Fa1 2009 1

Functional Analysis I
Term 1, 2009–2010
Vassili Gelfreich
Contents
1 Vector spaces 1
1.1 Definition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Examples of vector spaces . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Hamel bases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Normed spaces 6
2.1 Norms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Four famous inequalities . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Examples of norms on a space of functions . . . . . . . . . . . . . . 8
2.4 Equivalence of norms . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.5 Linear Isometries . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3 Convergence in a normed space 12

3.1 Definition and examples . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 Topology on a normed space . . . . . . . . . . . . . . . . . . . . . . 14
3.3 Closed sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.4 Compactness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4 Banach spaces 18
4.1 Completeness: Definition and examples . . . . . . . . . . . . . . . . 18
4.2 The completion of a normed space . . . . . . . . . . . . . . . . . . . 20
4.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
5 Lebesgue spaces 25
5.1 Integrable functions . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.2 Properties of Lebesgue integrals . . . . . . . . . . . . . . . . . . . . 28
5.3 Lebesgue space L1 (R) . . . . . . . . . . . . . . . . . . . . . . . . . 31
5.4 L p spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
6 Hilbert spaces 36
6.1 Inner product spaces . . . . . . . . . . . . . . . . . . . . . . . . . . 36
6.2 Natural norms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
6.3 Parallelogram law and polarisation identity . . . . . . . . . . . . . . 38
6.4 Hilbert spaces: Definition and examples . . . . . . . . . . . . . . . . 40
7 Orthonormal bases in Hilbert spaces 41

7.1 Orthonormal sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
7.2 Gram-Schmidt orthonormalisation . . . . . . . . . . . . . . . . . . . 42
7.3 Bessel’s inequality . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
7.4 Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
7.5 Orthonormal basis in a Hilbert space . . . . . . . . . . . . . . . . . . 45
iii
8 Closest points and approximations 48
8.1 Closest points in convex subsets . . . . . . . . . . . . . . . . . . . . 48
8.2 Orthogonal complements . . . . . . . . . . . . . . . . . . . . . . . . 49
8.3 Best approximations . . . . . . . . . . . . . . . . . . . . . . . . . . 51
8.4 Weierstrass Approximation Theorem . . . . . . . . . . . . . . . . . . 53
9 Separable Hilbert spaces 56

9.2 Isometry to `2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
10 Linear maps between Banach spaces 58

10.1 Continuous linear maps . . . . . . . . . . . . . . . . . . . . . . . . . 58
10.2 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
10.3 Kernel and range . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
11 Linear functionals 63
11.2 Riesz representation theorem . . . . . . . . . . . . . . . . . . . . . . 63
12 Linear operators on Hilbert spaces 65

12.1 Complexification . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
12.2 Adjoint operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
12.3 Self-adjoint operators . . . . . . . . . . . . . . . . . . . . . . . . . . 68
13 Introduction to Spectral Theory 71

13.1 Point spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
13.2 Invertible operators . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
13.3 Resolvent and spectrum . . . . . . . . . . . . . . . . . . . . . . . . . 73
13.4 Compact operators . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
13.5 Spectral theory for compact self-adjoint operators . . . . . . . . . . . 78
14 Sturm-Liouville problems 82
iv
Preface
These notes follow the lectures on Functional Analysis given in the Autumn Term of
2009. If you find a mistake or misprint please inform the author by sending an e-mail
to v.gelfreich@warwick.ac.uk. The author thanks James Robinson for his set of
notes and selection of exercises which significantly facilitated the preparation of the
lectures. The author also thanks all students who helped with proofreading the notes.
1 Vector spaces
1.1 Definition.
A vector space V over a field K is a set equipped with two binary operations called
vector addition and multiplication by scalars. Elements of V are called vectors and
elements of K are called scalars. The sum of two vectors x, y ∈ V is denoted x + y, the
product of a scalar α ∈ K and vector x ∈ V is denoted αx.
It is possible to consider vector spaces over an arbitrary field K, but we will con-
sider the fields R and C only. So we will always assume that K denotes either R or C
and refer to V as a real or complex vector space respectively.
In a vector space, addition and multiplication have to satisfy the following set of
axioms: Let x, y, z be arbitrary vectors in V , and α, β be arbitrary scalars in K, then
• Associativity of addition: x + (y + z) = (x + y) + z.
• Commutativity of addition: x + y = y + z.
• There exists an element 0 ∈ V , called the zero vector, such that x + 0 = x for all
x ∈ V.
• For all x ∈ V , there exists an element y ∈ V , called the additive inverse of x, such
that x + y = 0. The additive inverse is denoted −x.
• “Associativity” of multiplication:1 α(β x) = (αβ )x.
• Distributivity:
α(x + y) = αx + αy and (α + β )x = αx + β x.
• There is an element 1 ∈ K such that 1x = x for all x ∈ V . This element is called

the multiplicative identity in K.
1 The purist would not use the word “associativity” for this property as it includes two different
operations: αβ is a product of two scalars and β x involves a vector and scalar.
1
It is convenient to define two additional operations: subtraction of two vectors and
division by a (non-zero) scalar are defined by
x − y = x + (−y),
x/α = (1/α)x.
1.2 Examples of vector spaces

1. Rn is a real vector space.
2. Cn is a complex vector space.
3. Cn is a real vector space.
4. The set of all polynomials P is a vector space:
( )
n
P= ∑ αk x k : αk ∈ K, n ∈ N .
k=0
5. The set of all bounded sequences `∞ (K) is a vector space:

∞
` (K) = (x1 , x2 , . . .) : xk ∈ K for all k ∈ N, sup |xk | < ∞ .
k∈N
For two sequences x, y ∈ `∞ (K), we define x + y by

x + y = (x1 + y1 , x2 + y2 , . . .) .
For α ∈ K, we set
αx = (αx1 , αx2 , . . .) .
We will always use these definitions of addition and multiplication by scalars for
sequences.
6. Let 1 ≤ p < ∞. The set ` p (K) of all pth power summable sequences is a vector
space: ( )
∞
` p (K) = (x1 , x2 , . . .) : xk ∈ K, ∑ |xk | p < ∞ .
k=1
The definition of the multiplication by scalars and vector addition is the same
as in the previous example. Let us check that the sum x + y ∈ ` p (K) for any
x, y ∈ ` p (K). Indeed,
∞ ∞ ∞
∑ |xk + yk | p ≤ ∑ (|xk | + |yk |) p ≤ ∑ 2 p(|xk | p + |yk | p)
k=1 k=1 k=1
∞ ∞
≤ 2 p ∑ |xk | p + 2 p ∑ |yk | p < ∞ .
k=1 k=1
2
7. The space C[0, 1] of all real-valued continuous functions on the closed interval
[0, 1] is a vector space. The addition and multiplication by scalars are defined
naturally: for f , g ∈ C[0, 1] and α ∈ R we defined by f + g the function whose
values are given by
( f + g)(t) = f (t) + g(t) , t ∈ [0, 1] ,
and α f is the function whose values are
(α f )(t) = α f (t) , t ∈ [0, 1] ,
We will always use these definitions on all spaces of functions to be considered

later.
8. The set L̃1 (0, 1) of all real-valued continuous functions f on the open interval
(0, 1) for which
Z 1
| f (t)| dt < ∞
0
is a vector space.
If f ∈ C[0, 1] then f ∈ L̃1 (0, 1). Indeed, since [0, 1] is compact f is bounded (and
attains its lower and upper bounds). Then
Z 1
| f (t)| dt ≤ max | f (t)| < ∞ ,
0 t∈[0,1]
i.e. f ∈ L̃1 (0, 1).

We note that L̃1 (0, 1) contains some functions which do not belong to C[0, 1].
For example, f (t) = t −1/2 is not continuous on [0, 1] but it is continuous on
(0, 1) and
Z 1 Z 1 1
| f (t)| dt = |t|−1/2 dx = 2t 1/2 = 2 < ∞ ,

0 0 0
so f ∈ L̃1 (0, 1).

We conclude that C[0, 1] is a strict subset of f ∈ L̃1 (0, 1).
1.3 Hamel bases

Definition 1.1 The linear span of a subset E of a vector space V is the collection of
all finite linear combinations of elements of E:
( )
n
Span(E) = x ∈V :x = ∑ α j e j , n ∈ N, α j ∈ K, e j ∈ E .
j=1
We say that E spans V if V = Span(E), i.e. every element of V can be written as a

finite linear combination of elements of E.
3
Definition 1.2 A set E is linearly independent if any finite collection of elements of E
is linearly independent:
n
∑ α je j = 0 =⇒ α1 = α2 = · · · = αn = 0
j=1
for any choice of n ∈ N, e j ∈ E and α j ∈ K.
Definition 1.3 A Hamel basis E for V is a linearly independent subset of V which

spans V .
Examples:
1. Any basis in Rn is a Hamel basis.
2. The set E = 1, x, x2 , . . . is a Hamel basis in the space of all polynomials.

Lemma 1.4 If E is a Hamel basis for a vector space V then any element x ∈ V can be
uniquely written in the form
n
x= ∑ α je j
j=1
where n ∈ N, α j ∈ K and e j ∈ E.
Exercise: Prove the lemma.
Definition 1.5 We say that a set is finite if it consists of a finite number of elements.
Theorem 1.6 If V has a finite Hamel basis then every Hamel basis for V has the same
number of elements.
Definition 1.7 If V has a finite basis E then the dimension of V (denoted dimV ) is
the number of elements in E. If V has no finite basis then we say that V is infinite-
dimensional.
Example: In Rn any basis consists of n vectors. Therefore dim Rn = n.

Let V and W be two vector spaces over K.
Definition 1.8 A map L : V → W is called linear if for any x, y ∈ V and any α ∈ K
L(x + αy) = L(x) + αL(y) .
Definition 1.9 If a linear map L : V → W is a bijection, then L is called a linear

isomorphism. We say that V and W are linearly isomorphic if there is a bijective
linear map L : V → W .
4
Proposition 1.10 Any n-dimensional vector space over K is linearly isomorphic to
Kn .
Proof: If E = { e j : 1 ≤ j ≤ n } is a basis in V , then every element x ∈ V is represented
uniquely in the form
n
x= ∑ α je j .
j=1
The map L : x 7→ (α1 , . . . , αn ) is a linear bijection V → Kn . Therefore V is linearly
isomorphic to Kn .
In order to show that a vector space is infinite-dimensional it is sufficient to find an
infinite linearly independent subset. Let’s consider the following examples:
1. ` p (K) is infinite-dimensional (1 ≤ p ≤ ∞).
Proof. The set
E = { (1, 0, 0, 0, . . .), (0, 1, 0, 0, . . .), (0, 0, 1, 0, . . .), . . .}
is not finite and linearly independent. Therefore dim ` p (K) = ∞.
Remark: This linearly independent subset is not a Hamel basis. Indeed, the
sequence x = (x1 , x2 , x3 , . . .) with xk = e−k belongs to ` p (K) for any p ≥ 1 but
cannot be represented as a sum of finitely many elements of the set E.
2. C[0, 1] is infinite-dimensional.
Proof: The set E = { xk : k ∈ N } is linearly independent subset of C0 [0, 1]: In-
deed, suppose
n
p(x) = ∑ αk x k = 0 for all x ∈ [0, 1].
k=1
Differentiating the equality n times we get p(n) (x) = n!αn = 0. Which implies
αn = 0. Therefore p(x) ≡ 0 implies αk = 0 for all k.
The linearly independent sets provided in the last two examples are not Hamel
bases. This is not a coincidence: we will see later that ` p (K) and C[0, 1] (as well as
many other functional spaces) do not have a countable Hamel basis.
Theorem 1.11 Every vector space has a Hamel basis.
The proof of this theorem is based on Zorn’s Lemma.
We note that in many interesting vector spaces (called normed spaces), a very large
number of elements should be included into a Hamel basis in order to enable repre-
sentation of every element in the form of a finite sum. Then the basis is too large to
be useful for the study of the original vector space. A natural idea would be to allow
infinite sums in the definition of a basis. In order to use infinite sums we need to de-
fine convergence which cannot be done using the axioms of vector spaces only. An
additional structure on the vector space should be defined.
5
2 Normed spaces
2.1 Norms
Definition 2.1 A norm on a vector space V is a map k · k : V → R such that for any
x, y ∈ V and any α ∈ K:
1. kxk ≥ 0, and kxk = 0 ⇔ x = 0 (positive definiteness);
2. kαxk = |α| kxk (positive homogeneity);
3. kx + yk ≤ kxk + kyk (triangle inequality).
The pair (V, k · k) is called a normed space.
In other words, a normed space is a vector space equipped with a norm.

Examples:
1. Rn with each of the following norms is a normed space:

s
n
(a) kxk = ∑ |xk |2
k=1
!1/p
n
(b) kxk p = ∑ |xk | p ,1≤ p<∞
k=1
(c) kxk∞ = max |xk |.
1≤k≤n
2. ` p (K) is a vector space with the following norm (1 ≤ p < ∞)

!1/p
∞
kxk` p = ∑ |xk | p .
k=1
3. `∞ (K) is a vector space with the following norm
kxk`∞ = sup |xk | .

k∈N
In order to prove the triangle inequality for the ` p norm, we will state and prove
several inequalities.
6
2.2 Four famous inequalities
1
Lemma 2.2 (Young’s inequality) If a, b > 0, 1 < p, q < ∞, p + 1q = 1, then
a p bq
ab ≤ + .
p q
p
Proof: Consider the function f (t) = tp −t + 1q defined for t ≥ 0. Since f 0 (t) = t p−1 − 1
vanishes at t = 1 only, and f 00 (t) = (p−1)t p−2 ≥ 0, the point t = 1 is a global minimum
for f . Consequently, f (t) ≥ f (1) = 0 for all t ≥ 0. Now substitute t = ab−q/p :
a p b−q 1
f (ab−q/p ) = − ab−q/p + ≥ 0 .
p q
Multiplying the inequality by bq yields Young’s inequality.
1
Lemma 2.3 (Hölder’s inequality) If 1 ≤ p, q ≤ ∞, p + 1q = 1, x ∈ ` p (K), y ∈ `q (K),
then ∞
∑ |x j y j | ≤ kxk` p kyk`q .
j=1
Proof. If 1 < p, q < ∞, we use Young’s inequality to get that for any n ∈ N
n |x | |y | n
1 |x j | p 1 |y j |q

j j 1 1
∑ kxk` p kyk`q ≤ ∑ p kxk pp + q kykqq ≤ p + q = 1
j=1 j=1 ` `
Therefore for any n ∈ N

n
∑ |x j y j | ≤ kxk` p kyk`q .
j=1
Since the partial sums are monotonically increasing and bounded above, the series
converge and Hölder’s inequality follows by taking the limit as n → ∞.
If p = 1 and q = ∞:
n n
∑ |x j y j | ≤ max |y j |
1≤ j≤n
∑ |x j | ≤ kxk`1 kyk`∞ .
j=1 j=1
Therefore the series converges and Hölder’s inequality follows by taking the limit as
n → ∞.
Lemma 2.4 (Cauchy-Schwartz inequality) If x, y ∈ `2 (K) then

!1/2 !1/2
∞ ∞ ∞
∑ |x j y j | ≤ ∑ |x j |2 ∑ |y j |2 .
j=1 j=1 j=1
7
Proof: This inequality coincides with Hölder’s inequality with p = q = 2.
Now we state and prove the triangle inequality for the ` p norm.
Lemma 2.5 (Minkowski’s inequality) If x, y ∈ ` p (K) for 1 ≤ p ≤ ∞ then x + y ∈

` p (K) and
kx + yk` p ≤ kxk` p + kyk` p .
Proof: If 1 < p < ∞, define q from 1p + 1q = 1. Then using Hölder’s inequality (finite
sequences belong to ` p with any p) we get2
n n
∑ |x j + y j | p = ∑ |x j + y j | p−1|x j + y j |
j=1 j=1
n n
≤ ∑ |x j + y j | p−1|x j | + ∑ |x j + y j | p−1|y j |
j=1 j=1
!1/q !1/p
n n
≤ ∑ |x j + y j |(p−1)q ∑ |x j | p (Hölder’s inequality)
j=1 j=1
!1/q !1/p
n n
+ ∑ |x j + y j |(p−1)q ∑ |y j | p .
j=1 j=1
1/q
Dividing the inequality by ∑nj=1 |x j + y j | p and using that (p − 1)q = p and 1 −
1
q = 1p , we get for all n
!1/p !1/p !1/p
n n n
∑ |x j + y j | p ≤ ∑ |x j | p + ∑ |y j | p .
j=1 j=1 j=1
The series on the right hand side converge to kxk` p + kyk` p . Consequently the series
on the left hand side also converge, x + y ∈ ` p (K), and Minkowski’s inequality follows
by taking the limit as n → ∞.
Exercise: Prove Minkowski’s inequality for p = 1 and p = ∞.
2.3 Examples of norms on a space of functions

Each of the following formulae defines a norm on C[0, 1], the space of all continuous
functions on [0, 1]:
2 We do not start directly with n = ∞ because a priori we do not know convergence for some of the
series involved in the proof.
8
1. the “sup(remum) norm”
k f k∞ = sup | f (t)| ;
t∈[0,1]
2. the “L1 norm” Z 1

k f kL1 = | f (t)| dt ;
0
3. the “L2 norm” 1/2

1
Z
2
k f kL 2 = | f (t)| dt .
0
Exercise: Check that each of these formulae defines a norm. For the case of the L2
norm, you will need a Cauchy-Schwartz inequality for integrals.
Example: Let k ∈ N. The space Ck [0, 1] consists of all continuous real-valued func-
tions which have continuous derivatives up to order k. The norm on Ck [0, 1] is defined
by
k
k f kCk = ∑ sup | f ( j) (t)| ,
j=0 t∈[0,1]
where f ( j) denotes the derivative of order j.
2.4 Equivalence of norms

We have seen that various different norms can be introduced on a vector space. In order
to compare different norms it is convenient to introduce the following equivalence
relation.
Definition 2.6 Two norms k · k1 and k · k2 on a vector space V are equivalent if there
are constants c1 , c2 > 0 such that
c1 kxk1 ≤ kxk2 ≤ c2 kxk1 for all x ∈ V .
In this case we write k · k1 ∼ k · k2 .
Theorem 2.7 Any two norms on Rn are equivalent.3
Example: The norms k · kL1 and k · k∞ on C[0, 1] are not equivalent.

3 You already saw this statement in Analysis III and Differentiation in Year 2. The proof is based
on the observation that the unit sphere S ⊂ Rn is sequentially compact. Then we checked that f (x) =
kxk2 /kxk1 is continuous on S and consequently it is bounded and attains its lower and upper bounds on
S. We set c1 = minS f and c2 = maxS f .
9
Proof: Consider the sequence of functions fn (t) = t n with n ∈ N. Obviously fn ∈
C[0, 1]. We see that
k fn k∞ = max |t|n = 1 ,
t∈[0,1]
Z 1
n 1
k fn kL1 = t dt = .
0 n+1
Suppose the norms are equivalent. Then there is a constant c2 > 0 such that for all fn :
k f n k∞
= n + 1 ≤ c2 ,
k f n kL 1
which is not possible for all n. This contradiction implies the norms are not equivalent.

2.5 Linear Isometries

Suppose V and W are normed spaces.
Definition 2.8 If a linear map L : V → W preserves norms, i.e. kL(x)k = kxk for all
x ∈ V , it is called a linear isometry.
This definition implies L is injective, i.e., L : V → L(V ) is bijective, but it does not im-
ply L(V ) = W , i.e., L is not necessarily invertible. Note that sometimes the invertibility
property is included into the definition of the isometry. Finally, in Metric Spaces the
word “isometry” is used to denote distance-preserving transformations.
Definition 2.9 We say that two normed spaces are isometrically isomorphic (or simply
isometric), if there is an invertible linear isometry between them.
A linear invertible map can be used to “pull back” a norm as follows.
Proposition 2.10 Let (V, k · kV ) be a normed space, W a vector space, and L : W → V

a linear isomorphism. Then
kxkW := kL(x)kV
defines a norm on W .
Proof: For any x, y ∈ V and any α ∈ K we have:
kxkW = kL(x)kV ≥ 0 ,
kαxkW = kL(αx)kV = |α| kL(x)kV = |α| kxkW .
If kxkW = kL(x)kV = 0, then L(x) = 0 due to non-degeneracy of the norm k · kV . Since

L is invertible, we get x = 0. Therefore k · kW is non-degenerate.
10
Finally, the triangle inequality follows from the triangle inequality for k · kV :
kx + ykW = kL(x) + L(y)kV ≤ kL(x)kV + kL(y)kV = kxkW + kykW .
Therefore, k · kW is a norm.
Note that in the proposition the new norm is introduced in such a way that L :
(W, k · kW ) → (V, k · kV ) is a linear isometry.
Let V be a finite dimensional vector space and n = dimV . We have seen that V is
linearly isomorphic to Kn . Then the proposition implies the following statements.
Corollary 2.11 Any finite dimensional vector space V can be equipped with a norm.
Corollary 2.12 Any n-dimensional normed space V is isometrically isomorphic to Kn

equipped with a suitable norm.
Since any two norms on Rn (and therefore on Cn ) are equivalent we also get the
following statement.
Theorem 2.13 Let V be a finite-dimensional vector space. Then all norms on V are
equivalent.
11
3 Convergence in a normed space
3.1 Definition and examples
The norm on a vector space V can be used to measure distances between points x, y ∈ V .
So we can define the limit of a sequence.
n=1 , xn ∈ V , n ∈ N, converges to a limit x ∈ V if for any

Definition 3.1 A sequence (xn )∞
ε > 0 there is N ∈ N such that
kxn − xk < ε for all n > N.
Then we write xn → x.
We note that the sequence of vectors xn → x if and only if the sequence of non-
negative real numbers kxn − xk → 0.
Proposition 3.2 The limit of a convergent sequence is unique.
Exercise: Prove it.
Proposition 3.3 Any convergent sequence is bounded.
Exercise: Prove it.
Proposition 3.4 If xn converges to x, then kxn k → kxk.
Exercise: Prove it.

It is possible to check convergence of a sequence of real numbers without actually
finding its limit: it is sufficient to check that it satisfies the following definition:
Definition 3.5 (Cauchy sequence) A sequence (xn )∞

n=1 in a normed space V is Cauchy
if for any ε > 0 there is an N such that
kxn − xm k < ε for all m, n > N.
Theorem 3.6 A sequence of real numbers converges iff it is Cauchy.
Proposition 3.7 Any convergent sequence is Cauchy.
Exercise: Prove it.
Proposition 3.8 Any Cauchy sequence is bounded.
Exercise: Prove it.
12
Example: Consider the sequence fn ∈ C[0, 1] defined by fn (t) = t n .
1. fn → 0 in the L1 norm.
Proof: We have already computed the norms:
1
k fn kL1 = → 0.
n+1
Consequently, fn → 0.
2. fn does not converge in the sup norm.
Proof: If m > 2n ≥ 1 then
1 1 1
fn (2−1/n ) − fm (2−1/n ) = − m/n ≥ .
2 2 4
Consequently ( fn ) is not Cauchy in the sup norm and hence not convergent.
This example shows that the convergence in the L1 norm does not imply the point-
wise convergence and, as a results, does not imply the convergence in the sup norm
(often called the uniform convergence). Note that in contrast to the uniform and L1
convergences the notion of pointwise convergence is not based on a norm on the space
of continuous function.
Exercise: The pointwise convergence does not imply the L1 convergence.
Hint: Construct fn with a very small support but make the maximum of fn very
large to ensure that k fn kL1 > n. Therefore fn is not bounded in the L1 norm, hence not
convergent.
We can also make supp fn ∩ supp fm = 0/ for all m, n such that n 6= m. Then for any
t there is at most one n such that fn (t) 6= 0. The last property guarantees pointwise
convergence: fn (t) → 0 for any t.
Proposition 3.9 If fn ∈ C[0, 1] for all n ∈ N and fn → f in the sup norm, then fn → f
in the L1 norm, i.e.,
k fn − f k∞ → 0 =⇒ k fn − f kL1 → 0 .
Proof:
Z 1
0 ≤ k f n − f kL 1 = | fn (t) − f (t)| dt ≤ sup | fn (t) − f (t)| = k fn − f k∞ → 0 .
0 0≤t≤1
Therefore k fn − f kL1 → 0.
We have seen that different norms may lead to different conclusions about conver-
gence of a given sequence but sometime convergence in one norm implies convergence
in another one. The following lemma shows that equivalent norms give rise to the same
notion of convergence.
13
Lemma 3.10 Suppose k · k1 and k · k2 are equivalent norms on a vector space V . Then
for any sequence (xn ):
kxn − xk1 → 0 ⇔ kxn − xk2 → 0 .
Proof: Since the norms are equivalent, there are constant c1 , c2 > 0 such that
0 ≤ c1 kxn − xk1 ≤ kxn − xk2 ≤ c2 kxn − xk1
for all n. Then kxn − xk2 → 0 implies kxn − xk1 → 0, and vice versa.
3.2 Topology on a normed space

We say that a collection T of subsets of V is a topology on V if it satisfies the following
properties:
/ ∈T ;
1. 0,V
2. any finite intersection of elements of T belongs to T ;
3. any union of elements of T belongs to T .
A set equipped with a topology is called a topological space. The elements of T

are called open sets. The topology can be used to define a convergent sequence and
continuous function.
We note that each norm on V can be used to define a topology on V , i.e., to define
the notion of an open set.
Definition 3.11 A subset X ⊂ V is open, if for any x ∈ X there is ε > 0 such that the
ball of radius ε centred around x belongs to X:
B(x, ε) = {y ∈ V : ky − xk < ε} ⊂ X .
Example: In any normed space V :
1. The unit ball centred around the zero, B0 = { x : kxk < 1 }, is open.
2. Any open ball B(x, ε) is open.
3. V is open.
4. The empty set is open.
It is not too difficult to check that the collection of open sets defines a topology
on V . You can easily check from the definition that equivalent norms generate the
same topology, i.e., open sets are exactly the same. The notion of convergence can be
defined in terms of the topology.
14
Definition 3.12 An open neighbourhood of x is an open set which contains x.
Lemma 3.13 A sequence xn → x if and only if for any open neighbourhood X of x

there is N ∈ N such that xn ∈ X for all n > N.
Proof: ( =⇒ ). Let xn → x. Take any open X such that x ∈ X. Then there is ε > 0 such
that B(x, ε) ⊂ X. Since the sequence converges there is N such that kxn − xk < ε for
all n > N. Then xn ∈ B(x, ε) ⊂ X for the same values of n.
(⇐=). Take any ε > 0. The ball B(x, ε) is open, therefore there is N such that
xn ∈ B(x, ε) for all n > N. Hence kxn − xk < ε and xn → x.
3.3 Closed sets

Definition 3.14 A set X ⊂ V is closed if its complement V \ X is open.
Example: In any normed space V :

1. The unit sphere S = { x : kxk = 1 } is closed
2. Any closed ball
B(x, ε) = { y ∈ V : ky − xk ≤ ε }
is closed.
3. V is closed.
4. The empty set is closed.
Lemma 3.15 A subset X ⊂ V is closed if and only if any convergent sequence in X

has its limit in X.
Proof: The proof literally repeats the proof given in Differentiation.
Definition 3.16 We say that a subset L ⊂ V is a linear subspace, if it is a vector space

itself, i.e., if x1 , x2 ∈ L and λ ∈ K imply x1 + λ x2 ∈ L.
Proposition 3.17 Any finite dimensional linear subspace L of a normed space V is

closed.
Proof: Since L is finite-dimensional, it has a finite Hamel basis
E = { e1 , e2 , . . . , en } ⊂ L
such that L = Span(E). Suppose L is not closed, then by Lemma 3.15 there is a
convergent sequence xk → x∗ , xk ∈ L but x∗ ∈ V \ L. Then x∗ is linearly independent
from E (otherwise it would belong to L). Consequently
Ẽ = { e1 , e2 , . . . , en , x∗ }
15
is a Hamel basis in L̃ = Span(Ẽ). In this basis, the components of xk are given by
(α1k , . . . , αnk , 0) and x∗ corresponds to the vector (0, . . . , 0, 1). We get in the limit as
k→∞
(α1k , . . . , αnk , 0) → (0, . . . , 0, 1) ,
which is obviously impossible. Therefore L is closed.4
Example: The subspace of polynomial functions is linear but not closed in C[0, 1]
equipped with the sup norm.
3.4 Compactness
Definition 3.18 (sequential compactness) A subset K of a normed space (V, k · kV ) is
n=1 with xn ∈ K has a convergent subse-
(sequentially) compact if any sequence (xn )∞
quence xn j → x∗ with x∗ ∈ K.
Proposition 3.19 A compact set is closed and bounded.
Theorem 3.20 A subset of Rn is compact iff it is closed and bounded.
Corollary 3.21 A subset of a finite-dimensional vector space is compact iff it is closed

and bounded.
Example: The unit sphere in ` p (K) is closed, bounded but not compact.
Proof: Take the sequence e j such that
e j = (0, . . . , 0, |{z}
1 , 0, . . .) .
jth place
We note that ke j − ek k` p = 21/p for all j 6= k. Consequently, (e j )∞j=1 does not have any
convergent subsequence, hence S is not compact.
Lemma 3.22 (Riesz’ Lemma) Let X be a normed vector space and Y be a closed
linear subspace of X such that Y 6= X and α ∈ R, 0 < α < 1. Then there is xα ∈ X
such that kxα k = 1 and kxα − yk > α for all y ∈ Y .
Proof: Since Y ⊂ X and Y 6= X there is x ∈ X \Y . Since Y is closed, X \Y is open and

therefore
d := inf{ kx − yk : y ∈ Y } > 0 .
4 Thisproof implicitly uses equivalence of norms on Rn+1 to establish that convergence in the norm
obtained by restricting the original norm k · kV onto L̃ implies convergence of the components of the
vectors.
16
Since α −1 > 1 there is a point z ∈ Y such that kx − zk < dα −1 . Let xα = kx−zk
x−z
. Then
kxα k = 1 and for any y ∈ Y ,

x−z x − (z + kx − zky) d
kxα − yk = kx − zk − y =
> =α,
kx − zk dα −1
as z + kx − zky ∈ Y because Y is a linear subspace.
Theorem 3.23 A normed space is finite dimensional iff the unit sphere is compact.
Proof: Bolzano-Weierstrass Theorem and Lemma 3.15 imply that in a finite dimen-
sional normed space the unit sphere is compact (the unit sphere is bounded and closed).
So we only need to show that if the unit sphere S ⊂ B is sequentially compact, then
the normed space V is finite dimensional. Indeed, if V is infinite dimensional, then
Riesz’ Lemma can be used to construct an infinite sequence of xn ∈ S such that kxn −
xm k > α > 0 for all m 6= n. This sequence does not have a convergent subsequence
(none of the subsequences is Cauchy) and therefore S is not compact.
We construct xn inductively. Fix α ∈ (0, 1) and take any x1 ∈ S.
Suppose that for some n ≥ 1 we have found En = { x1 , . . . , xn } such that xk ∈ S and
kxl − xk k > α for all 1 ≤ k, l ≤ n, k 6= l (note that the second property is automatically
fulfilled for n = 1). The linear subspace Yn = Span(En ) is n-dimensional and hence
closed (see Proposition 3.17). Since X is infinite dimensional Yn 6= X. Then Riesz’
Lemma implies that there is xn+1 ∈ S such that kxn+1 − xk k > α for all 1 ≤ k ≤ n.
Repeating this argument we generate xn for all n ∈ N.
17
4 Banach spaces
4.1 Completeness: Definition and examples
Definition 4.1 (Banach space) A normed space V is called complete if any Cauchy
sequence in V converges to a limit in V . A complete normed space is called a Banach
space.
Theorem 4.2 Every finite-dimensional normed space is complete.
Proof: Theorem 3.6 implies that R is complete, i.e., every Cauchy sequence of num-
bers has a limit.
Now let V be a real vector space, dimV = n < ∞. Take any basis in V . Then
a sequence of vectors in V converges iff each component of the vectors converges,
and a sequence of vectors is Cauchy iff each component is Cauchy. Therefore each
component has a limit, and those limits constitute the limit vector for the original
sequence. Hence V is complete.
Considering C as a real vector space we conclude that it is also complete. There-
fore, any finite-dimensional complex vector space V is also complete.
In particular, Rn and Cn are complete.
Theorem 4.3 (` p is a Banach space) The space ` p (K) equipped with the standard ` p
norm is complete.
Proof: Suppose that xk = (x1k , x2k , . . .) ∈ ` p (K) is Cauchy. Then for every ε > 0 there is
N such that
∞
kxm − xn k` p = ∑ |xmj − xnj| p < ε
j=1
for all m, n > N. Consequently, for each j ∈ N the sequence xkj is Cauchy, and the
completeness of K implies that there is a j ∈ K such that
xkj → a j
as k → ∞. Let a = (a1 , a2 , . . .). First we note that for any M ≥ 1 and m, n > N:
M ∞
∑ |xmj − xnj| p ≤ ∑ |xmj − xnj| p < ε .
j=1 j=1
Taking the limit as n → ∞ we get

M
∑ |xmj − a j | p ≤ ε .
j=1
18
This holds for any M, so we can take the limit as M → ∞:
∞
∑ |xmj − a j | p ≤ ε .
j=1
We conclude that xm − a ∈ ` p (K). Since ` p (K) is a vector space and xm ∈ ` p (K), then
a ∈ ` p (K). Moreover, kxm − ak` p < ε for all m > N. Consequently xm → a in ` p (K)
with the standard norm, and so ` p (K) is complete.
Theorem 4.4 (C is a Banach space) The space C[0, 1] equipped with the sup norm is
complete.
Proof: Let fk be a Cauchy sequence. Then for any ε > 0 there is N such that
sup | fn (t) − fm (t)| < ε

t∈[0,1]
for all m, n > N. In particular, fn (t) is Cauchy for any fixed t and consequently has a
limit. Set
f (t) = lim fn (t) .
n→∞
Let’s prove that fn (t) → f (t) uniformly in t. Indeed, we already know that
| fn (t) − fm (t)| < ε
for all n, m > N and all t ∈ [0, 1]. Taking the limit as m → ∞ we get
| fn (t) − f (t)| ≤ ε
for all n > N and all t ∈ [0, 1]. Therefore fn converges uniformly:
k fn − f k∞ = sup | fn (t) − f (t)| < ε .

t∈[0,1]
for all n > N. The uniform limit of a sequence of continuous functions is continuous.
Consequently, f ∈ C[0, 1] which completes the proof of completeness.
Example: The space C[0, 2] equipped with the L1 norm is not complete.
Proof: Consider the following sequence of functions:
( n
t for 0 ≤ t ≤ 1,
fn (t) =
1 for 1 ≤ t ≤ 2 .
This is a Cauchy sequence in the L1 norm. Indeed for any n < m:

Z 1
1 1 1
k f n − f m kL 1 = (t n − t m ) dt = − < ,
0 n+1 m+1 n+1
19
and consequently for any m, n > N
1
k fn − fm kL1 < .
N
Now let us show that fn do not converge to a continuous function in the L1 norm.
Indeed, suppose such a limit exists and call it f . Then
Z 1 Z 2
n
k fn − f kL1 = |t − f (t)| dt + |1 − f (t)| dt → 0 .
0 1
Since
| f (t)| − |t n | ≤ |t n − f (t)| ≤ | f (t)| + |t n |
implies that
Z 1 Z 1 Z 1 Z 1 Z 1
n n
| f (t)| dt − t dt ≤ |t − f (t)| dt ≤ | f (t)| dt + t n dt ,
0 0 0 0 0
we have 01 |t n − f (t)| dt → 01 | f (t)| dt as n → ∞ and consequently

R R
Z 1 Z 2
| f (t)| dt + |1 − f (t)| dt = 0 .
0 1
As f is assumed to be continuous, it follows

0, 0 < t < 1,
f (t) =
1, 1 < t < 2.
We see that the limit function f cannot be continuous. This contradiction implies that
C[0, 2] is not complete with respect to the L1 norm.
4.2 The completion of a normed space

A normed space may be incomplete. However, every normed space X can be con-
sidered as a subset of a larger Banach space X̂. The minimal among these spaces5 is
called the completion of X.
Informally we can say that X̂ consists of limit points of all Cauchy sequences in X.
Of course, every point x ∈ X is a limit point of the constant sequence (xn = x for all
n ∈ N) and therefore X ⊂ X̂. If X is not complete, some of the limit points are not in
X, so X̂ is larger then the original set X.
Definition 4.5 (dense set) We say that a subset X ⊂ V is dense in V if for any v ∈ V
and any ε > 0 there is x ∈ X such that kx − vk < ε.
5 In this context “minimal” means that if any other space X̃ has the same property, then the minimal
X̂ is isometric to a subspace of X̃. It turns out that this property can be achieved by requiring X to be
dense in X̂.
20
Note that X is dense in V iff for every point v ∈ V there is a sequence xn ∈ X such
that xn → v.
Theorem 4.6 Let (X, k · kX ) be a normed space. Then there is a complete normed
space (X , k · kX ) and a linear map i : X → X such that i is an isometrical isomor-
phism between (X, k · kX ) and (i(X), k · kX ), and i(X) is dense in X .
Moreover, X is unique up to isometry, i.e., if there is another complete normed
space (X˜ , k·kX˜ ) with these properties, then X and X˜ are isometrically isomorphic.
Proof: The proof is relatively long so we break it into a sequence of steps.

Construction of X . Let Y be the set of all Cauchy sequences in X. We say that two
Cauchy sequences x = (xn )∞ n=1 , xn ∈ X, and y = (yn )n=1 , yn ∈ X, are equivalent, and
∞
write x ∼ y, if
lim kxn − yn kX = 0 .
n→∞
Let X be the space of all equivalence classes in Y , i.e., it is the factor space: X =
Y / ∼. The elements of X are collections of equivalent Cauchy sequences from X.
We will use [x] to denote the equivalence class of x.
Exercises: Show that X is a vector space.
Norm on X . For an η ∈ X take any representative x = (xn )∞ n=1 , xn ∈ X, of the
equivalence class η. Then the equation
kηkX = lim kxn kX . (4.1)

n→∞
defines a norm on X . Indeed:

1. Equation (4.1) defines a function X → R, i.e., for any η ∈ X and any repre-
sentative x ∈ η the limit exists and is independent from the choice of the repre-
sentative. (Exercise)
2. The function defined by (4.1) satisfies the axioms of norm. (Exercise)
Definition of i : X → X . For any x ∈ X let
i(x) = [(x, x, x, x, . . .)]
(the equivalence class of the constant sequence). Obviously, i is a linear isometry,

and it is a bijection X → i(X). Therefore the spaces X and i(X) are isometrically
isomorphic.
∞
Completeness of X . Let η (k) be a Cauchy sequence in (X , k · kX ). For every
k=1
k ∈ N take a representative x(k) ∈ η (k) . Note that x(k) ∈ Y is a Cauchy sequence in the
space (X, k · kX ). Then there is a strictly monotone sequence of integers nk such that

(k)

(k) 1
x j − xl ≤ for all j, l ≥ nk . (4.2)
X k
21
Now consider the sequence x∗ defined by
∞
∗ (k)
x = xnk .
k=1
Next we will check that x∗ is Cauchy, and consider its equivalence class η ∗ = [x∗ ] ∈
X . Then we will prove that η (k) → η ∗ in (X , k · kX ).
The sequence x∗ is Cauchy. Since the sequence of η (k) is Cauchy, for any ε > 0 there
is Mε such that
(k) (l)
lim kxn − xn kX = kη (k) − η (l) kX < ε for all k, l > Mε .
n→∞
Consequently, for every k, l > Mε there is Nεk,l such that

(k) (l)
kxn − xn kX < ε for all n > Nεk,l . (4.3)
3 k,l
Then fix any ε > 0. If j, l > ε and m > max{n j , nl , Nε/3 } we have
( j) (l)
kx∗j − xl∗ kX = kxn j − xnl kX
( j) ( j) ( j) (l) (l) (l)
≤ kxn j − xm kX + kxm − xm kX + kxm − xnl kX
1 ε 1
< + + <ε
j 3 l
where we used (4.3) and (4.2). Therefore x∗ is Cauchy and η = [x∗ ] ∈ X .
The sequence η (k) → [x∗ ]. Indeed, take any ε > 0 and k > 3ε −1 , then
(k) (k) ( j)
kη (k) − η ∗ kX = lim kx j − x∗j kX = lim kx j − xn j kX
j→∞ j→∞

(k) (k) (k) ( j)
≤ lim kx j − xnk kX + kxnk − xn j kX
j→∞
1
≤ + ε < 2ε
k
Therefore η (k) → η ∗ .
We have proved that any Cauchy sequence in X has a limit in X , so X is com-
plete.
Density of i(X) in X . Take an η ∈ X and let x ∈ η. Take any ε > 0. Since x is
Cauchy, there is Nε such that kxm − xk kX < ε for all k, m > Nε . Then
kη − i(xk )kX = lim kxm − xk kX ≤ ε.
m→∞
Therefore i(X) is dense in X .
Uniqueness of X up to isometry. I will not show you the details of the proof. The
proof uses the fact that i(X) is isometrically isomorphic to ĩ(X) (since each one is
isometrically isomorphic to X). Moreover, i(X) is dense in X and ĩ(X) is dense in
X˜ . In order to complete the proof one has to show that an isometry between two
dense subsets can be extended to an isometry between the sets.
22
4.3 Examples
The theorem provides an explicit construction for the completion of a normed space.
Often this description is not sufficiently convenient and a more direct description is
desirable.
1. Example: Consider the space P[0, 1] of all polynomial functions restricted to

the interval [0, 1] and equip this space with the sup norm. This space is not com-
plete. On the other hand any polynomial is continuous, and therefore P[0, 1]
can be considered as a subspace of C[0, 1] which is complete. The Weierstrass
approximation theorem states that any continuous function on [0, 1] can be uni-
formly approximated by polynomials. In other words, the polynomials are dense
in C[0, 1]. Then Theorem 4.6 implies that the completion of P[0, 1] is isometri-
cally isomorphic to C[0, 1] equipped with the sup norm.
2. Example: Let ` f (K) be the space of all sequences which have only a finite
number of non-zero elements. This space is not complete in the ` p norm. The
completion of ` f (K) in the ` p norm is isometric to ` p (K).
Indeed, we have already seen that ` p (K) is complete. So in order to prove the
claim you only need to check that ` f (K) is dense in ` p (K).
We see that the completion of a space depends both on the space and on the
norm.
3. Example: L1 (0, 1) is the completion of C[0, 1] in the L1 norm.

According to this definition any f ∈ L1 (0, 1) is an equivalence class of a Cauchy
sequence fn ∈ C[0, 1] equipped with the L1 norm. The norm of f is defined by
Z 1
k f kL1 = lim k fn kL1 = lim | fn (t)| dt .
n→∞ n→∞ 0
In spite of the fact that f can be considered as a limit of the sequence of continu-
ous functions fn in the L1 norm, we cannot define the value of f (t) for a given t
as the limit of fn (t), because the limit may not exist or may depend on the choice
of the representative in the equivalence class.
In spite of that we can define the integral of f by setting
Z 1 Z 1
f (t) dt := lim fn (t) dt . (4.4)
0 n→∞ 0
Indeed, the limit exists: the sequence of the integrals is Cauchy as

Z 1 Z 1 Z 1

fn (t) dt − fm (t) dt = fn (t) − fm (t) dt
0 0 0

Z 1
≤ | fn (t) − fm (t)| dt = k fn − fm kL1 ,
0
23
hence convergent. It is not difficult to check that the limit does not depend from
the choice of a representative in the equivalence class f .
Obviously, if fn is a constant sequence, i.e., fn (t) = f0 (t), t ∈ [0, 1], for all n ∈
N, then 01 f (t) = 01 f0 (t) dt. Therefore this definition can be considered as an
R R
extension of the classical Riemann’s integral used for continuous functions.

Taking into account the new definition of the integral we can write
Z 1
k f kL1 = | f (t)| dt .
0
Then we recall that the norm is non-degenerate, therefore f = g for f , g ∈

L1 (0, 1) if and only if k f − gkL1 = 0, i.e.,
Z 1
| f (t) − g(t)| dt = 0 ,
0
where the integral should be interpreted in the sense of the new definition. Note
that f and g are not continuous and consequently the equality above cannot be
used to deduce that f (t) = g(t) for all t.
Nevertheless a more direct description of L1 (0, 1) is possible: The space L1 (0, 1)

is isometrically isomorphic to the space of equivalence classes of Lebesgue in-
tegrable functions on (0, 1): the functions f and g are equivalent if
Z 1
| f (t) − g(t)| dt = 0 .
0
Here the integral should be considered as the Lebesgue integral. We note that it
coincides with the definition provided above but its construction is more direct.
Since the notion of the Lebesgue integration is very important for the functional
analysis and its applications we will discuss it in more details in the next few
lectures. A more detailed study of this topics is a part of MA359 Measure Theory
module.
24
5 Lebesgue spaces
The exposition of the Lebesgue integral is based on the book H.A.Priesly, Introduction
to integration, Oxford Sc.Publ., 1997, 306 p.
5.1 Integrable functions

Integrals of step functions
We say that ϕ : R → R is a step function if it is piecewise constant on a finite number
of intervals, i.e., it can be represented as a finite sum
n
ϕ(x) = ∑ c j χI j (x) ,
j=1
where c j ∈ R, I j ⊂ R is an interval, I j ∩ Ik = 0/ if k 6= j, and χI is the characteristic

function of I:
1, x ∈ I ,
χI (x) =
0, x 6∈ I .
We note that the intervals are allowed to be of any of the four possible types (e.g.
(a, b), [a, b), (a, b] or [a, b]).
We define the integral of a step function ϕ by
Z n
ϕ := ∑ c j |I j |,
j=1
where |I j | is the length of I j . We Rnote that this sum equals to the Riemann integral
which you studied in Year 1, i.e., ϕ is the “algebraic” area under the graph of the
step function ϕ (the area is counted negative on those intervals where ϕ(x) < 0).
Sets of measure zero

Definition 5.1 We say that a set A ⊂ R has measure zero if for any ε > 0 there is an
(at most countable) collection of intervals that cover A and whose total length is less
than ε:
∞
[ ∞
A⊂ [a j , b j ] and ∑ (b j − a j ) < ε .
j=1 j=1
Exercise: Show that a countable union of measure zero sets has measure zero. Hint:
for An choose a cover with εn = ε/2n .
Examples. The set Q of all rational numbers has measure zero. The Cantor set has
measure zero.
Definition 5.2 A property is said to hold “almost everywhere” or “for almost every
x” (and abbreviated to “a.e.”) if the set of points at which the property does not hold
has measure zero.
25
Almost everywhere convergence
Theorem 5.3 Let (ϕn (x))R∞ n=1 be an increasing sequence of step functions (ϕn+1 (x) ≥
ϕn (x) for all x) such that ϕn < K. Then ϕn (x) converges for a.e. x.
Proof: First note that an increasing sequence of numbers has a limit if and only if it is
bounded from above. So in order to prove the theorem it is sufficient to show that the
set
E = { x ∈ R : ϕn (x) → +∞ }
has measure zero.
Without loosing in generality, we can assume that ϕn (x) ≥ 0.6 Let us define the set
En,m = { x : ϕn (x) > m } .

(n)
This set is a finite union of intervals. Indeed, ϕn (x) = ∑ j c j χI (n) (x) is a step function.
j
(n)
Let In,m = { j : cj > m }. Then
[ (n)
En,m = Ij .
j∈In,m
(n)
The total length of those intervals is less than K/m. Indeed, since c j ≥ 0 for all j and
(n)
c j > n for j ∈ In ,
Z
(n) (n) (n)
K> ϕn = ∑ c j |I j | > m ∑ |I j | .
j j∈In,m
Finally, E ⊂ Em = ∞ l=1 El,m for every m. Since the sequence ϕn is increasing, En,m ⊂
S
En+1,m . Moreover En+1,m \ En,m consists of a finite number of intervals. Then

∞
[
Em = El,m \ El−1,m
l=1
/ Since En,m = nl=1 El,m \

S
is at most countable union of intervals (we denote E0,m = 0).
El−1,m and the total length of the intervals in En,m is less than K/m for all n, the total
length of the intervals in Em is not larger than K/m.
Since Em is at most countable union of intervals whose total length is less than
K/m and E ⊂ Em for all m, the set E has measure zero.
6 Otherwise replace ϕn by ϕn − ϕ1 which is non-negative. The new sequence satisfies the assumption
of the theorem (possibly with a different constant K). Moreover, the convergence of ϕn (x) − ϕ1 (x) is
equivalent to the convergence of ϕn (x).
26
Lemma 5.4 If ϕn and ψn are two increasing sequences of step functions which respec-
tively tend to f and g a.e. and f (x) ≥ g(x) a.e., then
Z Z
lim ϕn ≥ lim ψn .
n→∞ n→∞
Proof: The following sets

n o
E1 = x : lim ϕn (x) 6= f (x) ,
n o
E2 = x : lim ψn (x) 6= g(x) ,
n o
E3 = x : f (x) < g(x)
have measure zero. Let E = E1 ∪ E2 ∪ E3 . Assume x 6∈ E.

Fix arbitrary k ∈ N. The sequence ψk (x) − ϕn (x) is decreasing in n. Then as n → ∞
ψk (x) − ϕn (x) → ψk (x) − f (x) ≤ g(x) − f (x) ≤ 0 .
Consequently
(ψk − ϕn )+ (x) = max{ ψk (x) − ϕn (x), 0 } → 0
Since it is a decreasing sequence of non-negative step functions which converges to 0
a.e.,7 Z
(ψk − ϕn )+ → 0 .
Since ψk and ϕn are step functions

Z Z Z Z
ψk − ϕn = (ψk − ϕn ) ≤ (ψk − ϕn )+ .
Then taking the limit as n → ∞ we get

Z Z
ψk ≤ lim ϕn
n→∞
Finally, take the limit as k → ∞ to obtain the desired inequality.
Corollary 5.5 If ϕn and ψn are two increasing sequences of step functions which tend
to a function f a.e., then Z Z
lim ϕn = lim ψn .
n→∞ n→∞
Proof: Use the previous lemma with g = f .

7 This statement is provided without a proof.
27
Lebesgue integrable functions
Definition 5.6 If a function f : R → R can be represented as an a.e. limit of an in-
creasing sequence of step functions ϕn , then the integral of f is given by
Z Z
f = lim ϕn .
n→∞
If the limit is finite we write f ∈ Linc (R).
Note that the limit is independent of the choice of the increasing sequence of step
functions.
Unfortunately Linc (R) is not a vector space as f ∈ Linc (R) does not imply − f ∈
inc
L (R). Indeed, f is bounded from below by ϕ1 but not necessarily bounded from
above. Then − f is not bounded from below and therefore − f 6∈ Linc (R). For example,
( 1
√ , x 6= 0 , ∞
f (x) = |x| or f = ∑ k1/2 χ[(k+1)−1 ,k−1 ] .
0, x = 0, k=1
Definition 5.7 (Lebesgue integral) A function f : R → R is (Lebesgue) integrable if

f = g − h where g, h ∈ Linc (R) and
Z Z Z
f := g− h.
This integral is called the Lebesgue integral.8
Of course in the definition the choice of g and h is not unique. So we will have to
check that the value of the integral does not depend from this freedom.
5.2 Properties of Lebesgue integrals

The properties of the Lebesgue integral are based on the properties of the integrals for
the functions from Linc (R) .
Proposition 5.8 If f , g ∈ Linc (R) and λ ∈ R, λ ≥ 0, then
1. f + g, λ f , max{ f , g }, min{ f , g } ∈ Linc (R) .

R R R
2. ( f + g) = f + g.
R R
3. If additionally f (x) ≥ g(x) a.e., then f ≥ g.
8 This construction is different (but equivalent) to the original definition of the Lebesgue integral.
For an alternative approach see Measure Theory module.
28
Exercise: Prove the proposition.
First we state the main elementary properties of the Lebesgue integration.
Theorem 5.9 If f , f1 , f2 are integrable and λ ∈ R, then

R R R
1. f1 + λ f2 is also integrable and ( f1 + λ f2 ) = f1 + λ f2 .
R R
2. | f (x)| is also integrable and | f | ≤ | f |.
R
3. If additionally f (x) ≥ 0 a.e., then f ≥ 0.
Proof:
1. Since f1 , f2 are integrable, there are functions g1 , g2 , h1 , h2 ∈ Linc (R) such that
fk = gk − hk , k = 1, 2. If λ ≥ 0 then g1 + λ g2 , h1 + λ h2 ∈ Linc (R) and
Z Z Z
( f1 + f2 ) = (g1 + λ g2 ) − (h1 + λ h2 )
Z Z Z Z Z Z
= g1 + λ g2 − h1 − λ h2 = f1 + λ f2 .
The case of λ < 0 can be reduced to the previous one. Indeed, in this case
we can write f1 + λ f2 = f1 + (−λ )(− f2 ) and observe that − f2 = h2 − g2 and
consequently is also integrable. Linearity is proved.
2. Since f is integrable, f = g − h with g, h ∈ Linc (R). Obviously for every x
| f (x)| = max{ g(x), h(x) } − min{ g(x), h(x) } ,
then Proposition 5.8 implies that the maximum and the minimum belong to
Linc (R) and hence | f | is integrable. The inequality
g(x) + min{ g(x), h(x) } ≤ h(x) + max{ g(x), h(x) }
is valid for every x and in combination with Proposition 5.8 implies

Z Z Z Z
g+ min{ g, h } ≤ h+ max{ g, h } .
Consequently
Z Z Z Z Z Z
f= g− h≤ max{ g, h } − min{ g, h } = | f |.
Applying this result for f replaced by − f we conclude

Z Z Z Z
− f= (− f ) ≤ |− f| = | f |.
R R
Consequently, | f | ≤ | f |.
29
3. Since f (x) = g(x) − h(x) with g, h ∈ Linc (R) and Rf (x) ≥
R
0 a.e., we conclude
g(x) ≥ h(x) a.e.. Then Proposition 5.8 implies that g ≥ h. Consequently
Z Z Z
f= g− h ≥ 0.
We note that | f | is integrable does not imply that f is integrable.9

Exercise: If f is integrable than f+ = max{ f , 0 } and f− = min{ f , 0 } are integrable.
(Hint: f+ = ( f + | f |)/2 and f− = ( f − | f |)/2.)
Integrals and limits

R
You should be careful when swapping lim and :
Examples:
Z Z
lim nχ 1 = 1 6= lim nχ 1
= 0.
n→∞ 0, n n→∞ 0, n
1 1
Z Z
lim χ = 1 6= lim χ(0,n) = 0 .
n→∞ n (0,n) n→∞ n
The following two theorems establish conditions which allow swapping the limit
and integration. They play the fundamental role in the theory of Lebesgue integrals.
Theorem 5.10 (MonotoneR Convergence Theorem) Suppose that fn are integrable,
fn (x) ≤ fn+1 (x) a.e., and fn < K for some constant independent of n. Then there is
an integrable function g such that fn (x) → g(x) a.e. and
Z Z
g = lim fn .
n→∞
R
Corollary 5.11 If f is integrable and | f | = 0, then f (x) = 0 a.e.
Proof: Let fn (x) = n| f (x)|. This sequence satisfies MCT (integrable, increasing and
R
fn = 0 < 1), consequently there is an integrable g(x) such that fn (x) → g(x) for a.e.
x. Since the sequence is increasing, fn (x) ≤ g(x) a.e. which implies | f (x)| ≤ g(x)/n
for all n and a.e. x. Consequently f (x) = 0 a.e.
Theorem 5.12 (Dominated Convergence Theorem) Suppose that fn are integrable
functions and fn (x) → f (x) for a.e. x.. If there is an integrable function g such that
| fn (x)| ≤ g(x) for every n and a.e. x, then f is integrable and
Z Z
f = lim fn .
n→∞
9 Indeed, we can sketch an example. It is based on partitioning the interval [0, 1] into two very nasty
subsets. So let f (x) = 0 outside [0, 1], for x ∈ [0, 1] let f (x) = 1 if x belongs to the Vitali set and
f (x) = −1 otherwise. Then | f | = χ[0,1] but f is not integrable.
30
It is also possible to integrate complex valued functions: f : R → C is integrable if
its real and imaginary parts are both integrable, and
Z Z Z
f := Re f + i Im f .
The MCT has no meaning for complex valued functions. The DCT is valid without
modifications (and indeed follows easily from the real version).
5.3 Lebesgue space L1 (R)

Definition 5.13 The Lebesgue space L1 (R) is the space of integrable functions modulo
the following equivalence relation: f ∼ g iff f (x) = g(x) a.e. The Lebesgue space is
equipped with the L1 norm: Z
k f kL1 = | f |.
It is convenient to think about elements of L1 (R) as functions R → R interpreting

the equality f = g as f (x) = g(x) a.e.
From the viewpoint of Functional Analysis, the equivalence relation is introduced
to
R
ensure non-degeneracy of the L1 norm. Indeed, in the space of integrable functions
| f | = 0 is equivalent to f (x) = 0 a.e. and therefore does not imply the latter equality
for all x.
Theorem 5.14 L1 (R) is a Banach space.
Revise the properties of the Lebesgue integral which imply that L1 (R) is a normed
space. The completeness of L1 (R) follows from the combination of the following two
statements: The first lemma gives a criterion for completeness of a normed space,
and the second one implies that the assumptions of the first lemma are satisfied for
X = L1 (R).
Lemma 5.15 If (X, k · kX ) is a normed space in which

∞
∑ ky j kX < ∞
j=1
implies the series ∑∞j=1 y j converges, then X is complete.
Proof: Let x j ∈ X be a Cauchy sequence. Then there is a monotone increasing se-

quence nk ∈ N such that for every k ∈ N
kx j − xl kX < 2−k for all k, l ≥ nk .
31
Let y1 = xn1 and yk = xnk − xnk−1 for k ≥ 2. Since kyk kX ≤ 21−k for k ≥ 2,
∞ ∞
∑ kyk kX ≤ ky1kX + ∑ 21−k = ky1kX + 2 < ∞ .
k=1 k=1
Consequently there is x∗ ∈ X such that

∞
x∗ = ∑ yj .
j=1
On the other hand

k k
∑ y j = xn1 + ∑ (xn j − xn j−1 ) = xnk
j=1 j=2
and therefore xnk → x∗ . Consequently xk → x∗ and the space X is complete.
∞
k=1 is a sequence of integrable functions such that ∑k=1 k f k kL1 <
Lemma 5.16 If ( fk )∞
∞, then
1. ∑∞
k=1 | f k (x)| converges a.e. to an integrable function,
2. ∑∞
k=1 f k (x) converges a.e. to an integrable function.
Proof: The first statement follows from MCT applied to the sequence gn = ∑nk=1 | fk |
and K = ∑∞k=1 k f k kL1 . So there is an integrable function g(x) such that
∞
g(x) = ∑ | fk (x)|
k=1
for almost all x. For these values of x the partial sums hn (x) = ∑nk=1 fk (x) obviously
converge, so let
∞
h(x) = ∑ fk (x).
k=1
Moreover
n n ∞
|hn (x)| = ∑ fk (x) ≤ ∑ | fk (x)| ≤ ∑ | fk (x)| = g(x) .

k=1 k=1 k=1
Therefore the partial sums hn satisfy DCT and the second statement follows.
In addition to L1 (R) we will sometimes consider the Lebesgue spaces L1 (I) where
I is an interval. We say that f ∈ L1 (I) if χI f ∈ L1 (R), i.e., we extend the function by
zero outside its original domain.
Proposition 5.17 The space C[0, 1] is dense in L1 (0, 1).
Proof: First show that step functions are dense in L1 (0, 1). Then check that every step
function can be approximated by a piecewise linear continuous function.
Consequently the space we constructed in this section is isometrically isomorphic
to the completion of C[0, 1] in the L1 norm.
32
5.4 L p spaces
Another important class of Lebesgue spaces consists of L p spaces for 1 ≤ p < ∞,
among those the L2 space is the most remarkable (it is also a Hilbert space, see the
next chapter for details). In this section we will sketch the main definitions of those
spaces noting that the full discussion requires more knowledge of Measure Theory
than we can fit into this module).
If I = (a, b) is an interval, then L p (I) can be defined in terms of the integration
procedure developed earlier in this chapter. This definition is equivalent to the standard
one which will be given a bit later.
The Lebesgue space L p (I) is the space of all integrable functions such that
Z
k f kLp p = | f |p < ∞
I
modulo the equivalence relation: f = g iff f (x) = g(x) a.e. We note that in this case
L p (I) ⊂ L1 (I). The definition of L p (R) is slightly different. We say that f ∈ L p (R) if f
is locally integrable (i.e., f ∈ L1 (I) for any interval I)10 and its pth power is integrable.
The norm is defined by the same formula:
Z
k f kLp p = | f |p < ∞ .
R
We note that although L1 (R) ∩ L2 (R) 6= 0/ (e.g. both spaces contain all step functions)
none of those spaces is a subset of the other one. For Rexample, f (x) = 1/(1 + |x|)
2 1
R 2
belongs to L (R) but not to L (R). Indeed, f < ∞ but f = ∞ so it is not integrable
on R. On the other hand
χ(0,1) (x)
g(x) =
|x|1/2
belongs to L1 (R) but not to L2 (R).
Theorem 5.18 L p (R) and L p (I) are Banach spaces for p ≥ 1 and any interval I.
We will not give a complete proof but sketch the main ideas instead.
Let 1p + q1 = 1 and f ∈ L p (R), g ∈ Lq (R). Then the Hölder inequality states11 that
Z
| f g| ≤ k f k p kgkq .
10 This is an important requirement. It is not sufficient to define L p as a set of all functions such
that f p is integrable: this space would not be a vector space. Indeed, let p = 2, g = χ[0,1] and f be
the non-integrable function from the footnote9 . Then f 2 = g2 = χ[0,1] is integrable. But ( f + g)2 =
f 2 + 2 f g + g2 = 2 + 2 f is not integrable. Therefore if we followed this definition f , g ∈ L2 would not
imply f + g ∈ L2 .
11 We will not discuss the proof of this inequality in these lectures.
33
Note that the characteristic function χI ∈ Lq (R) for any interval I and any q ≥ 1, more-
1
over kχI kLq = |I| q where |I| = b − a is the length of I. The Hölder inequality with
g = χI implies that Z Z
χI | f | = | f | ≤ |I|1/q k f k p .
I
The left hand side of this inequality is the norm of f in L1 (I):
k f kL1 (I) ≤ |I|1/q k f kL p (I) .
Consequently any Cauchy sequence in L p (I) is automatically a Cauchy sequence in

L1 (I). Since L1 is complete the Cauchy sequence converges to a limit in L1 (I) (and
consequently converges a.e.). In order to proof completeness of L p it is sufficient to
show that the pth power of this limit is integrable. This can be done on the basis of the
Dominated Convergence Theorem.
Exercise: The next two exercises show that L2 (R) is complete (compare with the proof
of completeness for L1 (R)).
1. Let ( fk )∞ 2
k=1 be a sequence in L (R) such that
∞
∑ k f k kL 2 < ∞ .
k=1
Applying the MCT to the sequence

!2
n
gn = ∑ | fk |
k=1
show that ∑k fk converges to a function f with integrable f 2 .

2
2. Now use the DCT applied to hn = f − ∑nk=1 fk to deduce that ∑k fk converges

in the L2 norm to a function in L2 .
If you look into a textbook, you will probably see a differently looking definition
of the Lebesgue spaces. Traditionally a function f is asked to be measurable instead
of locally integrable. Local integrability is a stronger property: every locally inte-
grable function is measurable but there are measurable functions which
R 1 −2
are not locally
−2
integrable, e.g. x is measurable but not locally integrable since −1 x = +∞. Nev-
ertheless the two alternative definitions of the Lebesgue space are equivalent.
Let us discuss the notion of a measurable function from the perspective of our
definitions.
34
First we need to define the measure, which can be considered as a generalisation
of the length of an interval. We say that a subset A ⊂ R has finite Lebesgue measure
µ(A) if the characteristic function χA is Lebesgue integrable. Then
Z
µ(A) := χA ≥ 0 .
Obviously µ([a, b]) = b − a for an interval [a, b] and consequently its Lebesgue mea-
sure coincides with the length.
In order to study large sets (like R) we need to extend this definition to allow
measuring sets with infinitely large measures. We say that A ⊂ R is measurable if χA
is locally integrable. In particular, if A is measurable then An = A ∩ [−n, n] has finite
measure for each n. Since µ(An ) is an increasing sequence the following limit exists
(but can be +∞)
µ(A) = lim µ(An ) ≤ +∞ .
n→∞
R
Note that if χA < ∞ MCT implies that the limit coincides with the previous definition
of µ(A).
For example R is measurable and µ(R) = +∞.
Definition 5.19 A function f : R → R is measurable if preimage of any interval is

measurable.
We note that sums, products and pointwise limits of measurable functions are mea-
surable.
Consider the set of all measurable functions from R to R (or C) whose absolute
value raised to the pth power has a finite Lebesgue integral, i.e.,
Z 1/p
p
k f kL p := |f| < ∞.
This space modulo the equivalence relation “ f = g iff f (x) = g(x) a.e.” is called the
Lebesgue space L p (R).
35
6 Hilbert spaces
6.1 Inner product spaces
You have already seen the inner product on Rn .
Definition 6.1 An inner product on a vector space V is a map (·, ·) : V ×V → K such

that for all x, y, z ∈ V and for all λ ∈ K:
(i) (x, x) ≥ 0, and (x, x) = 0 iff x = 0;
(ii) (x + y, z) = (x, z) + (y, z);
(iii) (λ x, y) = λ (x, y);
(iv) (x, y) = (y, x).

A vector space equipped with an inner product is called an inner product space.
• In a real vector space the complex conjugate in (iv) is not necessary.
• If K = C, then (iv) with y = x implies that (x, x) is real and therefore the require-
ment (x, x) ≥ 0 make sense.
• (iii) and (iv) imply that (x, λ y) = λ (x.y).
1. Example: Rn is an inner product space

n
(x, y) = ∑ xk yk .
k=1
2. Example: Cn is an inner product space

n
(x, y) = ∑ xk yk .
k=1
3. Example: `2 (K) is an inner product space

∞
(x, y) = ∑ xk yk .
k=1
Note that the sum converges because ∑k |xk yk | ≤ 12 ∑k |xk |2 + |yk |2 .

4. Example: L2 (a, b) is an inner product space

Z b
( f , g) = f (x)g(x) dx .
a
36
6.2 Natural norms
Every inner product space is a normed space as well.
Proposition 6.2 If V is an inner product space, then
p
kvk = (v, v)
defines a norm on V .
p
Definition 6.3 We say that kxk = (x, x) is the natural norm induced by the inner
product.
Lemma
p 6.4 (Cauchy-Schwartz inequality) If V is an inner product space and kvk =
(v, v) for all v ∈ V , then
|(x, y)| ≤ kxk kyk for all x, y ∈ V .
Proof of the lemma: The inequality is obvious if y = 0. So suppose that y 6= 0. Then
for any λ ∈ K:
0 ≤ (x − λ y, x − λ y) = (x, x) − λ (y, x) − λ (x, y) + |λ |2 (y, y) .
Then substitute λ = (x, y)/kyk2 :
|(x, y)| |(x, y)| |(x, y)|
0 ≤ (x, x) − 2 2
+ 2
= kxk2 − ,
kyk kyk kyk2
which implies the desired inequality.
Proof of the proposition: Now we can complete the proof of Proposition 6.2. We note
that positive definiteness and homogeneity of k · k easily follow from (i), and (iii), (iv)
in the definition of the inner product. In order to establish the triangle inequality we
use the Cauchy-Schwartz inequality. Let x, y ∈ V . Then
kx + yk2 = (x + y, x + y) = (x, x) + (x, y) + (y, x) + (y, y)
≤ kxk2 + 2kxk kyk + kyk2 = (kxk + kyk)2 ,
and the triangle inequality follows by taking the square root.
Therefore k · k is a norm.
We have already proved the Cauchy-Schwartz inequality for `2 (K) using a different
strategy (see Lemma 2.4).
The Cauchy-Schwartz inequality in L2 (a, b) takes the form
Z b Z b 1/2 Z b 1/2
2 2

f (x)g(x) dx ≤ | f (x)| dx |g(x)| dx .
a a a

In particular, it states that f , g ∈ L2 (a, b) implies f g ∈ L1 (a, b).
37
Lemma 6.5 If V is an inner product space equipped with the natural norm, then xn →
x and yn → y imply that
(xn , yn ) → (x, y) .
Proof: Since any convergent sequence is bounded, the inequality
|(xn , yn ) − (x, y)| = |(xn − x, yn ) + (x, yn − y)|

≤ |(xn − x, yn )| + |(x, yn − y)|
≤ kxn − xk kyn k + kxk kyn − yk
implies that (xn , yn ) → (x, y).
The lemma implies that we can swap inner products and limits.
6.3 Parallelogram law and polarisation identity

Natural norms have some special properties.
Lemma 6.6 (Parallelogram law) If V is an inner product space with the natural norm
k · k, then
kx + yk2 + kx − yk2 = 2 kxk2 + kyk2

for all x, y ∈ V .
Proof: The linearity of the inner product implies that for any x, y ∈ V
kx + yk2 + kx − yk2 = (x + y, x + y) + (x − y, x − y)
= (x, x) + (x, y) + (y, x) + (y, y)
+(x, x) − (x, y) − (y, x) + (y, y)
= 2 kxk2 + kyk2

Example (some norms are not induced by an inner product): There is no inner
product which induces the following norms on C[0, 1]:
Z 1
k f k∞ = sup | f (t)| or k f kL1 = | f (t)| dt .
t∈[0,1] 0
Indeed, these norms do not satisfy the parallelogram law, e.g., take f (x) = x and g(x) =
1 − x, obviously f , g ∈ C[0, 1] and
k f k∞ = kgk∞ = k f − gk∞ = k f + gk∞ = 1 ,
substituting these numbers into the parallelogram law we see 2 6= 4.

Exercise: Is the parallelogram law for the L1 norm satisfied for these f , g?
38
Lemma 6.7 (Polarisation identity) Let V be an inner product space with the natural
norm k · k. Then
1. If V is real
4(x, y) = kx + yk2 − kx − yk2 ;
2. If V is complex
4(x, y) = kx + yk2 − kx − yk2 + ikx + iyk2 − ikx − iyk2 .
Proof: Plug in the definition of the natural norm into the right hand side and use
linearity of the inner product.
Lemma 6.7 shows that the inner product can be restored from its natural norm.
Although the right hand sides of the polarisation identities is meaningful for any norm,
we should not rush to the conclusion that any normed space is automatically an inner
product space. Indeed, the example above implies that for some norms these formulae
cannot define an inner product. Nevertheless, if the norm satisfy the parallelogram law,
we indeed get an inner product:
Proposition 6.8 Let V be a real normed space with the norm k · k satisfying the par-
allelogram law, then
kx + yk2 − kx − yk2 kx + yk2 − kxk2 − kyk2
(x, y) = =
4 2
defines an inner product on V .12
Proof: Let us check that (x, y) satisfy the axioms of inner product. Positivity and
symmetry are straightforward (Exercise). The linearity:
4(x, y) + 4(z, y) = kx + yk2 − kx − yk2 + kz + yk2 − kz − yk2

= 12 (kx + 2y + zk2 + kx − zk2 ) − 12 (kx − 2y + zk2 + kx − zk2 )
1 2 1 2
= 2 kx + 2y + zk − 2 kx − 2y + zk
1 2 2 2
= 2 (2kx + y + zk + 2kyk − kx + zk )
− 21 (2kx − y + zk2 + 2kyk2 − kx + zk2 )
2 2
= kx + y + zk − kx − y + zk = 4(x + z, y) .
We have proved that

(x, y) + (z, y) = (x + z, y) .
Applying this identity several times and setting z = x/m we obtain
n(x/m, y) = (nx/m, y) and m(x/m, y) = (x, y)

12 Can you find a simpler proof?
39
n
for any m ∈ Z and n ∈ N. Consequently, for any rational λ = m
(λ x, y) = λ (x, y) .
We note that the right hand side of the definition involves the norms only, which com-
mute with the limits. Any real number is a limit of rational numbers and therefore the
linearity holds for all λ ∈ R.
6.4 Hilbert spaces: Definition and examples

Definition 6.9 A Hilbert space is a complete inner product space (equipped with the
natural norm).
Of course, any Hilbert space is a Banach space.
1. Example: Rn is a Hilbert space

!1/2
n n
(x, y) = ∑ xk yk , kxk = ∑ xk2 .
k=1 k=1
2. Example: Cn is a Hilbert space

!1/2
n n
(x, y) = ∑ xk yk , kxk = ∑ |xk |2 .
k=1 k=1
3. Example: `2 (K) is a Hilbert space

!1/2
∞ ∞
(x, y) = ∑ xk yk , kxk = ∑ |xk |2 .
k=1 k=1
4. Example: L2 (a, b) is a Hilbert space

Z b Z b
1/2
2
( f , g) = f (x)g(x) dx, kxk = | f (x)| dx .
a a
40
7 Orthonormal bases in Hilbert spaces
The goal of this section is to discuss properties of orthonormal bases in a Hilbert space
H. Unlike Hamel bases, the orthonormal ones involve a countable number of elements:
i.e. a vector x is represented in the form of an infinite sum
∞
x= ∑ αk ek
k=1
for some αk ∈ K.
We will mainly consider complex spaces with K = C. The real case K = R is not
very different. We will use (·, ·) to denote an inner product on H, and k · k will stand
for the natural norm induced by the inner product.
7.1 Orthonormal sets

Definition 7.1 Two vectors x, y ∈ H are called orthogonal if (x, y) = 0. Then we write
x ⊥ y.
Theorem 7.2 (Pythagoras theorem) If x ⊥ y then kx + yk2 = kxk2 + kyk2 .
Proof: Since (x, y) = 0
kx + yk2 = (x + y, x + y) = (x, x) + (x, y) + (y, x) + (y, y) = kxk2 + kyk2 .
Definition 7.3 A set E is orthonormal if kek = 1 for all e ∈ E and (e1 , e2 ) = 0 for all
e1 , e2 ∈ E such that e1 6= e2 .
Note that this definition does not require the set E to be countable.
Exercise: Any orthonormal set is linearly independent.
Indeed, suppose ∑nk=1 αk ek = 0 with ek ∈ E and αk ∈ K. Multiplying this equality
by e j we get !
n n
0= ∑ αk ek , e j = ∑ αk (ek , e j ) = α j .
k=1 k=1
Since α j = 0 for all j, we conclude that the set E is linearly independent.
Definition 7.4 (Kronecker delta) The Kronecker delta is the function defined by

1, if j = k,
δ jk =
0, if j 6= k .
41
Example: For every j ∈ N, let e j = (δ jk )∞k=1 (it is an infinite sequence of zeros with 1
at the jth position). The set E = { e j : j ∈ N } is orthonormal in `2 . Indeed, from the
definition of the scalar product in `2 we see that (e j , ek ) = δ jk for all j, k ∈ N.
Example: The set13

eikx

E= fk = √ : k ∈ Z
2π
is an orthonormal set in L2 (−π, π). Indeed, since | fk (x)| = √1 for all x:
2π
Z π
k fk k2L2 = | fk (x)|2 dx = 1 ,
−π
and if j 6= k
x=π
1 π i(k− j)x ei(k− j)x
Z π Z
( fk , f j ) = fk (x) f j (x) dx = e dx = = 0.
2π i(k − j)

−π −π
x=−π
Lemma 7.5 If {e1 , . . . , en } is an orthonormal set in an inner product space V , then for
any α j ∈ K
2
n n
∑ α j e j = ∑ |α j |2 .

j=1 j=1
Proof: The following computation is straightforward:

2 !
n n n n n
∑ α je j = ∑ α j e j , ∑ αl el = ∑ ∑ α j α l (e j , el )

j=1 j=1 l=1 j=1 l=1
n n n
= ∑ ∑ α j α l δ jl = ∑ α jα j .
j=1 l=1 j=1
7.2 Gram-Schmidt orthonormalisation

Lemma 7.6 (Gram-Schmidt orthonormalisation) Let V be an inner product space
and (vk ) be a sequence of linearly independent vectors in V (finite or infinite). Then
there is an orthonormal sequence (ek ) such that
Span{ v1 , . . . , vk } = Span{ e1 , . . . , ek } for all k.

13 Remember that for any x ∈ R and any k ∈ Z: eikx = cos kx + i sin kx. Then |eikx | = 1 and e±ikπ =
cos kπ ± i sin kπ = (−1)k .
42
v1
Proof: Let e1 = kv1 k . Then
Span{ v1 } = Span{ e1 }
and the statement is true for n = 1 as the set E1 = { e1 } is obviously orthonormal.
Then we continue inductively.14 Suppose that for some k ≥ 2 we have found an
orthonormal set Ek−1 = { e1 , . . . , ek−1 } such that its span coincides with the span of
{ v1 , . . . , vk−1 }. Then set
k−1
ẽk = vk − ∑ (vk , e j )e j .
j=1
Since ∑k−1 ∈ Span(Ek−1 ) = Span{ v1 , . . . , vk−1 } and v1 , . . . , vk are linearly

j=1 (vk , e j )e j
independent, we conclude that ẽk 6= 0. For every j < k
k−1
(ẽk , e j ) = (vk , el ) − ∑ (vk , e j )(e j , el ) = (vk , el ) − (vk , el ) = 0
j=1
which implies that ẽk ⊥ e j . Finally let ek = ẽk /kẽk k. Then { e1 , . . . , ek } is an orthonor-
mal set such that
Span{ e1 , . . . , ek } = Span{ v1 , . . . , vk } .
If the original sequence is finite, the orthonormalisation procedure will stop after a
finite number of steps. Otherwise, we get an infinite sequence of ek .
Corollary 7.7 Any infinite-dimensional inner product space contains a countable or-
thonormal sequence.
Corollary 7.8 Any finite-dimensional inner product space has an orthonormal basis.
Proposition 7.9 Any finite dimensional inner product space is isometric to Cn (or Rn
if the space is real) equipped with the standard inner product.
Proof: Let n = dimV and e j , j = 1, . . . , n be an orthonormal basis in V . Note that
(ek , e j ) = δk j . Any two vectors x, y ∈ V can be written as
n n
x= ∑ xk ek and y= ∑ y je j .
k=1 j=1
Then !
n n n n n
(x, y) = ∑ xk ek , ∑ y j e j = ∑ ∑ xk y j (ek , e j ) = ∑ xk yk .
k=1 j=1 k=1 j=1 k=1
Therefore the map x 7→ (x1 , . . . , xn ) is an isometry.
We see that an arbitrary inner product, when written in orthonormal coordinates,
takes the form of the “canonical” inner product on Cn (or Rn if the original space is
real).
14 For example, let k = 2. We define ẽ2 = v2 −(v2 , e1 )e1 . Then (ẽ2 , e1 ) = (v2 , e1 )−(v2 , e1 )(e1 , e1 ) = 0.
Since v1 , v2 are linearly independent ẽ2 6= 0. So we can define e2 = kẽẽ2 k .
2
43
7.3 Bessel’s inequality
Lemma 7.10 (Bessel’s inequality) If V is an inner product space and E = (ek )∞
k=1 is
an orthonormal sequence, then for every x ∈ V
∞
∑ |(x, ek )|2 ≤ kxk2 .
k=1
Proof: We note that for any n ∈ N:

2 !
n n n
x − ∑ (x, ek )ek = x − ∑ (x, ek )ek , x − ∑ (x, ek )ek

k=1 k=1 k=1

n n
= kxk2 − 2 ∑ |(x, ek )|2 + ∑ |(x, ek )|2
k=1 k=1
n
= kxk2 − ∑ |(x, ek )|2 .
k=1
Since the left hand side is not negative,

n
∑ |(x, ek )|2 ≤ kxk2
k=1
and the lemma follows by taking the limit as n → ∞.
Corollary 7.11 If E is an orthonormal set in an inner product space V , then for any
x ∈ V the set
Ex = { e ∈ E : (x, e) 6= 0 }
is at most countable.
Proof: For any m ∈ N the set Em = { e : |(x, e)| > m1 } has a finite number of elements.
Otherwise there would be an infinite sequence (ek )∞ k=1 with ek ∈ Em , then the series
∑k=1 |(x, ek )| = +∞ which contradicts to Bessel’s inequality. Therefore Ex = ∪∞
∞ 2
m=1 Em
is a countable union of finite sets and hence at most countable.
7.4 Convergence
In this section we will discuss convergence of series which involve elements from an
orthonormal set.
Lemma 7.12 Let H be a Hilbert space and E = (ek )∞ k=1 an orthonormal sequence.
The series ∑∞ α e converges iff ∞
|α |2 < +∞. Then
k=1 k k ∑k=1 k
2
∞ ∞
∑ αk ek = ∑ |αk |2 . (7.1)

k=1 k=1
44
Proof: Let xn = ∑nk=1 αk ek and βn = ∑nk=1 |αk |2 . Lemma 7.5 implies that kxn k2 = βn
and that for any n > m
2
n n
kxn − xm k2 = ∑ αk ek = ∑ |αk |2 = βn − βm .

k=m+1 k=m+1
Consequently, xn is a Cauchy sequence in H iff βn is Cauchy in R. Since both spaces

are complete, the sequences converge or diverge simultaneously.
If they converge, we take the limit as n → ∞ in the equality kxn k2 = βn to get (7.1)
(the limit commutes with k · k2 ).
Definition 7.13 A series ∑∞ n=1 xn in a Banach space X is unconditionally convergent

if for every permutation σ : N → N the series ∑∞n=1 xσ (n) converges.
Every absolutely convergent series is unconditionally convergent, but the converse

implication does not hold in general. In Rn a series is unconditionally convergent if
and only if it is absolutely convergent.
Lemma 7.12 and Bessel’s inequality imply:
Corollary 7.14 If H is a Hilbert space and E = (ek )∞

k=1 is an orthonormal sequence,
then for every x ∈ H the sequence
∞
∑ (x, ek )ek
k=1
converges unconditionally.
Lemma 7.15 Let H be a Hilbert space, E = (ek )∞

k=1 an orthonormal sequence and
∞
x ∈ H. If x = ∑k=1 αk ek , then
αk = (x, ek ) for all k ∈ N.
Proof: Exercise.
7.5 Orthonormal basis in a Hilbert space

Definition 7.16 A set E is a basis for H if every x ∈ H can be written uniquely in the
form
∞
x= ∑ αk e k
k=1
for some αk ∈ K and ek ∈ E. If additionally E is an orthonormal set, then E is an
orthonormal basis.
45
If E is a basis, then it is a linearly independent set. Indeed, if ∑nk αk ek = 0 then
αk = 0 due to the uniqueness.
Note that in this definition the uniqueness is a delicate point. Indeed, the sum
∑∞
k=1 αk ek is defined as a limit of partial sums xn = ∑nk=1 αk ek . A permutation of ek
changes the partial sums and may lead to a different limit. In general, we cannot even
guarantee that after a permutation the series remains convergent.
If E is countable, we can assume that the sum involves all elements of the basis
(some αk can be zero) and that the summation is taken following the order of a selected
enumeration of E. The situation is more difficult if E is uncountable since in this case
there is no natural way of numbering the elements.
The situation is much simpler if E is orthonormal as in this case the series converge
unconditionally and the order of summations is not important.
Proposition 7.17 Let E = { e j : j ∈ N } be an orthonormal set in a Hilbert space H.

Then the following statements are equivalent:
(a) E is a basis in H;
(b) x = ∑∞
k=1 (x, ek )ek for all x ∈ H;
(c) kxk2 = ∑∞ 2
k=1 |(x, ek )| ;
(d) (x, en ) = 0 for all n ∈ N implies x = 0;
(e) the linear span of E is dense in H.
Proof:
(a) ⇐⇒ (b): use Lemma 7.15.
(b) =⇒ (c): use Lemma 7.12.
(c) =⇒ (d): Let (x, ek ) = 0 for all k, then (c) implies that kxk = 0 hence x = 0.
(d) =⇒ (b): let y = x − ∑∞
k=1 (x, ek )ek . Corollary 7.14 implies that the series con-
verges. Then Lemma 6.5 implies we can swap the limit and the inner product to get
for every n
!
∞
(y, en ) = x − ∑ (x, ek )ek , en
k=1
∞
= (x, en ) − ∑ (x, ek )(ek , en ) = (x, en ) − (x, en ) = 0 .
k=1
Since (y, en ) = 0 for all n, then (d) implies that y = 0 which is equivalent to x =
∑∞k=1 (x, ek )ek as required.
46
(e) =⇒ (d): since Span(E) is dense in H for any x ∈ H there is a sequence xn ∈
Span(E) such that xn → x. Take x such that (x, en ) = 0 for all n. Then (xn , x) = 0 and
consequently
kxk2 = lim xn , x = lim (xn , x) = 0 .
n→∞ n→∞
Therefore x = 0.
(a) =⇒ (e): Since E is a basis any x = limn→∞ xn with xn = ∑nk=1 αk ek ∈ Span(E).

Example: The orthonormal sets from examples of Section 7.1 are also examples of
orthonormal bases.
47
8 Closest points and approximations
8.1 Closest points in convex subsets
Definition 8.1 A subset A of a vector space V is convex if λ x + (1 − λ )y ∈ A for any
two vectors x, y ∈ V and any λ ∈ [0, 1].
Lemma 8.2 Let A be a non-empty closed convex subspace of a Hilbert space H and
x ∈ H. Then there is a unique a∗ ∈ A such that
kx − a∗ k = inf kx − ak .
a∈A
Proof: The parallelogram rule implies:
k(x − u) + (x − v)k2 + k(x − u) − (x − v)k2 = 2kx − uk2 + 2kx − vk2 .
Then
ku − vk2 = 2kx − uk2 + 2kx − vk2 − 4kx − 12 (u + v)k2 .
Let d = infa∈A kx − ak. Since A is convex, 21 (u + v) ∈ A for any u, v ∈ A, and conse-
quently kx − 12 (u + v)k ≥ d. Then
ku − vk2 ≤ 2kx − uk2 + 2kx − vk2 − 4d 2 . (8.1)
Since d is the infinum, for any n there is an ∈ A such that kx − an k2 < d 2 + n1 . Then
equation (8.1) implies that
2 2 2 2
kan − am k ≤ 2d 2 + + 2d 2 + − 4d 2 = + .
n m n m
Consequently (an ) is Cauchy and, since H is complete, it converges to some a∗ . Since
A is closed, a∗ ∈ A. Then
kx − a∗ k2 = lim kx − an k2 = d 2 .
n→∞
Therefore a∗ is the point closest to x. Now suppose that there is another point ã ∈ A
such that kx − ãk = d, then (8.1) implies
ka∗ − ãk ≤ 2kx − a∗ k2 + 2kx − ãk2 − 4d 2 = 2d 2 + 2d 2 − 4d 2 = 0 .
So ã = a∗ and a∗ is unique.
48
8.2 Orthogonal complements
In an infinite dimensional space a linear subspace does not need to be closed. For
example the space ` f of all sequences with only a finite number of non-zero elements
is a linear subspace of `2 but it not closed in `2 (e.g. consider the sequence xn =
(1, 2−1 , 2−2 , . . . , 2−n , 0, 0, . . .)).
Definition 8.3 Let X ⊆ H. The orthogonal complement of X in H is the set
X ⊥ = {u ∈ H : (u, x) = 0 for all x ∈ X } .
Proposition 8.4 If X ⊆ H, then X ⊥ is a closed linear subspace of H.
Proof: If u, v ∈ X ⊥ and α ∈ K then
(u + αv, x) = (u, x) + α(v, x) = 0
for all x ∈ X. Therefore X ⊥ is a linear subspace. Now suppose that un ∈ X ⊥ and

un → u ∈ H. Then for all x ∈ X
(u, x) = ( lim un , x) = lim (un , x) = 0 .

n→∞ n→∞
Consequently, u ∈ X ⊥ and so X ⊥ is closed.
Exercises:
1. If E is a basis in H, then E ⊥ = { 0 }.
2. If Y ⊆ X, then X ⊥ ⊆ Y ⊥ .
3. X ⊆ (X ⊥ )⊥
4. If X is a closed linear subspace in H, then X = (X ⊥ )⊥
Definition 8.5 The closed linear span of E ⊂ H is a minimal closed set which contains
Span(E):
Span(E) = { u ∈ H : ∀ε > 0 ∃x ∈ Span(E) such that kx − uk < ε } .
Proposition 8.6 If E ⊂ H then E ⊥ = (Span(E))⊥ = (Span(E))⊥ .
Proof: Since E ⊆ Span(E) ⊆ Span(E) we have (Span(E))⊥ ⊆ (Span(E))⊥ ⊆ E ⊥ . So

we need to prove the inverse inclusion. Take u ∈ E ⊥ and x ∈ Span(E). Then there is
xn ∈ Span(E) such that xn → x. Then
(x, u) = ( lim xn , u) = lim (xn , u) = 0 .

n→∞ n→∞
Consequently, u ∈ (Span(E))⊥ and we proved E ⊥ ⊆ (Span(E))⊥ .
49
Theorem 8.7 If U is a closed linear subspace of a Hilbert space H then
1. any x ∈ H can be written uniquely in the form x = u + v with u ∈ U and v ∈ U ⊥ .
2. u is the closest point to x in U.
3. The map PU : H → U defined by PU x = u is linear and satisfies
PU2 x = PU x and kPU (x)k ≤ kxk for all x ∈ H .
Definition 8.8 The map PU is called the orthogonal projector onto U.

Proof: Any linear subspace is obviously convex. Then Lemma 8.2 implies that there
is a unique u ∈ U such that
kx − uk = inf kx − ak .
a∈U
Let v = x − u. Let us show that v ∈ U ⊥ . Indeed, take any y ∈ U and consider the
function ∆ : C → R defined by
∆(t) = kv + tyk2 = kx − (u − ty)k2 .
Since the definition of u together with u − ty ∈ U imply that ∆(t) ≥ ∆(0) = kx − uk2 ,
the function ∆ has a minimum at t = 0. On the other hand
∆(t) = kv + tyk2 = (v + ty, v + ty)
= (v, v) + t(y, v) + t¯(v, y) + |t|2 (y, y) .
d∆
First suppose that t is real. Then t¯ = t and dt (0) = 0 implies
(y, v) + (v, y) = 0 .
d∆
Then suppose that t is purely imaginary, Then t¯ = −t and dt (0) = 0 implies
(y, v) − (v, y) = 0 .
Taking the sum of these two equalities we conclude
(y, v) = 0 for every y ∈ U.
Therefore v ∈ U ⊥ .
In order to prove the uniqueness of the representation suppose x = u1 + v1 = u + v
with u1 , u ∈ U and v1 , v ∈ U ⊥ . Then u1 − u = v − v1 . Since u − u1 ∈ U and v − v1 ∈ U ⊥ ,
kv − v1 k2 = (v − v1 , v − v1 ) = (v − v1 , u1 − u) = 0 .
Therefore u and v are unique.
Finally x = u + v with u ⊥ v implies kxk2 = kuk2 + kvk2 . Consequently kPU (x)k =
kuk ≤ kxk. We also note that PU (u) = u for any u ∈ U. So PU2 (x) = PU (x) as PU (x) ∈ U

Corollary 8.9 If U is a closed linear subspace in a Hilbert space H and x ∈ H, then
PU (x) is the closest point to x in U.
50
8.3 Best approximations
Theorem 8.10 Let E be an orthonormal sequence: E = { e j : j ∈ J } where J is
either finite or countable set. Then for any x ∈ H, the closest point to x in Span(E) is
given by
y = ∑ (x, e j )e j .
j∈J
Corollary 8.11 If E is an orthonormal basis in a closed subspace U ⊂ H, then the

orthogonal projection onto U is given by
PU (x) = ∑ (x, e j )e j .
j∈J
Proof: Corollary 7.14 implies that u = ∑ j∈J (x, e j )e j converges. Then obviously u ∈
Span(E) which is a closed linear subset. Let v = x−u. Since (v, ek ) = (x, ek )−(u, ek ) =
0 for all k ∈ J, we conclude v ∈ E ⊥ = (Span(E))⊥ (Lemma 8.6). Theorem 8.7 implies
that u is the closest point.
Example: The best approximation of an element x ∈ `2 in terms of the elements of the

standard basis (e j )nj=1 is given by
n
∑ (x, e j )e j = (x1, x2, . . . , xn, 0, 0, . . .) .
k=1
Example: Let (e j )∞j=1 be an orthonormal basis in H. The best approximation of an

element x ∈ H in terms of the first n elements of the orthonormal basis is given by
n
∑ (x, e j )e j .
k=1
Now suppose that the set E is not orthonormal. If the set E is finite or countable we
can use the Gram-Schmidt orthonormalisation procedure to construct an orthonormal
basis in Span(E). After that the theorem above gives us an explicit expression for the
best approximation. Let’s consider some examples.
Example: Find the best approximation of a function f ∈ L2 (−1, 1) with polynomials

of degree up to n. In other words, let E = { 1, x, x2 , . . . , xn }. We need to find u ∈
Span(E) such that
k f − ukL2 = inf k f − pkL2 .
p∈Span(E)
The set E is not orthonormal. Let’s apply the Gram-Schmidt orthonormalisation pro-
cedure to construct an orthonormal basis in Span(E). For the sake of shortness, let’s
write k · k = k · kL2 (−1,1) .
51
√
First note that k1k = 2 and let
1
e1 = √ .
2
R1 R1 2
Then (1, x) = −1 x dx = 0 and kxk2 = 2
−1 |x| dx = 3 so let
r
3
e2 = x.
2
Then
ẽ3 = x2 − (x2 , e2 )e2 − (x2 , e1 )e1

r Z r
3 1 3 1
Z 1
1
2 2
= x − x t t dt − √ t 2 √ dt
2 −1 2 2 −1 2
Z 1
1 1
= x2 − t 2 dt = x2 − .
2 −1 3
Taking into account that
Z 1 2
2 2 1 8
kẽ3 k = x − dt =
−1 3 45
we obtain r
ẽ3 5
e3 = = (3x2 − 1) .
kẽ3 k 8
q
7 3
Exercise: Show that e4 = 8 (5x − 3x) is orthogonal to e1 , e2 and e3 .
The best approximation of any function f ∈ L2 (−1, 1) by a polynomial of third

degree is given by
7 3 1 Z
5
Z 1
(5x − 3x) f (t)(5t 3 − 3t)dt + (3x2 − 1) f (t)(3t 2 − 1)dt
8 −1 8 −1
Z 1
3 1 1
Z
+ x t f (t) dt + f (t) dt
2 −1 2 −1
For example, if f (x) = |x| its best approximation by a third degree polynomial is
15x2 + 3
p3 = .
16
We can check (after computing the corresponding integral);
3
k f − p3 k2 = .
16
52
Note that the best approximation in the L2 norm is not necessarily the best approx-
imation in the sup norm. Indeed, for example,
2 + 3

15x > 3
sup |x| −
x∈[−1,1] 16 16
(the supremum is larger than the values at x = 0). At the same time

2 1 1
sup |x| − x +
= .
x∈[−1,1] 8 8
8.4 Weierstrass Approximation Theorem

In this section we will prove an approximation theorem which is independent from the
discussions of the previous lectures. This theorem implies that polynomials are dense
in the space of continuous functions on an interval.
Theorem 8.12 If f : [0, 1] → R is continuous on [0, 1] then the sequence of polynomials

n
n
Pn (x) = ∑ f (p/n)x p (1 − x)n−p
p=0 p
uniformly converges to f on [0, 1].
Proof: The binomial theorem states that

n
n n p n−p
(x + y) = ∑ x y .
p=0 p
Differentiating with respect to x and multiplying by x we get

n
n−1 n p n−p
nx(x + y) = ∑p x y .
p=0 p
Differentiating the original identity twice with respect to x and multiplying by x2 we

get
n
2 n−2 n p n−p
n(n − 1)x (x + y) = ∑ p(p − 1) x y .
p=0 p
Now substitute y = 1 − x and denote

n p
r p (x) = x (1 − x)n−p .
p
53
We get
n
∑ r p(x) = 1,
p=0
n
∑ pr p(x) = nx .
p=0
n
∑ p(p − 1)r p(x) = n(n − 1)x2 .
p=0
Consequently,
n n n n
∑ (p − nx)2r p(x) = ∑ p2 r p (x) − 2nx ∑ pr p (x) + n2 x2 ∑ r p(x)
p=0 p=0 p=0 p=0
= n(n − 1)x2 + nx − 2(nx)2 + n2 x2 = nx(1 − x) .
Let M = supx∈[0,1] | f (x)|. Note that f is uniformly continuous, i.e., for every ε > 0
there is δ > 0 such that
|x − y| < δ =⇒ | f (x) − f (y)| < ε .
Now we can estimate

n
| f (x) − Pn (x)| = f (x) − ∑ f (p/n)r p (x)

p=0

n
= ∑ ( f (x) − f (p/n))r p (x)

p=0

≤ ∑ ( f (x) − f (p/n))r p (x)

|x−p/n|<δ

+ ∑ ( f (x) − f (p/n))r p (x)

|x−p/n|>δ
The first sum is bounded by

∑ ( f (x) − f (p/n))r p (x) ≤ ε ∑ r p (x) < ε

|x−p/n|<δ |x−p/n|<δ
54
The second sum is bounded by

∑ ( f (x) − f (p/n))r p (x) ≤ 2M ∑ r p (x)

|x−p/n|>δ |nx−p|>nδ
n
(p − nx)2
≤ ∑ 2 2
r p (x)
p=0 n δ
2Mx(1 − x) 2M
= 2
≤ 2
nδ nδ
2M
which is less than ε for any n > nδ 2 ε
. Therefore for these values of n
| f (x) − Pn (x)| < 2ε .
Consequently,
sup | f (x) − Pn (x)| → 0 as n → ∞.

x∈[0,1]
Corollary 8.13 The set of polynomials is dense in C[0, 1] equipped with the supremum
norm.
55
9 Separable Hilbert spaces
Definition 9.1 A normed space is separable if it contains a countable dense subset.
In other words, a space H is separable if there is a countable set { xn ∈ H : n ∈ N }

such that for any u ∈ H and any ε > 0 there is n ∈ N such that
kxn − uk < ε .
Examples: R is separable (Q is dense). Rn is separable (Qn is dense), Cn is separable

(Qn + iQn is dense).
Example: `2 is separable. Indeed, the set of sequences (x1 , x2 , . . . , xn , 0, 0, 0, . . .) with

x j ∈ Q is dense and countable.
Example: The space C[0, 1] is separable. Indeed, the Weierstrass approximation the-
orem states that every continuous function can be approximated (in the sup norm) by
a polynomial. The dense countable set is given by polynomials with rational coeffi-
cients.
Example: L2 (0, 1) is separable. Indeed, continuous functions are dense in L2 (0, 1)

(in the L2 -norm). The polynomials are dense in C[0, 1] (in the supremum norm and
therefore in the L2 norm as well). The set of polynomials with rational coefficients is
dense in the set of all polynomials and, consequently, it is also dense in L2 [0, 1] (in the
L2 norm).
9.2 Isometry to `2
If H is a Hilbert space, then its separability is equivalent to existence of a countable
orthonormal basis.
Proposition 9.2 An infinite-dimensional Hilbert space is separable iff it has a count-

able orthonormal basis.
Proof: If a Hilbert space has a countable basis, then we can construct a countable dense
set by taking finite linear combinations of the basis elements with rational coefficients.
Therefore the space is separable.
If H is separable, then it contains a countable dense subset V = {xn : n ∈ N}.
Obviously, the closed linear span of V coincides with H. First we construct a linear
independent set Ṽ which has the same linear span as V by eliminating from V those
xn which are not linearly independent from { x1 , . . . , xn−1 }. Then the Gram-Schmidt
process gives an orthonormal sequence with the same closed linear span, i.e., it is a
basis by characterisation (e) of Proposition 7.17.
56
The following theorem shows that all infinite dimensional separable spaces are
isometric. So in some sense `2 is essentially the “only” separable infinite-dimensional
space.
Theorem 9.3 Any infinite-dimensional Hilbert space is isometric to `2 .
Proof: Let { e j } be an orthonormal basis in H. The map A : H → `2 defined by
A : u → ((u, e1 ), (u, e2 ), (u, e3 ), . . . }
is invertible. Indeed, the image of A is in `2 due to Lemma 7.12, and the inverse map
is given by
∞
A−1 : (xk )∞
k=1 7→ ∑ xk ek .
k=1
The characterisation of a basis in Proposition 7.17 implies that kukH = kA(u)k`2 .

Note that there are Hilbert spaces which are not separable.
Example: Let J be uncountable. The space of all functions f : J → R such that
∑ | f ( j)|2 < ∞
j∈J
is a Hilbert space. It is not separable.15
15 How do we define the sum over an uncountable set? For any n ∈ N the set J = { j ∈ J : | f ( j)| >
n
1
n } is finite (otherwise the sum is obviously infinite). Consequently, the set J ( f ) := { j ∈ J : | f ( j)| >
0 } is countable because it is a countable union of finite sets: J ( f ) = ∪∞
n=1 Jn . Therefore, the number
of non-zero terms in the sum is countable and the usual definition of an infinite sum can be used.
57
10 Linear maps between Banach spaces
A linear map on a vector space is traditionally called a linear operator. All linear
functions defined on a finite-dimensional space are continuous. This statement is no
longer true in the case of an infinite dimensional space.
We will begin our study with continuous operators: this class has a rich theory and
numerous applications. We will only slightly touch some of them (the most remarkable
examples will be the shift operators on `2 , and integral operators and multiplication
operators on L2 ).
Of course many interesting linear maps are not continuous, i.e., the differential
operator A : f 7→ f 0 on the space of continuously differentiable functions. More accu-
rately, let D(A) = C1 [0, 1] ⊂ L2 (0, 1) be the domain of A Obviously A : D(A) → L2 (0, 1)
is linear but not continuous. Indeed, consider the sequence xn (t) = n−1 sin(nt). Obvi-
ously kxn kL2 ≤ n−1 so xn → 0, but A(xn ) = cos(nt) does not converge to A(0) = 0 in
the L2 norm so A is not continuous.
Some definitions and properties from the theory of continuous linear operators can
be literally extended onto unbounded ones, but sometimes subtle differences appear:
e.g., we will see that a bounded operator is self-adjoint iff it is symmetric, which is
no longer true for unbounded operators. In a study of unbounded operators a special
attention should be paid to their domains.
10.1 Continuous linear maps

Let U and V be vector spaces over K.
Definition 10.1 A function A : U → V is called a linear operator if
A(αx + β y) = αA(x) + β A(y) for all x, y ∈ U and α, β ∈ K.
We will often write Ax to denote A(x).

The collection of all linear operators from U to V is a vector space. If A, B : U → V
are linear operators and α, β ∈ K then we define
(αA + β B)(x) = αAx + β Bx .
Obviously, αA + β B is also linear.
Definition 10.2 A linear operator A : U → V is bounded if there is a constant M such

that
kAxkV ≤ MkxkU for all x ∈ U. (10.1)
If an operator is bounded, then the image of a bounded set is also bounded.
Lemma 10.3 A linear operator A : U → V is continuous iff it is bounded.
58
Proof: Suppose A is bounded. Then there is M > 0 such that
kA(x) − A(y)k = kA(x − y)k ≤ Mkx − yk
for all x, y ∈ V and consequently A is continuous.

Now suppose A is continuous. Obviously A(0) = 0. Then for ε = 1 there is δ > 0
such that kA(x)k < ε = 1 for all kxk < δ . For any u ∈ U, u 6= 0,

2kuk δ
A(u) = A u .
δ 2kuk

Since 2kuk u = 2 < δ we get kA(u)k ≤ 2kuk
δ δ
δ and consequently A is bounded.
The space of all bounded linear operators from U to V is denoted by B(U,V ).
Definition 10.4 The operator norm of A : U → V is

kA(x)kV
kAkB(U,V ) = sup .
x6=0 kxkU
We will often write kAkop instead of kAkB(U,V ) .
Since A is linear
kAkB(U,V ) = sup kA(x)kV .
kxkU =1
We note that kAkB(U,V ) is the smallest M such that (10.1) holds: indeed, it is easy to
see that the definition of operator norm implies
kA(x)kV ≤ kAkB(U,V ) kxkU
and (10.1) holds with M = kAkB(U,V ) . On the other hand, (10.1) implies M ≥ kAxk V
kxkU for
any x 6= 0 and consequently M ≥ kAkB(U,V ) .
Theorem 10.5 Let U be a normed space and V be a Banach space. Then B(U,V ) is a
Banach space.
n=1 be a Cauchy sequence in B(U,V ). Take a vector u ∈ U. The

Proof: Let (An )∞
sequence vn = An (u) is a Cauchy sequence in V :
kvn − vm k = kAn (u) − Am (u)k = k(An − Am )(u)k ≤ kAn − Am kop kuk.
Since V is complete there is v ∈ V such that vn → v. Let A(u) = v.

The operator A is linear. Indeed,
A(α1 u1 + α2 u2 ) = lim An (α1 u1 + α2 u2 ) = lim (α1 An (u1 ) + α2 An (u2 ))

n→∞ n→∞
= α1 lim An u1 + α2 lim An u2 = α1 Au1 + α2 Au2 .
n→∞ n→∞
59
The operator A is bounded. Indeed, An is Cauchy and hence bounded: there is constant
M ∈ R such that kAn kop < M for all n. Taking the limit in the inequality kAn uk ≤ Mkuk
implies kAuk ≤ Mkuk. Therefore A ∈ B(U,V ).
Finally, An → A in the operator norm. Indeed, Since An is Cauchy, for any ε > 0
there is N such that kAn − Am kop < ε or
kAn (u) − Am (u)k ≤ εkuk for all m, n > N.
Taking the limit as m → ∞
kAn (u) − A(u)k ≤ εkuk for all n > N.
Consequently kAn − Ak ≤ ε and so An → A. Therefore B(U,V ) is complete.
10.2 Examples
1. Example: Shift operator: Tl , Tr : `2 → `2 :
Tr (x) = (0, x1 , x2 , x3 , . . .) and Tl (x) = (x2 , x3 , x4 , . . .) .
Both operators are obviously linear. Moreover,

∞
kTr (x)k2`2 = ∑ |xk |2 = kxk`2 .
k=1
Consequently, kTr kop = 1. We also have

∞
kTl (x)k2`2 = ∑ |xk |2 ≤ kxk`2 .
k=2
Consequently, kTl kop ≤ 1. However, if x = (0, x2 , x3 , x4 , . . .) then kTl (x)k`2 =

kxk`2 . Therefore kTl kop = 1.
2. Example: Multiplication operator: Let f be a continuous function on [a, b]. The

equation
(Ax)(t) = f (t)x(t)
defines a bounded linear operator A : L2 [a, b] → L2 [a, b]. Indeed, A is obviously
linear. It is bounded since
Z b Z b
kAxk2 = | f (t)x(t)|2 dt ≤ k f k2∞ |x(t)| dt = k f k2∞ kxkL2 .
a a
Consequently kAkop ≤ k f k∞ . Now let t0 be a maximum of f . If t0 6= b, consider

the characteristic function
xε = χ[t0 ,t0 +ε] .
60
(If t0 = b let xε = χ[t0 −ε,t0 ] .) Since f is continuous,
Z t0 +ε
kAxε k 1
= | f (t)|2 dt → | f (t)|2 as ε → 0.
kxε k ε t0
Therefore kAkop = k f k∞ .
3. Example: Integral operator on L2 (a, b):

Z b
(Ax)(t) = K(t, s)x(s) ds for all t ∈ [a, b],
a
where Z bZ b
|K(s,t)| ds dt < +∞ .
a a
Let us estimate the norm of A:
Z b Z b 2
2

kAxk = K(t, s)x(s)ds dt
a a

Z b Z b Z b
2 2
≤ |K(t, s)| ds |x(s)| ds dt (Cauchy-Schwartz)
a a a
Z bZ b
= |K(t, s)|2 dsdt kxk2 .
a a
Consequently
Z bZ b
kAk2op ≤ |K(t, s)|2 dsdt .
a a
Note that this example requires a bit more from the theory of Lebesgue integrals
than we discussed in Section 5. If you are not taking Measure Theory and feel
uncomfortable with these integrals, you may assume that x, y and K are continu-
ous functions.
10.3 Kernel and range

Definition 10.6 Kernel of A:
Ker A = { x ∈ U : Ax = 0 }
Range of A:
Range A = { y ∈ V : ∃x ∈ U such that y = Ax }
We note that 0 ∈ Ker A for any linear operator A. We say that Ker A is trivial if
Ker A = { 0 }.
Proposition 10.7 If A ∈ B(U,V ) then Ker A is a closed linear subspace of U.
61
Proof: If x, y ∈ Ker A and α, β ∈ K, then
K(αx + β y) = αK(x) + β K(y) = 0 .
Consequently αx + β y ∈ Ker A and it is a linear subspace. Furthermore if xn → x and

A(xn ) = 0 for all n, then A(x) = 0 due to continuity of A.
Note that the range is a linear subspace but not necessarily closed (see Examples
3).
62
11 Linear functionals
Definition 11.1 If U is a vector space then a linear map U → K is called a linear
functional on U,
Definition 11.2 The space of all continuous functionals on a normed space U is called
the dual space, i.e., U ∗ = B(U, K) .
The dual space equipped with the operator norm is Banach. Indeed, K = R or C
which are both complete. Then Theorem 10.5 implies that U ∗ is Banach.
1. Example: δx ( f ) = f (x), x ∈ [a, b], is a bounded linear functional on C[a, b].

Rb
2. Example: Let φ ∈ C[a, b] (or φ ∈ L2 (a, b)). Then `φ (x) = a φ (t)x(t) dt is a
bounded linear functional on L2 (a, b).
11.2 Riesz representation theorem

Let H be a Hilbert space. Then for any y ∈ H
`y (x) = (x, y)
is a bounded functional `y : H → K (by the Cauchy-Schwartz inequality). The follow-

ing theorem is one of the fundamental results of Functional Analysis: it states that the
map y 7→ `y is an isometry between H and its dual space H ∗ .
Theorem 11.3 (Riesz Representation Theorem) Let H be a Hilbert space. For any
bounded linear functional f : H → K there is a unique y ∈ H such that
f (x) = (x, y) for all x ∈ H.
Moreover, k f kH ∗ = kykH .
Proof: Let K = Ker f . It is a closed linear subspace of H. If K = H then f (x) = 0

for all x and the statement of the theorem is true with y = 0. Otherwise K ⊥ 6= {0} and
there is a vector z ∈ K ⊥ with kzkH = 1.
Now we show that dim K ⊥ = 1. Indeed, let u ∈ K ⊥ . Since K ⊥ is a linear subspace
v = f (z)u − f (u)z ∈ K ⊥ . On the other hand
f (v) = f ( f (z)u − f (u)z) = f (z) f (u) − f (u) f (z) = 0
and so v ∈ K. For any linear subspace K ∩ K ⊥ = { 0 }, and so v = 0. Then f (z)u −

f (u)z = v = 0, i.e. u = ff (u) ⊥
(z) z. Consequently { z } is the basis in K .
63
Theorem 8.7 implies that every vector x ∈ H can be written uniquely in the form
x = u+v where u ∈ K and v ∈ K ⊥ .
Since K ⊥ is one dimensional and kzk = 1, u = (x, z)z. Moreover,
f (x) = f (u) + f (v) = f (u) = (x, z) f (z) = (x, f (z) z) .
Set y = f (z) z to get the desired equality:
f (x) = (x, y) .
If there is another y0 ∈ H such that f (x) = (x, y0 ) for all x ∈ H, then (x, y) = (x, y0 ) for
all x, i.e., (x, y − y0 ) = 0. Setting x = y − y0 we conclude ky − y0 k2 = 0, i.e. y = y0 is
unique.
Finally, the Cauchy-Schwartz inequality implies
| f (x)| = |(x, y)| ≤ kxk kyk,
i.e., k f kH ∗ = k f kop ≤ kyk. On the other hand,
| f (y)| |(y, y)|

k f kop ≥ = = kyk .
kyk kyk
Consequently, k f kH ∗ = kykH .
64
12 Linear operators on Hilbert spaces
12.1 Complexification
In the next lectures we will discuss the spectral theory of linear operators. The spectral
theory looks more natural in complex spaces. For example, a part of the theory studies
eigenvalues and eigenvectors of linear maps (i.e. non-zero solutions of the equation
Ax = λ x). In the finite-dimensional space a linear operator can be describe by a ma-
trix. You already know that a matrix (even a real one) can have complex eigenvalues.
Fortunately a real Hilbert space can always be considered as a part of a complex one
due to the “complexification” procedure.
Definition 12.1 Let H be a Hilbert space of R. The complexification of H is the vector

space
HC = { x + iy : x, y ∈ H }
where the addition and multiplication are respectively defined by
(x + iy) + (u + iw) = (x + u) + i(y + w)
(α + iβ )(x + iy) = (αx − β y) + i(αy + β x) .
The inner product is defined by
(x + iy, u + iw) = (x, u) − i(x, w) + i(y, u) + (y, w) .
Exercise: Show that HC is a Hilbert space.
Exercise: Show that kx + iyk2HC = kxk2 + kyk2 for all x, y ∈ H.
The following lemma states that any bounded operator on H can be extended to a
bounded operator on HC .
Lemma 12.2 Let H be a real Hilbert space and A : H → H be a bounded operator.

Then
AC (x + iy) = A(x) + iA(y)
is a bounded operator HC → HC .
Exercise: Prove the lemma.
12.2 Adjoint operators

Theorem 12.3 If A : H → H is a bounded linear operator on a Hilbert space H, then
there is a unique bounded operator A∗ : H → H such that
(Ax, y) = (x, A∗ y) for all x, y ∈ H .
Moreover, kA∗ kop ≤ kAkop .
65
Proof: Let y ∈ H and f (x) = (Ax, y) for all x ∈ H. The map f : H → K is linear and
| f (x)| = |(Ax, y)| ≤ kAxk kyk ≤ kAkop kxk kyk
where we have used the Cauchy-Schwartz inequality. Consequently, f is a bounded

functional on H. The Riesz representation theorem implies that there is a unique z ∈ H
such that
(Ax, y) = (x, z) for all x ∈ H.
Define the function A∗ : H → H by A∗ y = z. Then
(Ax, y) = (x, A∗ y) for all x, y ∈ H.
First, A∗ is linear since for any x, y1 , y2 ∈ H and α1 , α2 ∈ K
(x, A∗ (α1 y1 + α2 y2 )) = (Ax, α1 y1 + α2 y2 ) = ᾱ1 (Ax, y1 ) + ᾱ2 (Ax, y2 )

= ᾱ1 (x, A∗ y1 ) + ᾱ2 (x, A∗ y2 ) = (x, α1 A∗ y1 + α2 A∗ y2 ) .
Since the equality is valid for all x ∈ H, it implies
A∗ (α1 y1 + α2 y2 ) = α1 A∗ y1 + α2 A∗ y2 .
Second, A∗ is bounded since
kA∗ yk2 = (A∗ y, A∗ y) = (AA∗ y, y) ≤ kAA∗ yk kyk ≤ kAkop kA∗ yk kyk .
Dividing by kA∗ yk (do not forget to consider the case A∗ y = 0 separately), we conclude
that
kA∗ yk ≤ kAkop kyk .
Therefore A∗ is bounded and kA∗ kop ≤ kAkop .
Definition 12.4 The operator A∗ from Theorem 12.3 is called the adjoint operator.
1. Example: If A : Cn → Cn , then A∗ is the Hermitian conjugate of A, i.e. if

A∗ = ĀT (the complex conjugate of the transposed matrix).
2. Example: Integral operator on L2 (0, 1)

Z 1
(Ax)(t) = K(t, s)x(s)ds
0
The adjoint operator

Z 1
(A∗ y)(s) = K(t, s)y(t)dt
0
66
Indeed, for any x, y ∈ L2 (0, 1):
Z 1 Z 1
(Ax, y) = K(t, s)x(s) ds ȳ(t) dt
0 0
Z 1Z 1
= K(t, s)x(s)ȳ(t) ds dt
0 0
Z 1
!
Z 1
= x(s) K(t, s)y(t) dt ds
0 0
= (x, A∗ y) .
Note that we used Fubini’s Theorem to change the order of integration.
3. Example: Shift operators: Tl∗ = Tr and Tr∗ = Tl . Indeed,

∞
(Tr x, y) = ∑ xk ȳk+1 = (x, Tl y) .
k=1
The following lemma states some elementary properties of adjoint operators.
Lemma 12.5 If A, B : H → H are bounded operators on a Hilbert space H and α, β ∈

C, then
1. (αA + β B)∗ = ᾱA∗ + β̄ B∗
2. (AB)∗ = B∗ A∗
3. (A∗ )∗ = A
4. kA∗ k = kAk
5. kA∗ Ak = kAA∗ k = kAk2
Proof: Statements 1—3 follow directly from the definition of an adjoint operator (Ex-
ercise). Statement 4 follows from 3 and the estimate of Theorem 12.3: indeed,
kA∗ k ≤ kAk = k(A∗ )∗ k ≤ kA∗ k.
Finally since
kAxk2 = (AX, Ax) = (x, A∗ Ax) ≤ kxk kA∗ Axk ≤ kA∗ Ak kxk2
implies kAk2 ≤ kAA∗ k and on the other hand kA∗ Ak ≤ kA∗ k kAk = kAk2 , it follows
that kA∗ Ak = kAk2 .
67
12.3 Self-adjoint operators
Definition 12.6 A linear operator A is self-adjoint, if A∗ = A.
Lemma 12.7 An operator A ∈ B(H, H) is self-adjoint iff
(x, Ay) = (Ax, y) for all x, y ∈ H.
1. Example: H = Rn , a linear map defined by a symmetric matrix is self-adjoint.
2. Example: H = Cn , a linear map defined by a Hermitian matrix is self-adjoint.
3. Example: A : L2 (0, 1) → L2 (0, 1)

Z 1
A f (t) = K(t, s) f (s) ds
0
with real symmetric K, K(t, s) = K(s,t), is self-adjoint.
Theorem 12.8 Let A be a self-adjoint operator on a Hilbert space H. Then all eigen-
values of A are real and the eigenvectors corresponding to distinct eigenvalues are
orthogonal.
Proof: Suppose Ax = λ x with x 6= 0. Then
λ kxk2 = (λ x, x) = (Ax, x) = (x, A∗ x) = (x, Ax) = (x, λ x) = λ̄ kxk2 .
Consequently, λ is real.
Now if λ1 and λ2 are distinct eigenvalues and Ax1 = λ1 x1 , Ax2 = λ2 x2 , then
0 = (Ax1 , x2 ) − (x1 , Ax2 ) = (λ1 x1 , x2 ) − (x1 , λ2 x2 ) = (λ1 − λ2 ) (x1 , x2 ).
Since λ1 − λ2 6= 0, we conclude (x1 , x2 ) = 0.
Exercise: Let A be a self-adjoint operator on a real space. Show that the complexifi-
cation of A is also self-adjoint and has the same eigenvalues as the original operator
A.
Theorem 12.9 If A is a bounded self-adjoint operator then
1. (Ax, x) is real for all x ∈ H
2. kAkop = supkxk=1 |(Ax, x)|
68
Proof: For any x ∈ H
(Ax, x) = (x, Ax) = (Ax, x)
which implies (Ax, x) is real. Now let
M = sup |(Ax, x)| .

kxk=1
The Cauchy-Schwartz inequality implies
|(Ax, x)| ≤ kAxk kxk ≤ kAkop kxk2 = kAkop
for all x ∈ H such that kxk = 1. Consequently M ≤ kAkop . On the other hand, for any
u, v ∈ H we have
4 Re(Au, v) = (A(u + v), u + v) − (A(u − v), u − v)

≤ M ku + vk2 + ku − vk2

= 2M kuk2 + kvk2

using the parallelogram law. If Au 6= 0 let
kuk
v= Au
kAuk
to obtain, since kuk = kvk, that
kuk kAuk ≤ Mkuk2 .
Consequently kAuk ≤ Mkuk (for all u, including those with Au = 0) and kAkop ≤ M.
Therefore kAkop = M.
69
Unbounded operators and their adjoint operators16
The notions of adjoint and self-adjoint operators play an important role in the general
theory of linear operators. If an operator is not bounded a special care should be taken
in the consideration of its domain of definition.
Let D(A) be a linear subspace of a Hilbert space H, and A : D(A) → H be a linear
operator. If D(A) is dense in H we say that A is densely defined.
Example: Consider the operator A( f ) = ddtf on the set of all continuously differentiable
functions, i.e., D(A) = C1 [0, 1] ⊂ L2 (0, 1). This operator is densely defined.
Given a densely defined linear operator A on H, its adjoint A∗ is defined as follows:
• D(A∗ ), the domain of A∗ , consists of all vectors x ∈ H such that
y 7→ (x, Ay)
is a continuous linear functional D(A) → K. By continuity and density of D(A),

it extends to a unique continuous linear functional on all of H.
• By the Riesz representation theorem, if x ∈ D(A∗ ), there is a unique vector z ∈ H

such that
(x, Ay) = (z, y) for all y ∈ D(A).
This vector z is defined to be A∗ x.
It can be shown that the dependence of A∗ : D(A∗ ) → H is linear.

Note that two properties play a key role in this definition: the density of the domain
of A in H, and the uniqueness part of the Riesz representation theorem.
A linear operator is symmetric if (Ax, y) = (x, Ay) for all x, y ∈ D(A). From this
definition we see that D(A) ⊆ D(A∗ ) and symmetric A coincides with the restriction of
A∗ onto D(A). An operator is self adjoint if A = A∗ , i.e., it is symmetric and D(A) =
D(A∗ ). In general, the condition for a linear operator on a Hilbert space to be self-
adjoint is stronger than to be symmetric. If an operator is bounded then it is normally
assumed that D(A) = D(A∗ ) = H and therefore a symmetric operator is self-adjoint.
The Hellinger-Toeplitz theorem states that an everywhere defined symmetric oper-
ator on a Hilbert space is bounded.
16 Optional topic
70
13 Introduction to Spectral Theory
13.1 Point spectrum
Let H be a complex Hilbert space and A : H → H a linear operator. If Ax = λ x for
some x ∈ H, x 6= 0, and λ ∈ C, then λ is an eigenvalue of A and x is an eigenvector.
The space
Eλ = { x ∈ H : Ax = λ x }
is called the eigenspace.
Exercise: Prove the following: If A ∈ B(H, H) and λ is an eigenvalue of A, then Eλ is
a closed linear subspace in H. Moreover, Eλ is invariant, i.e., A(Eλ ) = Eλ .
Definition 13.1 The point spectrum of A consists of all eigenvalues of A:
σ p (A) = { λ ∈ C : Ax = λ x for some x ∈ H, x 6= 0 } .
Proposition 13.2 If A : H → H is bounded and λ is its eigenvalue then
kλ k ≤ kAkop .
Proof: If Ax = λ x with x 6= 0, then
kAyk kAxk
kAkop = sup ≥ = |λ | .
y6=0 kyk kxk
Examples:
1. A linear map on an n-dimensional vector space has at least one and at most n
different eigenvalues.
2. The right shift Tr : `2 → `2 has no eigenvalues, i.e., the point spectrum is empty.
Indeed, suppose Tr x = λ x, then
(0, x1 , x2 , x3 , x4 , . . .) = λ (x1 , x2 , x3 , x4 , . . . )
implies 0 = λ x1 , x1 = λ x2 , x2 = λ x3 , . . . If λ 6= 0, we divide by λ and conclude
x1 = x2 = · · · = 0. If λ = 0 we also get x = 0. Consequently
σ p (Tr ) = 0.
/
3. The point spectrum of the left shift Tl : `2 → `2 is the open unit disk. Indeed,
suppose Tl x = λ x with λ ∈ C. Then
(x2 , x3 , x4 , . . .) = λ (x1 , x2 , x3 , x4 , . . . )
is equivalent to x2 = λ x1 , x3 = λ x2 , x4 = λ x3 , . . . Consequently, x = (xk )∞
k=1 with
xk = λ k−1 x1 for all k ≥ 2. This sequence belongs to `2 if and only if ∑∞ 2
k=1 |xk | =
2k
∑∞k=1 |x1 | |λ | converges or equivalently |λ | < 1. Therefore
σ p (Tl ) = { λ ∈ C : |λ | < 1 } .
71
13.2 Invertible operators
Let us discuss the concept of an inverse operator.
Definition 13.3 (injective operator) We say that A : U → V is injective if the equation
Ax = y has a unique solution for every y ∈ Range(A).
Definition 13.4 (bijective operator) We say that A : U → V is bijective if the equation
Ax = y has exactly one solution for every y ∈ V .
Definition 13.5 (inverse operator) We say that A is invertible if it is bijective. Then
the equation Ax = y has a unique solution for all y ∈ V and we define A−1 y = x.
1. Exercise: Show that A−1 is a linear operator.
2. Exercise: Show that if A−1 is invertible, then A−1 is also invertible and
(A−1 )−1 = A.
3. Exercise: Show that if A and B are two invertible linear operators, then AB is
also invertible and (AB)−1 = B−1 A−1 .
Proposition 13.6 A linear operator A : U → V is invertible iff
Ker(A) = { 0 } and Range(A) = V.
Exercise: prove the proposition.
We will use IV : V → V to denote the identity operator on V , i.e., IV (x) = x for all
x ∈ V . Moreover, we will skip the subscript V if there is no danger of a mistake. It is
easy to see that if A : U → V is invertible then
AA−1 = IV and A−1 A = IU .
Example: The right shift Tr : `2 → `2 has a trivial kernel and

Tl Tr = I.
but it is not invertible since Range(Tr ) 6= `2 . (Indeed, any sequence in the range of Tr
has a zero on the first place). Consequently, the equality AB = I along does not imply
that B = A−1 .
Lemma 13.7 If A : U → V and B : V → U are linear operators such that
AB = IV and BA = IU
then A and B are both invertible and B = A−1 .
Proof: The equality ABy = y for all y ∈ V implies that Ker B = { 0 } and Range A = V .
On the other hand BAx = x for all x ∈ U implies Ker A = { 0 } and Range B = U.
Therefore both A and B satisfy the definition of invertible operator.
72
13.3 Resolvent and spectrum
Let A : V → V be a linear operator on a vector space V . A complex number λ is an
eigenvalue of A if Ax = λ x for some x 6= 0. This equation is equivalent to (A−λ I)x = 0.
Then we immediately see that A − λ I is not invertible since 0 has infinitely many
preimages: α x with α ∈ C.
If V is finite dimensional the reversed statement is also true: if A − λ I is not invert-
ible then λ is an eigenvalue of A (recall the Fredholm alternative from the first year
Linear Algebra). In the infinite dimensional case this is not necessarily true.
Definition 13.8 (resolvent set and spectrum) The resolvent set of a linear operator
A : H → H is defined by
R(A) = { λ ∈ C : (A − λ I)−1 ∈ B(H, H) } .
The resolvent set consists of regular values. The spectrum is the complement to the
resolvent set in C:
σ (A) = C \ R(A) .
Note that the definition of the resolvent set assumes existence of the inverse opera-
tor (A−λ I)−1 for λ ∈ R(A). If λ ∈ σ p (A) then (A−λ I) is not invertible. Consequently
any eigenvalue λ ∈ σ (A) and
σ p (A) ⊆ σ (A) .
The spectrum of A can be larger than the point spectrum.
Example: The point spectrum of the right shift operator Tr is empty but since
Range Tr 6= `2 it is not invertible and therefore 0 ∈ σ (Tr ). So σ p (Tr ) 6= σ (Tr ).
Technical lemmas
Lemma 13.9 If T ∈ B(H, H) and kT k < 1, then (I − T )−1 ∈ B(H, H). Moreover
(I − T )−1 = I + T + T 2 + T 3 + . . .
and
k(I − T )−1 k ≤ (1 − kT k)−1 .
Proof: Consider the sequence Vn = I + T + T 2 + · · · + T n . Since

kT n xk ≤ kT k kT n−1 xk
we conclude that kT n k ≤ kT kn . Consequently for any m > n we have
kVm −Vn k = T n+1 + T n+2 + · · · + T m

≤ kT kn+1 + kT kn+2 + · · · + kT km
kT kn+1 − kT km+1 kT kn+1
= ≤ .
1 − kT k 1 − kT k
73
Since kT k < 1, Vn is a Cauchy sequence in the operator norm. The space B(H, H) is
complete and there is V ∈ B(H, H) such that Vn → V . Moreover,
kV k ≤ 1 + kT k + kT k2 + · · · = (1 − kT k)−1 .
Finally, taking the limit as n → ∞ in the equalities
Vn (I − T ) = Vn −Vn T = I − T n+1 ,
(I − T )Vn = Vn − TVn = I − T n+1
and using that T n+1 → 0 in the operator norm we get V (I − T ) = (I − T )V = I.

Lemma 13.7 implies (I − T )−1 = V .
Lemma 13.10 Let H be a Hilbert space and T, T −1 ∈ B(H, H). If U ∈ B(H, H) and
kUk < kT −1 k−1 , then the operator T +U is invertible and
(T +U)−1 ≤
kT −1 k
.
1 − kUk kT −1 k
Proof: Consider the operator V = T −1 (T +U) = I + T −1U. Since
kT −1Uk ≤ kT −1 k kUk < 1,
Lemma 13.9 implies that V is invertible and

−1
kV −1 k ≤ 1 − kT −1 k kUk .
Moreover, since both V and T are invertible
I = V −1V = V −1 T −1 (T +U)
implies that T +U is invertible with
(T +U)−1 = V −1 T −1 .
Finally, k(T +U)−1 k ≤ kV −1 k kT −1 k implies the desired upper bound for the norm of
the inverse operator.
74
Properties of the spectrum
Lemma 13.11 If A is bounded and λ ∈ σ (A) then λ̄ ∈ σ (A∗ ).
Proof: If λ ∈ R(A) then A − λ I has a bounded inverse:
(A − λ I)(A − λ I)−1 = I = (A − λ I)−1 (A − λ I) .
Taking adjoints we obtain

∗ ∗
(A − λ I)−1 (A∗ − λ̄ I) = I = (A∗ − λ̄ I) (A − λ I)−1 .
∗
Consequently, (A∗ − λ̄ I) has a bounded inverse (A − λ I)−1 (an adjoint of a bounded
operator). Therefore λ ∈ R(A) iff λ̄ ∈ R(A∗ ). Since the spectrum is the complement
of the resolvent set we also get λ ∈ σ (A) iff λ̄ ∈ σ (A∗ ).
Proposition 13.12 If A is bounded and λ ∈ σ (A) then |λ | ≤ kAkop .
Proof: Take λ ∈ C such that |λ | > kAkop . Since kλ −1 Akop < 1 Lemma 13.9 implies
that I −λ −1 A is invertible and the inverse operator is bounded. Consequently, A−λ I =
−λ (I − λ −1 A) also has a bounded inverse and so λ ∈ R(A). The proposition follows
immediately since σ (A) is the complement of R(A).
Proposition 13.13 If A is bounded then R(A) is open and σ (A) is closed.
Proof: Let λ ∈ R(A). Then T = (A − λ I) has a bounded inverse. Set U = −δ I.

Obviously, kUk = |δ |. If
|δ | ≤ kT −1 k−1
Lemma 13.10 implies that T + U = A − (λ + δ )I also has a bounded inverse and so
λ + δ ∈ R(A). Consequently R(A) is open and σ (A) = C \ R(A) is closed.
Example: The spectrum of Tl and of Tr are both equal to the closed unit disk on the
complex plane.
Indeed, σ p (Tl ) = { λ ∈ C : |λ | < 1 }. Since σ p (Tl ) ⊂ σ (Tl ) and σ (Tl ) is closed, we
conclude that σ (Tl ) includes the closed unit disk. On the other hand, Proposition 13.12
implies that σ (Tl ) is a subset of the closed disk |λ | ≤ kTl kop = 1. Therefore
σ (Tl ) = { λ ∈ C : |λ | ≤ 1 } .
Since Tr = Tl∗ and σ (Tl ) is invariant, Lemma 13.11 implies σ (Tr ) = σ (Tl ).
75
13.4 Compact operators
Definition 13.14 Let X be a normed space and Y be a Banach space. Then a linear
operator A : X → Y is compact if the image of any bounded sequence has a convergent
subsequence.
Obviously a compact operator is bounded. Indeed, otherwise there is a sequence

xn with kxn k = 1 such that kAxn k > n for each n. The sequence Axn does not contain
a convergent subsequence (it does not even contain a bounded subsequence) and so is
not sequentially compact.
Lemma 13.15 Let X be a normed space and Y be Banach. A linear operator A : X →

Y is compact iff the image of the unit sphere is sequentially compact.
Example: Any bounded operator with finite-dimensional range is compact. Indeed, in

a finite dimensional space any bounded sequence has a convergent subsequence.
Theorem 13.16 If X is a normed space and Y is a Banach space, then compact linear
operators form a closed linear subspace in B(X,Y ).
Proof: If K1 , K2 are compact operators and α1 , α2 ∈ K, then α1 K1 + α2 K2 is also com-

pact. Indeed, take any bounded sequence (xn ) in H. There is a subsequence xn1 j
such that K1 xn1 j converges. This subsequence is also bounded, so it contains a subse-
quence xn2 j such that K2 xn2 j converges. Obviously K1 xn2 j also converges and therefore
α1 K1 xn2 j + α2 K2 xn2 j is convergent and consequently α1 K1 + α2 K2 is compact There-
fore the compact operators form a linear subspace.
Let us prove that this subspace is closed. Let Kn be a convergent sequence of
compact operators: Kn → K in B(H, H). Take any bounded sequence (xn ) in X. Since
K1 is compact, there is a subsequence xn1 j such that K1 xn1 j converges. Since xn1 j is
bounded and K2 is compact, there is a subsequence xn2 j such that K2 xn2 j converges.
Repeat this inductively: for each k there is a subsequence xnk j of the original sequence
such that Kl xnk j converges as j → ∞ for all l ≤ k.
Consider the diagonal sequence y j = xn j j . Obviously (y j )∞j=k is a subsequence of
(xnk j )∞j=1 . Consequently Kl y j converges as j → ∞ for every l.
In order to show that K is compact it is sufficient to prove that Ky j is Cauchy:
kKy j − Kyl k ≤ kKy j − Kn y j k + kKn y j − Kn yl k + kKyl − Kn yl k

≤ kK − Kn k ky j k + kyl k + kKn y j − Kn yl k .
Given ε > 0 choose n sufficiently large to ensure that the first term is less than ε2 ,
then choose N sufficiently large to guarantee that the second term is less than ε2 for
all j, l > N. So Ky j is Cauchy and consequently converges. Therefore K is a compact
operator, and the subspace formed by compact operators is closed.
76
Proposition 13.17 The integral operator A : L2 (a, b) → L2 (a, b) defined by
Z b Z bZ b
(A f )(t) = K(t, s) f (s) ds with |K(t, s)|2 dsdt < ∞
a a a
is compact.
Proof: Let { ϕk : k ∈ N } be an orthonormal basis in L2 (a, b). Let κ jk = (Aϕ j , ϕk ) and

n ∈ N, and define an operator
n n
An f = ∑ ∑ κ jk ( f , ϕ j )ϕk .
k=1 j=1
Obviously dim Range(An ) = n and consequently An is compact. In order to complete

the proof we need to show that An → A in the operator norm.17
17 We will not discuss this proof further.
77
13.5 Spectral theory for compact self-adjoint operators
Lemma 13.18 Let H be an infinitely dimensional Hilbert space and T : H → H a
compact self-adjoint operator. Then at least one of λ± = ±kT kop is an eigenvalue of
T.
Proof: Assume T 6= 0 (otherwise the lemma is trivial). Since
kT kop = sup |(T x, x)|

kxk=1
there is a sequence xn ∈ H such that kxn k = 1 and (T xn , xn ) → ±kT kop = α. Since T

is compact, yn = T xn has a convergent subsequence. Relabel this subsequence as xn
and let y = limn→∞ T xn . Then
kT xn − αxn k2 = kT xn k2 − 2α(T xn , xn ) + α 2 ≤ 2α 2 − 2α(T xn , xn ) .
The right hand side converges to 0 as n → ∞. Consequently T xn − αxn → 0. On the

other hand T xn → y and consequently
xn → x = α −1 y .
The operator T is continuous and consequently T x = αx. Finally, since kxn k = 1 for
all n, we have kxk = 1, and consequently α is an eigenvalue.
Proposition 13.19 Let H be an infinitely dimensional Hilbert space and T : H → H

a compact self-adjoint operator. Then σ p (T ) is either a finite set or countable se-
quence tending to zero. Moreover, every non-zero eigenvalue corresponds to a finite
dimensional eigenspace.
Proof: Suppose there is ε > 0 such that T has infinitely many different eigenvalues
with |λn | > ε. Let xn be corresponding eigenvectors with kxn k = 1. Since the operator
is self-adjoint, this sequence is orthonormal and for any n 6= m
kT xn − T xm k2 = kλn xn − λm xm k2
= (λn xn − λm xm , λn xn − λm xm ) = |λn |2 + |λm |2 > 2ε .
Consequently, (T xn ) does not have a convergent subsequence. This contradicts to

the compactness of T . Consequently, σ p (T ) is either finite or a converging to zero
sequence.
Now let λ 6= 0 be an eigenvalue and Eλ the corresponding eigenspace. Let Ã :
Eλ → Eλ be the restriction of A onto Eλ . Since Ãx = λ x for any x ∈ Eλ , the operator
Ã maps the unit sphere into the sphere of radius λ . Since A is compact, the image of
the unit sphere is sequentially compact. Therefore the sphere of radius λ is compact.
Since Eλ is a Hilbert (and consequently Banach) space itself, Theorem 3.23 implies
that Eλ is finite dimensional.
78
Theorem 13.20 (Hilbert-Schmidt theorem) Let H be a Hilbert space and T : H → H
be a compact self-adjoint operator. Then there is a finite or countable orthonormal
sequence (en ) of eigenvectors of T with corresponding real eigenvalues (λn ) such that
T x = ∑ λ j (x, e j )e j for all x ∈ H.

j
Proof: We construct the sequence e j inductively. Let H1 = H and T1 = T : H1 →

H1 . Lemma 13.18 implies that there is an eigenvector e1 ∈ H1 with ke1 k = 1 and an
eigenvalue λ1 ∈ R such that |λ1 | = kT1 kB(H1 ,H1 ) .
Then let H2 = {x ∈ H1 : x ⊥ e1 }. If x ∈ H2 then T x ∈ H2 . Indeed, since T is self
adjoint
(T x, e1 ) = (x, Te1 ) = λ1 (x, e1 ) = 0
and T x ∈ H2 . Therefore the restriction of T onto H2 is an operator T2 : H2 → H2 .
Since H2 is an orthogonal complement, it is closed and so a Hilbert space itself.
Lemma 13.18 implies that there is an eigenvector e2 ∈ H2 with ke2 k = 1 and an eigen-
value λ2 ∈ R such that |λ2 | = kT2 kB(H2 ,H2 ) . Then let H3 = { x ∈ H2 : x ⊥ e2 } and repeat
the procedure as long as Tn is not zero.
Suppose Tn = 0 for some n ∈ N. Then for any x ∈ H let
n−1
y = x − ∑ (x, e j )e j .
j=1
Applying T to the equality we get:

n−1 n−1
Ty = T x − ∑ (x, e j )Te j = T x − ∑ (x, e j )λ j e j .
j=1 j=1
Since y ⊥ e j for j < n we have y ∈ Hn and consequently Ty = Tn y = 0. Therefore

n−1
Tx = ∑ (x, e j )λ j e j
j=1
which is the required formula for T .

Suppose Tn 6= 0 for all n ∈ N. Then for any x ∈ H and any n consider
n−1
yn = x − ∑ (x, e j )e j .
j=1
Since yn ⊥ e j for j < n we have y ∈ Hn and

n−1
kxk2 = kyn k2 + ∑ |(x, e j )|2
j=1
79
implies kyn k2 ≤ kxk2 . On the other hand

n−1
T x − ∑ (x, e j )λ j e j = kTyn k ≤ kTn k kyn k ≤ |λn | kxk

j=1

and since λn → 0 as n → ∞ we have

∞
Tx = ∑ (x, e j )λ j e j .
j=1
Corollary 13.21 Let H be an infinite dimensional separable Hilbert space and T :

H → H a compact self-adjoint operator. Then there is an orthonormal basis E = { e j :
j ∈ N } in H such that Te j = λ j e j for all j ∈ N and
∞
Tx = ∑ λ j (x, e j )e j for all x ∈ H.
j=1
Exercise: Deduce that operators with finite range are dense among compact self-
adjoint operators.
Theorem 13.22 If H is an infinite dimensional Hilbert space and T : H → H is a

compact self-adjoint operator, then σ (T ) = σ p (T ).
Proposition 13.19 implies that zero is the only possible limit point of σ p (T ). There-
fore the theorem means that either σ (T ) = σ p (T ) or σ (T ) = σ p (T ) ∪ { 0 }. In partic-
ular, σ (T ) = σ p (T ) if zero is an eigenvalue. Note that if zero is not an eigenvalue,
then there is a sequence of eigenvalues which accumulates to zero since H is infinite
dimensional.
Proof: According to the Hilbert-Schmidt theorem
∞
Tx = ∑ λ j (x, e j )e j
j=1
where { e j } is an orthonormal basis in H.18 Then x = ∑∞j=1 (x, e j )e j and for any µ ∈ C
∞
(T − µI)x = ∑ (λ j − µ)(x, e j )e j .
j=1
18 The proof uses a countable basis in H, therefore it assumes that His separable. The theorem remain
valid for a non-separable H but the proof should be slightly modified. The modification is based on the
following observation: Proposition 13.19 implies that (Ker T )⊥ has a countable orthonormal basis of
eigenvectors { e j }. Then for any vector x ∈ H write x = PKer T (x) + ∑∞j=1 (x, e j )e j where PKer T is the
orthogonal projection on the kernel of T . Then follow the arguments of the proof (adding µ −1 PKer T (y)
to the definition of S).
80
Let µ ∈ C \ σ p (T ) which is an open subset of C. Consequently there is ε > 0 such that
|µ − λ | > ε for all λ ∈ σ p (T ) ⊂ σ p (T ). Consider an operator S defined by
∞
(y, e )
Sy = ∑ λk −kµ ek .
k=1
Lemma 7.12 implies that the series converges since |λk − µ| > ε and
(y, ek ) 2

∞
∞
2
kSyk = ∑ ≤ ε −2 ∑ |(y, ek )|2 = ε −2 kyk2 .
k=1 λk − µ

j=k
In particular we see that S is bounded with kSkop ≤ ε −1 . Moreover S = (T − µI)−1 .

Indeed,
∞ ∞ λj − µ
(T − µI)Sy = ∑ (λ j − µ)(Sy, e j )e j = ∑ λ j − µ (y, e j )e j = y .
j=1 j=1
and ∞ ((T − µI)x, e j ) ∞ λj − µ

S(T − µI)x = ∑ λj − µ ej = ∑ λ j − µ (x, e j )e j = x .
j=1 j=1
Then S = (T − µI)−1 and µ ∈ R(T ), and so σ (T ) ⊆ σ p (T ). On the other hand,

σ p (T ) ⊆ σ (T ). We conclude σ (T ) = σ p (T ).
81
14 Sturm-Liouville problems
In this chapter we will study the Sturm-Liouville problem:

d du
− p(x) + q(x)u = λ u with u(a) = u(b) = 0
dx dx
where p and q are given function. The values of λ for which the problem has a non-
trivial solution are called eigenvalues of the Sturm-Liouville problem and the corre-
sponding solutions u are called eigenfunctions. We will prove that the eigenfunctions
form an orthonormal basis in L2 (a, b).
We assume that p ∈ C1 [a, b], q ∈ C[a, b] and
p(x) > 0 and q(x) ≥ 0 for x ∈ [a, b].
For a function u ∈ C2 [a, b] we define

d du
L(u) = − p(x) + q(x)u .
dx dx
Obviously L : C2 [a, b] → C0 [a, b] is linear. We will see that Range L = C0 [a, b] and
Ker L = Span{ u1 , u2 } where u1 , u2 are two linear independent solutions of the equation
Lu = 0. Therefore the operator is not invertible.
We will restrict L onto the space
D(L) := u ∈ C2 [a, b] : u(a) = u(b) = 0

and show that L : D(L) → C0 [a, b] is invertible. We will prove that L−1 is a restriction
on Range L of a compact self-adjoint operator A : L2 (a, b) → L2 (a, b). The Sturm-
Liouville theorem states that the eigenfunction of A form an orthonormal basis in
L2 (a, b). Moreover, we will see that A and L have the same eigenfunctions.
Lemma 14.1 If both u1 and u2 are solutions of the equation Lu = 0, then
Wp (u1 , u2 )(x) = p(x)(u01 (x)u2 (x) − u1 (x)u02 (x))
is constant. Moreover, the solutions u1 and u2 are linearly independent iff
Wp (u1 , u2 ) 6= 0.
Proof: The equation Lu = 0 implies pu00 = −p0 u0 + qu. Then differentiate W with
respect to x:
Wp0 = p0 (u01 u2 − u1 u02 ) + p(u001 u2 − u1 u002 )

= p0 (u01 u2 − u1 u02 ) + ((−p0 u01 + qu1 )u2 − (−p0 u02 + qu2 )u1 ) = 0 .
82
Therefore Wp is constant.
Finally, since p 6= 0, Wp = 0 implies u01 (x)u2 (x) − u1 (x)u02 (x) = 0 for all x. Conse-
quently at all points where u1 and u2 do not vanish
u01 u02 d ln |u1 | d ln |u2 |
= ⇔ = ,
u1 u2 dx dx
which implies that u1 = Cu2 for some constant C, i.e., if Wp = 0 then u1 and u2 are not
linearly independent. In the reverse direction the statement is straightforward.
Lemma 14.2 The equation Lu = 0 has two linearly independent solutions, u1 , u2 ∈
C2 [a, b], such that u1 (a) = u2 (b) = 0.
Proof: Let u1 , u2 be solutions of the Cauchy problems
Lu1 = 0 u1 (a) = 0, u01 (a) = 1,
Lu2 = 0 u2 (b) = 0, u02 (b) = 1 .
According to the theory of linear ordinary differential equations u1 and u2 exist, belong
to C2 [a, b] and are unique.
Moreover, u1 and u2 are linearly independent. Indeed, suppose Lu = 0 for some
u ∈ C2 [a, b] such that u(a) = u(b) = 0. Then
Z b
−(pu0 )0 u + qu2 dx

0 = (Lu, u) =
a
b Z b
0
p(u0 )2 + qu2 dx

= p(x)u (x)u(x) + (integration by parts)

a a
Z b
p(u0 )2 + qu2 dx

=
a
Since p > 0 on [a, b], we conclude that u0 ≡ 0. Then u(a) = u(b) = 0 implies u(x) = 0
for all x ∈ [a, b].
Consequently since u2 (b) = 0 and u2 is not identically zero, u2 (a) 6= 0 and so
Wp (u1 , u2 ) = p(a)(u01 (a)u2 (a) − u1 (a)u02 (a)) = p(a)u01 (a)u2 (a) 6= 0.
Therefore u1 , u2 are linearly independent by Lemma 14.1.
Proposition 14.3 If u1 and u2 are linearly independent solutions of the equation Lu =
0 such that u1 (a) = u2 (b) = 0 and

1 u1 (x)u2 (y), a ≤ x < y ≤ b,
G(x, y) =
Wp (u1 , u2 ) u 1 (y)u2 (x), a ≤ y ≤ x ≤ b,
then for any f ∈ C0 [a, b] the function
Z b
u(x) = G(x, y) f (y) dy
a
belongs to C2 [a, b], satisfies the equation Lu = f and the boundary conditions u(a) =
u(b) = 0.
83
Proof: The statement is proved by a direct substitution of
Z x Z b
u2 (x) u1 (x)
u(x) = u1 (y) f (y) dy + u2 (y) f (y) dy
Wp (u1 , u2 ) a Wp (u1 , u2 ) x
into the differential equation. Moreover, since u1 (a) = u2 (b) = 0, we get u(a) = u(b) =
0.
Lemma 14.4 The operator A : L2 (a, b) → L2 (a, b) defined by

Z b
(A f )(x) = G(x, y) f (y) dy .
a
is compact and self-adjoint. Moreover, Range A is dense in L2 (a, b), Ker A = { 0 } and
all its eigenfunctions, Au = µu, belong to C2 [a, b] and satisfy u(a) = u(b) = 0.
Proof: The operator A is compact by Proposition 13.17. Moreover, G is symmetric

and so A is self-adjoint. Proposition 14.3 implies the range of A contains all functions
from C2 [a, b] such that u(a) = u(b) = 0. This set is dense in L2 (a, b).
Now suppose Au = 0 for some u ∈ L2 [a, b]. Then for any v ∈ L2
0 = (Au, v) = (u, Av) ,
which implies u = 0 because u is orthogonal to a dense set.

Finally, let u be an eigenfunction of A, i.e., Au = µu. Since µ 6= 0 we can write
u = µ −1 Au = µ −2 A2 u. A priori u ∈ L2 (a, b), then Au is continuous and A2 u is C2 .
Therefore, the eigenvectors of A are smooth.
Proposition 14.5 The operator L : D(L) → C0 [a, b] has a bounded inverse (in the op-
erator norm induced by the L2 norm on both spaces).
Proof: Lemma 14.2 implies that Ker L = { 0 }. Consequently, L is injective. Proposi-

tion 14.3 implies that Range L = C0 [a, b] and the inverse operator is defined by
Z b
−1
(L f )(x) = G(x, y) f (y) dy .
a
Lemma 14.4 states that this operator is bounded. In other words, L−1 coincides with
the restriction of A onto Range L.
Theorem 14.6 The eigenfunctions of the Sturm-Liouville problem form an orthonor-

mal basis in L2 (a, b).
84
Proof: Since A : L2 (a, b) → L2 (a, b) is compact and self adjoint, Theorem 13.21 im-
plies that its eigenvectors form an orthonormal basis in L2 (a, b). If u is an eigenfunc-
tion of A, then Au = µu, µ 6= 0 and u ∈ C2 [a, b]. Consequently Lu = λ u = µ −1 u, i.e.,
u is also an eigenvector of L which corresponds to the eigenvalue λ = µ −1 .
Example: An application for Fourier series. Consider the Strum-Liouville problem
d2u
− = λ u, u(0) = u(1) = 0 .
dx2
It corresponds to the choice p = 1, q = 0. Theorem 14.6 implies that the normalised
eigenfunctions of this problem form an orthonormal basis in L2 (0, 1). In this example
the eigenfunctions are easy to find:

1
√ sin kπx : k ∈ N .
2
Consequently any function f ∈ L2 (0, 1) can be written in the form

∞
f (x) = ∑ αk sin kπx
k=1
where Z 1
1
αk = f (x) sin kπx dx .
2 0
The series converges in the L2 norm.
85

Fa1 2009 1

Transféré par

Informations du document

Description originale:

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Fa1 2009 1

Transféré par

Droits d'auteur :

Formats disponibles

Functional Analysis I

3 Convergence in a normed space 12

7 Orthonormal bases in Hilbert spaces 41

9 Separable Hilbert spaces 56

10 Linear maps between Banach spaces 58

12 Linear operators on Hilbert spaces 65

13 Introduction to Spectral Theory 71

• “Associativity” of multiplication:1 α(β x) = (αβ )x.

• There is an element 1 ∈ K such that 1x = x for all x ∈ V . This element is called

operations: αβ is a product of two scalars and β x involves a vector and scalar.

1.2 Examples of vector spaces

5. The set of all bounded sequences `∞ (K) is a vector space:

For two sequences x, y ∈ `∞ (K), we define x + y by

( f + g)(t) = f (t) + g(t) , t ∈ [0, 1] ,

and α f is the function whose values are

(α f )(t) = α f (t) , t ∈ [0, 1] ,

We will always use these definitions on all spaces of functions to be considered

i.e. f ∈ L̃1 (0, 1).

so f ∈ L̃1 (0, 1).

1.3 Hamel bases

We say that E spans V if V = Span(E), i.e. every element of V can be written as a

for any choice of n ∈ N, e j ∈ E and α j ∈ K.

Definition 1.3 A Hamel basis E for V is a linearly independent subset of V which

1. Any basis in Rn is a Hamel basis.

2. The set E = 1, x, x2 , . . . is a Hamel basis in the space of all polynomials.

Exercise: Prove the lemma.

Example: In Rn any basis consists of n vectors. Therefore dim Rn = n.

Definition 1.8 A map L : V → W is called linear if for any x, y ∈ V and any α ∈ K

L(x + αy) = L(x) + αL(y) .

Definition 1.9 If a linear map L : V → W is a bijection, then L is called a linear

1. kxk ≥ 0, and kxk = 0 ⇔ x = 0 (positive definiteness);

2. kαxk = |α| kxk (positive homogeneity);

3. kx + yk ≤ kxk + kyk (triangle inequality).

The pair (V, k · k) is called a normed space.

In other words, a normed space is a vector space equipped with a norm.

1. Rn with each of the following norms is a normed space:

2. ` p (K) is a vector space with the following norm (1 ≤ p < ∞)

3. `∞ (K) is a vector space with the following norm

kxk`∞ = sup |xk | .

Therefore for any n ∈ N

Lemma 2.4 (Cauchy-Schwartz inequality) If x, y ∈ `2 (K) then

Lemma 2.5 (Minkowski’s inequality) If x, y ∈ ` p (K) for 1 ≤ p ≤ ∞ then x + y ∈

2.3 Examples of norms on a space of functions

2. the “L1 norm” Z 1

3. the “L2 norm” 1/2

where f ( j) denotes the derivative of order j.

2.4 Equivalence of norms

c1 kxk1 ≤ kxk2 ≤ c2 kxk1 for all x ∈ V .

In this case we write k · k1 ∼ k · k2 .

Theorem 2.7 Any two norms on Rn are equivalent.3

Example: The norms k · kL1 and k · k∞ on C[0, 1] are not equivalent.

2.5 Linear Isometries

A linear invertible map can be used to “pull back” a norm as follows.

Proposition 2.10 Let (V, k · kV ) be a normed space, W a vector space, and L : W → V

Proof: For any x, y ∈ V and any α ∈ K we have:

If kxkW = kL(x)kV = 0, then L(x) = 0 due to non-degeneracy of the norm k · kV . Since

kx + ykW = kL(x) + L(y)kV ≤ kL(x)kV + kL(y)kV = kxkW + kykW .

Corollary 2.12 Any n-dimensional normed space V is isometrically isomorphic to Kn

n=1 , xn ∈ V , n ∈ N, converges to a limit x ∈ V if for any

kxn − xk < ε for all n > N.

Proposition 3.2 The limit of a convergent sequence is unique.

Exercise: Prove it.

3. the “L2 norm” 1/2