Linear Functional Analysis: Lecture 1: Introduction Rynne and Youngson 1.1, 1.2

Linear Functional
Analysis
Lecture 1: Introduction
Rynne and Youngson §1.1, §1.2
1
2n
0 = (e1 , e2m
History for all m ∈ N, and hence

thonormality of the sequen
We now give various co
• Functional analysis is the child ofTheorem

the 20th
3.47
century (Stefan Banach, Hilbert,Let
Lebesgue)
H be a Hilbert space
• Fourier, Riemann, Lebesgue

following conditions are eq
(a) {en : n ∈ N}⊥ = {0};
(b) Sp {en : n ∈ N} = H;
$∞
(c) $x$ = n=1 |(x, en )|2
2
$∞
(d) x = n=1 (x, en )en for
2
Ingredients
• Linear algebra (vector spaces)

• Analysis (calculus)
• Measure and integration
3
2n
0 = (e1 , e2m
Why? for all m ∈ N, and hence

thonormality of the sequen
We now give various co
• We want problems to have solutions

Theorem 3.47
• We want the solutions to be unique
Let H be a Hilbert space
• We want to be able to calculate (a) {e : n ∈ N} = {0};

following conditions are eq
⊥
(approximate) the solution n
(b) Sp {en : n ∈ N} = H;
$∞
(c) $x$ = n=1 |(x, en )|2
2
$∞
(d) x = n=1 (x, en )en for
4
Examples
• We want problems to have solutions
• We want the solutions to be unique
• We want to be able to calculate (approximate) the solution
• Solving equations:
N → Z → Q → R → C ...
• Existence, uniqueness, computation of Fourier series ...

connections to solutions of differential equations ...
eigen-values and eigen-functions ...
waves sound light heat quantum physics ...
5
Bryan P. Rynne, Martin A. Youngson
Huishoudelijke
Linear Functional Analysi
Bryan P. Rynne, Martin A. Youngson
• Blackboard Linear Functional Analysis

• Homework
• Rynne and Youngson
• Schedule (Gill , de Jeu)
Springer-Verlag
Berlin Heidelberg NewYork
London Paris Tokyo
Hong Kong Barcelona
Budapest
Springer-Verlag 6
in the space. As real or complex-valued functions are often called functionals,
the term functional analysis came to be used for this topic.
We now briefly outline the contents of the book. In Chapter 1 we present
(for reference and to establish our notation) various basic ideas that will be
required throughout the book. Specifically, we discuss the results from elemen-
tary linear algebra and the basic theory of metric spaces which will be required
in later chapters. We also give a brief summary of the elements of the theory
of Lebesgue measure and integration. Of the three topics discussed in this in-
troductory chapter, Lebesgue integration is undoubtedly the most technically
difficult and the one which the prospective reader is leastLinear
VI likelyFunctional
to have encoun-
Analysis
V
tered before. Unfortunately, many of the most important spaces which arise in
functional analysis are spaces of integrable functions, and it is necessary to use
the Lebesgue integral to overcome various drawbacks of the elementary Rie-
mann integral, commonly taught in real analysis courses. The reader who has
not met Lebesgue integration before can still read this book by accepting that
an integration process exists which coincides with the Riemann integral when
this is defined, but extends to a larger class of functions, and which has the
properties described in Section 1.3.
In Chapter 2 we discuss the fundamental concept of functional analysis, the
normed vector space. As mentioned above, a norm on a vector space is simply an
extension of the idea of the length of a vector to a rather more abstract setting.
7
sequence lies in. only obtain a “standard” inner product when there is s
basis for the space.
Now letbut
A rather more complicated, (X, Σ, µ) be important,
extremely a measure example
space, and recall
is the fol- t
lowing. from Definition 1.53.
Example 3.39 Example 3.7

−1/21einx for n ∈ Z, is an or-
The set of functions {e ∈ L2 (X)
If nf,},gwhere en (x) = (2π)
then
" f g ∈ L (X) and the function (· ,
2
thonormal sequence indefined
the spaceby L C [−π,
(f, g) =π]. f g dµ is an inner product on L2
X
Linear Functional Analysis
Linear Functional Analysis
will be called the standard inner product on L2 (X).
Solution
aThis follows from Solution
∞
! Let f, !g π∈ L2 (X). Then "by Hölder’s inequality, with p
, ex =n
n )e (x, en )en 1 (3.5) i(m−n)x (3.5) 2 1, if m = n,
(em , en )and
= the definition of L (X),
e dx =
n=1 2π −π 0, if m #= n. %
$
# $# %1/2 $#
sd was
in Definition 2.29).
defined in However,2.29).
Definition in theHowever, in the
|f g| dµ ≤ |f |2
dµ |g|2
dµ
major questions associated with this
here are two major questions associated X with this X X
The orthonormal sequence in Example 3.39, and related "sequences of
trigonometric functions, L1 (X)
g ∈ be
so fwill and theinformula
considered (f, g)in=Section
more detail X
f g dµ
3.5ison
we
? Fourier series. We alsothat
notethethat, strictly
above speaking,
formula to use
defines anthe wordproduct
inner “sequence”
on L
in Example 3.39 we should
all thechoose an ordering
properties of the functions
in Definition 3.1 or 3.3so hold.
that they are8
It follow
Ch. 1 Section 2: metric
spaces
• M – compact, metric space
• F – field R or C
• C (M) [ aka C(M) ] – space of continuous functions
F
from compact metric space M to field F endowed with
the uniform metric
• C(M) is complete
• Stone-Weierstrass: for M subset R,
P R is dense in CR(M)
9

Linear Functional Analysis: Lecture 1: Introduction Rynne and Youngson 1.1, 1.2

Uploaded by

Linear Functional Analysis: Lecture 1: Introduction Rynne and Youngson 1.1, 1.2

Uploaded by

Linear Functional

History for all m ∈ N, and hence

We now give various co

• Functional analysis is the child ofTheorem

• Fourier, Riemann, Lebesgue

• Linear algebra (vector spaces)

Why? for all m ∈ N, and hence

We now give various co

• We want problems to have solutions

• We want to be able to calculate (a) {e : n ∈ N} = {0};

• We want the solutions to be unique

• We want to be able to calculate (approximate) the solution

• Existence, uniqueness, computation of Fourier series ...

Bryan P. Rynne, Martin A. Youngson

• Blackboard Linear Functional Analysis

Example 3.39 Example 3.7

You might also like