0) = 0.
Hence
g(x) = Acos(x/) with g(1) = 0.
This gives that
k
=
1
2
+k
, k Z.
Check if all these s are eigenvalues. We calculate
(Tg
k
)(x) =
_
1x
0
cos(t/
k
)dt
=
k
[sin(t/
k
)]
1x
0
=
k
sin
__
2
+k
_
(1 x)
_
=
k
(1)
k
g
k
(x).
Hence, =
2l
, l Z are the eigenvalues of T, i.e.,
p
(T) =
2l
, l Z.
We know that (T) is closed and (T) 0
p
(T). This yields
(T) = 0
p
(T).
www.MATHVN.com
141
Problem 131
Let T be a compact operator on a Hilbert space H and (
n
) be a sequence of
complex numbers. Suppose there exists a nested sequence of distinct subspaces
(M
n
) such that for all n N
(i) M
n
M
n+1
(ii) (T
n
I)M
n+1
M
n
.
Prove that lim
n
n
= 0.
Solution.
Since M
n
is a subspace of M
n+1
, we can write M
n+1
= M
n
(M
n
)
M
, where (M
n
)
M
is
the orthogonal complement of M
n
in M
n+1
. For short we write (M
n
)
M
= M
n+1
M
n
.
Let e
n
be a sequence of unit vectors dened by
e
1
M
1
, e
n+1
M
n+1
M
n
, n N.
Clearly, that is an orthonormal system. Moreover,
(T
n
I)e
n
, e
n
_
= 0, for all n 2,
which implies that
|Te
n
| [Te
n
, e
n
)[ =
(T
n
I)e
n
, e
n
_
+[
n
e
n
, e
n
)[
= 0 +[
n
e
n
, e
n
)[ = [
n
[.
Since T is compact and e
n
w
0, it follows that lim
n
Te
n
= 0. Thus
lim
n
n
= 0.
Problem 132
Let T be a compact operator on a Hilbert space H and any C > 0. Prove
that there is a nite number of linearly independent eigenvectors x
1
, ..., x
n
of T
corresponding to eigenvalues
1
, ...,
n
such that
i
> C for all i = 1, ..., n.
Solution.
We can rescale to get |x
i
| = 1 for all i = 1, ..., n. Suppose to the contrary that
www.MATHVN.com
there is an innite sequence x
n
of unit vectors, and a sequence of eigenvalues
n
satisfying
n
> C and Tx
n
=
n
x
n
, n N.
Let M
n
= Spanx
1
, ..., x
n
, then M
n
is a nested sequence of subspaces of H and
the inclusions M
n
M
n+1
are strict. Let x M
n
, then there are c
1
, ..., c
n
C such
that x =
n
i=1
c
i
x
i
. So we have
(T
n
I)x = (T
n
I)
n
i=1
c
i
x
i
=
n
i=1
c
i
(T
n
I)x
i
=
n
i=1
c
i
(Tx
i
n
x
i
)
=
n
i=1
c
i
(
i
n
)x
i
M
n1
.
This implies that
(T
n
I)M
n
M
n1
, n 2.
From Problem 130, we obtain
lim
n
n
= 0.
This contradicts the assumption
n
> C for all n N.
Note: Argument in problems 130, 131 is the proof of Proposition 4 in the review at the beginning
of this chapter.
www.MATHVN.com
Chapter 10
Bounded Self Adjoint Operators
and Their Spectra
Review some main points.
Bounded self adjoint operators on Hilbert spaces were dened and considered before. This chapter
is devoted to their spectral properties.
Denition:
Let T B(H) where H is a complex Hilbert space. The adjoint operator of T is the operator
T
: H H dened by
Tx, y) = x, T
y), x, y H.
T is said to be self adjoint if T = T
. We can say that T is self adjoint if and only if

Tx, y) = x, Ty), x, y H.
Another equivalent condition is:
T is self adjoint if and only if Tx, x) is real for all x H.
Let T B(H) be a bounded self adjoint operator on the complex Hilbert space H.
Proposition 12 (Eigenvalues and eigenvectors)
1. All eigenvalues of T (if they exist) are real.
2. Eigenvectors corresponding to dierent eigenvalues of T are orthogonal.
Proposition 13 (Spectrum)
1. (T) R.
2. (T) [m, M] where m = inf
x=1
Tx, x) and M = sup
x=1
Tx, x).
3. m, M (T).
Proposition 14 (Norm)
|T| = max[m[, [M[ = sup
x=1
[Tx, x)[.
143
www.MATHVN.com
144CHAPTER 10. BOUNDEDSELF ADJOINT OPERATORS ANDTHEIR SPECTRA
Problem 133
Let T : H H be a bounded self-adjoint linear operator on a complex Hilbert
space H. Prove that the residual spectrum of T is empty, that is,
r
(T) = .
Solution.
Assume that
r
(T) ,= . Take
r
(T). By the denition of
r
(T), T
1
exists but
its domain is not dense in H. Hence, by the projection theorem, there is a y ,= 0 in
H such that y is perpendicular to the domain D(T
1
) of T
1
. But D(T
1
) is the
range of T
, hence
T
x, y) = 0, x H.
Since is real and T is self-adjoint, we obtain
x, T
y) = 0, x H.
Taking x = T
y, we get |T
y|
2
= 0, so that
T
y = Ty y = 0.
Since y ,= 0, this shows that is an eigenvalue of T. But this contradicts
r
(T).
Second solution:
By Problem 91, noting that T is self adjoint, for any C we have
Image(T I)
= ker(T
I) = ker(T
I).
And hence, if Image(T I) ,= H, then

is an eigenvalue of T. Since T is self
adjoint, is real. Thus =

is an eigenvalue of T. Therefore does not belong to
the residual spectrum of T.
Problem 134
Let T : H H be a bounded self-adjoint linear operator on a complex Hilbert
space H. Prove that
() (T) c > 0 : x H, |T
x| c|x|.
www.MATHVN.com
145
Solution.
If (T) then R
:= T
1
: H H exists and is bounded, say |R
| = k > 0.
Now since I = R
, we have for every x H

|x| = |R
x| |R
| |T
x| = k|T
x|.
This gives
|T
x| c|x|, where c =
1
k
.
Conversely, suppose () holds. We shall show:
(a) T
: H T
(H) is bijective;
(b) T
(H) is dense in H;
(c) T
(H) is closed in H;
so that T
(H) = H and R
:= T
1
is bounded by the bounded inverse theorem.

1
(a) By (7.1), we have for x
1
, x
2
H
|T
x
1
T
x
2
| = |T
(x
1
x
2
)| c|x
1
x
2
|.
Therefore,
T
x
1
= T
x
2
=x
1
= x
2
.
Thus T
: H T
(H) is bijective.
(b) We show that x
0
T
(H) implies x
0
= 0, so that T
(H) = H by the projection

theorem.
2
Let x
0
T
(H) Then for all x H we have

0 = T
x, x
0
) = Tx, x
0
) x, x
0
).
Since T is self-adjoint,
Tx, x
0
) = x, Tx
0
).
It follows that
x, Tx
0
) = x, x
0
) = x,
x
0
).
Thus Tx
0
=

x
0
. So x
0
= 0 since otherwise,

= would be an eigenvalue of T,
and T
x
0
= 0, which would imply
0 = |T
x
0
| c|x
0
| > 0 : a contradiction.
1
A bounded linear operator T from a Banach space X onto a Banach space Y is an open
mapping. Hence if T is bijective, then T is continuous and thus bounded.
2
If Y is a closed subspace of H, then H = Y Y
.
www.MATHVN.com
(c) To show T
(H) is closed we show

y T
(H) =y T
(H).
Let y T
(H). There is a sequence (y

n
) in T
(H) which converges to y. For every

n we have y
n
= T
x
n
for some x
n
H. By (7.1),
|x
n
x
m
|
1
c
|T
(x
n
x
m
)| =
1
c
|y
n
y
m
|.
Hence (x
n
) is Cauchy. Since H is complete, (x
n
) converges, say, x
n
x. Since
T
is continuous, y
n
= T
x
n
T
x. Since the limit is unique, T
x = y. Hence
y T
(H). Thus T
(H) is closed in H.
Problem 135
(a) Let A B(X) where X is a Banach space. Suppose there exists m > 0 such
that
|Ax| m|x|, x X.
Show that Image A is closed in X.
(b) Let A B(H) be self adjoint, where H is a Hilbert space. Let C such
that Im ,= 0. Prove that
|Ax x| [Im[ |x|, x H.
Prove that is a regular point of A.
Solution.
(a) Let (x
n
) be a sequence in X. Suppose Ax
n
y as n . From the hypothesis
we get
|Ax
n
Ax
m
| m|x
n
x
m
| for n ,= m.
Since (Ax
n
) is a Cauchy sequence in X, it follows that (x
n
) is also a Cauchy sequence
in X. Hence x x as n . But A is continuous, so Ax
n
Ax as n . By
uniqueness of the limit, we obtain y = Ax. This shows that Image A is closed.
(b) Let = a +ib with a, b R and b = Im ,= 0. We have for all x H
|Ax x|
2
= Ax (a +ib)x, Ax (a +ib)x)
= (A aI)x ibx, (A aI)x ibx)
= |(A aI)x|
2
+b
2
|x|
2
,
www.MATHVN.com
147
which implies that
|Ax x| [b[|x|, x H.
By part (a), Image(A I) is closed and therefore /
c
(A). By Problem 91
r
(A) = , thus is a regular point of A.
Problem 136
Let A B(H) be self adjoint, where H is a Hilbert space.
(a) Prove that
|A| = sup
x,=0
[Ax, x)[
|x|
2
.
(b) Prove that at least one of |A| or |A| is an element of (A).
Solution.
(a) Using the Cauchy-Schwarz inequality, we have
[Ax, x)[ |Ax| |x| |A| |x|
2
, x H.
Hence
sup
x,=0
[Ax, x)[
|x|
2
|A|. (i)
Now we establish the reverse. Notice rst that for all x, y H we have
A(x +y), x +y
_
A(x y), x y
_
= 2
_
Ax, y) +Ay, x)
.
Using the triangle inequality, we get
2
Ax, y) +Ay, x)
A(x +y), x +y)
A(x y), x y)
.
If we let C = sup
x,=0
[Ax,x)[
|x|
2
, then the Parallelogram Law gives that
Ax, y) +Ay, x)

1
2
C
_
|x +y|
2
+|x y|
2
_
= C
_
|x|
2
+|y|
2
_
. ()
Now let x be any vector with |x| = 1 and let y =
Ax
|Ax|
( the case Ax = 0 does not
give the supremum, hence we may assume that Ax ,= 0). Then |y| = 1. From ()
we get
Ax, Ax)
|Ax|
+
Ax, Ax)
|Ax|
2C.
www.MATHVN.com
Hence |Ax| C. This holds for all x H with |x| = 1. Thus |A| C. (ii)
Combine (i) and (ii) we obtain
|A| = sup
x,=0
[Ax, x)[
|x|
2
.
(b) If we take x arbitrary with |x| = 1, then we get
|A| = sup
|x|=1
[Ax, x)[. ()
Let (x
n
) be a sequence in H with |x
n
| = 1 such that [Ax
n
, x
n
)[ converges to
|A| (this is possible by ()). Let Ax
n
, x
n
) (it may be necessary to pass to
subsequence). Clearly, = |A|. Now,
0 |Ax
n
x
n
|
2
= |Ax
n
|
2
2Ax
n
, x
n
) +
2
|x
n
|
2
2
2
2Ax
n
, x
n
),
which converges to 0 as n . Thus (A) (see Problem 109).

To close this chapter, we introduce a well known theorem relative to compact self adjoint operators
on Hilbert spaces:
The spectral theorem for compact self adjoint operators on Hilbert spaces.
Problem 137
Let T B(H) be a compact self adjoint operator on a Hilbert space H.
(a) There exists a system (nite or innite) of orthonormal eigenvectors
e
1
, e
2
, ... of T and corresponding eigenvalues
1
,
2
, ... of T such that [
1
[
[
2
[ .... If the system is innite then
n
0 as n .
(b) Eigenvectors and eigenvalues mentioned above satisfy the following equation:
Tx =
k=1
k
x, e
k
)e
k
, x H.
Solution.
(a) We use Proposition 13-3 (that we proved in Problem 136b) repeatedly for con-
structing eigenvalues and eigenvectors.
www.MATHVN.com
149
Let H
1
= H and T
1
= T. Then by Proposition 13-3, there exists an eigenvalue
1
of T
1
and a corresponding eigenvector e
1
such that
|e
1
| = 1 and [
1
[ = |T
1
|.
Now Spane
1
is a closed subspace of H
1
hence, by the projection theorem,
H
1
= Spane
1
Spane
1
.
Let H
2
= Spane
1
. Clearly H
2
is a closed subspace of H
1
and T(H
2
) H
2
.
Indeed, if x H
2
then xe
1
, hence Tx = x Txe
1
. Let T
2
be the restriction
of T
1
on H
2
, that is, T
2
= T
1
[
H
2
= T[
H
2
. Then T
2
is a compact and self adjoint
operator in B(H
2
). If T
2
= 0, then there is nothing to prove. Assume that T
2
,= 0.
Then by Proposition 13-3, there exists an eigenvalue
2
of T
2
and a corresponding
eigenvector e
2
such that
|e
2
| = 1 and [
2
[ = |T
2
|.
Since T
2
is a restriction of T
1
,
[
2
[ = |T
2
| |T
1
| =
1
.
By construction e
1
and e
2
are orthonormal.
Now let H
3
= Spane
2
, e
2
. Clearly H
3
H
2
and T(H
3
) H
3
. The operator
T
3
= T[
H
3
is compact and self adjoint....If we continue to proceed in this way, either
after some stage, say n, we get T
n
= 0 or there exists an innite sequence (
n
) of
eigenvalues of T and corresponding eigenvectors (e
n
) satisfying
|e
n
| = 1, n N and [
1
[ [
2
[ ...
If the sequence (
n
) is innite, we show that
n
0 as n . Suppose that
n
0 as n . Then there exists > 0 such that [
n
[ > for innitely many
n. For n ,= m, we have
|Te
n
Te
m
|
2
= |
n
e
n
m
e
m
|
2
=
2
n
+
2
m
>
2
.
This shows that the sequence (Te
n
) has no convergent subsequence, a contradiction
to the compactness of T. Hence
n
0 as n .
(b) There are two cases to consider:
Case 1. T
n
= 0 for some n.
Let x
n
= x

n
k=1
x, e
k
)e
k
for all x H. Then x
n
e
k
for 1 k n, since
e
1
, e
2
, ... is an orthonormal system and
x
n
, e
k
) = x, e
k
) x, e
k
) = 0.
www.MATHVN.com
Hence
0 = T
n
x
n
= Tx
n
k=1
x, e
k
)T
k
e
k
=
n
k=1
k
x, e
k
)e
k
.
That is,
Tx =
n
k=1
k
x, e
k
)e
k
=
k=1
k
x, e
k
)e
k
, x H.
Case 2. T
n
,= 0 for innitely many n.
For x H, by Case 1, we have
_
_
_
_
_
Tx
n
k=1
k
x, e
k
)e
k
_
_
_
_
_
= |T
n
x
x
| |T
n
| |x
n
|
= [
n
[ |x
n
| 0 as n .
Hence
Tx =
k=1
k
x, e
k
)e
k
, x H.
THANK YOU and GOOD LUCK!
Anh Le
leqanh36@gmail.com
www.MATHVN.com
Bibliography
[1] Conway, J.B. A course in Functional Analysis. Second edition. Springer. New
York, 1990.
[2] Kreyszig. E. Introductory Functional Analysis with Applications. John Wiley
and Son. New York, 1989.
[3] Reed, M.; Simon, B. Methods of modern Mathematics Physics. I. Functional
Analysis. Second edition. Academic Press, Inc. New York, 1980.
[4] Royden. H.L. Real Analysis. Third edition. Prentice Hall. NJ, 1988.
[5] Rudin. W Functional Analysis. McGraw-Hill, Inc. New York, 1991.
[6] Yeh, J. Real Analysis. Theory of measure and integration. Second edition. World
Scientic. NJ, 2006.
151
www.MATHVN.com

Functional Problems AnhLe Full

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Functional Problems AnhLe Full

Uploaded by

Copyright:

Available Formats

Functional Analysis Problems with Solutions

ANH QUANG LE, Ph.D.

: the space of all bounded (continuous) linear functionals on X.

be a basis for X. For x X there exist numbers c

is the space of all linear functionals on X. We have

= d. For the space X

= n. From the relation X (X

we conclude that d n. Thus, n = d, and so X = (X

. Then |x| = 1 and

equipped with the norm |.|

is complete. Assume that the sequence (x

. That is, for all > 0, there exists N N such that

< for all j.

. Now for all j N, n N,

is NOT a Banach space.

. Now consider the sequence

norm is not a Banach

is separable. Then there is a set A = a

. So, for all

as well, we conclude that x + y = 0. Hence, = = 0 and therefore [x], [y] are

is the linear subspace of all

, then for every m M and n N we have

, and hence (M +N)

in the place of M and F

in the place of N, we obtain

and , F. Then for any a A,

is closed. Suppose that the sequence (x

. This implies that x (A

. Then xy for all y B. But A B, so xy for all y A.

= 0. If V is not dense in H, that is, V H, pick

, the dual space of L

be the bounded linear operator given by

f(x) = f(x), x G and

we can obtain a corresponding bounded

, so that we can write

. We nally prove that the norm of f is the norm of (

. Thus there is an isometric isomorphism between (c

such that || 1, (y) = d and (z) = 0 for all

and || 1. By the Hahn-Banach theorem

with the same norm.

such that f(x

is closed) and therefore

, and assume that x

be a Hahn-Banach extension for f given by the Hahn-

be a sequence in H. Prove that the following two statements are equivalent:

is called the adjoint operator of A if

= P, then P is called an orthogonal projection.

is the -essential supremum norm.

is self adjoint? unitary?

is evident. We show that M

a.e. []. Thus we can assume that is a bounded measurable and

is self-adjoint if and only if =

is unitary if and only if [[ = 1 a.e. [].

. This proves that Image A (ker A

, then for all y H, we have

x = 0, that is, x ker A

. This prove that (Image A)

is the integral operator with kernel k

(x, y) = k(x, y).

(x, y) = k(x, y).

A = B exists. It is also a self-adjoint operator on H. To show

A is compact. And so is the operator A =

A is compact. For any sequence (x