Scribd
Upload
13 views3 pages

Hilbert Spaces

The document discusses Hilbert spaces, defining them as vector spaces with inner products that are also Banach spaces. It covers key concepts such as the Cauchy-Schwarz inequality, the Parallelogram Identity, and orthogonal projections within Hilbert spaces. The document emphasizes the properties and significance of these mathematical structures in relation to convergence and completeness.

Uploaded by

maroaka10
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
13 views3 pages

Hilbert Spaces

The document discusses Hilbert spaces, defining them as vector spaces with inner products that are also Banach spaces. It covers key concepts such as the Cauchy-Schwarz inequality, the Parallelogram Identity, and orthogonal projections within Hilbert spaces. The document emphasizes the properties and significance of these mathematical structures in relation to convergence and completeness.

Uploaded by

maroaka10
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Table des matières

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Hilbert spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Orthogonal projections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Introduction
Hilbert spaces are the closest generalization to infinite dimensional spaces of the Euclidean
spaces.
First, we consider a normed space and we see that if the space is finite dimensional all the norms
defined on it generate the same topology, so that the convergence of sequences does not depend
on the norm that is used. Later we consider linear transformations defined in a normed space
and we see that all of them are continuous if the space is finite dimensional. Continuous linear
transformations are called bounded operators and, by introducing a norm in the space of all
bounded operators, we convert it into a Banach space under certain conditions. Once we have
clear the concepts of convergence of a sequence and completeness of the space, we define Hilbert
spaces and consider some of their properties. Mainly we focus on the Pythagorean theorem, the
Cauchy-Schwarz inequality, the Parallelogram Identity and in the introduction of the concept
of basis for a Hilbert space.

Hilbert spaces
Inner product
For simplicity from now on we only consider vector spaces over C. Let V be a vector space,
an inner product is a function h, i : V × V → C, such that for u, v, w ∈ V and λ ∈ C,
(i) hu, vi = hv, ui
(ii) hλu, vi = λhu, vi ;
(iii) hu + v, wi = hu, wi + hv, wi ;
(iv) hu, ui ≥ 0 and hu, ui = 0 if and only if u = 0.

Cauchy-Schwarz inequality
One of the most important inequalities in an inner product space is the Cauchy-Schwarz
inequality. It states the following,
for any u, v in an inner product space

|hu, vi| ≤ kukkvk


p
Using last inequality it is easy to prove that kuk = hu, ui is a norm in V .

Hilbert space
A Hilbert space is a vector space with inner product such that it is a Banach space with
the norm induced by the inner product.
The simplest example of a Hilbert space is Cn with the inner product :
n
X
hx, yi = xi yi ,
i=1

where x = (x1 , . . . , xn ) and y = (y1 , . . . , yn ). The metric induced by this inner product is the
usual Euclidean metric in Cn .

The space `2 is an infinite dimensional vector space that is a Hilbert space with the product :
X
ha, bi = an bn , where a = (a1 , a2 , . . .) , b = (b1 , b2 , . . .)
n

It is easy to prove that if kk is a norm induced by an inner product then,

2kuk2 + 2kvk2 = ku + vk2 + ku − vk2

The latter equation is called the Parallelogram Identity and is an important tool to determine
if a Banach space is a Hilbert space or not.

Orthogonal projections
Let H be a Hilbert space and M ⊂ H. For a point x ∈ H the distance from x to M is
defined by
d(x, M ) = inf{d(x, m) | m ∈ M }.
Theorem : Let H be a Hilbert space and M a closed subspace of H. Let x ∈ H\M , then there
exists a unique point z ∈ M such that

kx − zk = inf{d(h, m) | m ∈ M } = d(x, M ) := d.

Proof : Let (yn ) be a sequence of points in M such that

lim kx − ym k = d.
n→∞

Then, using the Parallelogram Identity, we have that

kyn + ym − 2xk2 + kyn − ym k2 = 2 kyn − xk2 + 2 kym − xk2 .

2
This fact implies that
2
yn + ym
kyn − ym k2 = 2 kyn − xk2 + 2 kym − xk2 − 4 −x .
2

Then,

kyn − ym k2 ≤ 2 kyn − xk2 + 2 kym − xk2 − 4d2 → 0 when n, m → ∞

Thus, there exists z ∈ M such that yn → z. Then,

d(x, M ) = lim kx − yn k = x − lim yn = kx − zk


n→∞ n→∞

Assume that there exists z1 ∈ M such that d = kx − z1 k . Then, the sequence z, z1 , z, z1 , z, z1 , . . .


is such that d = limn→∞ kx − wn k where,
(
z if n is even
wn =
z1 if n is odd

Since (x − wn ) converges, it is a Cauchy sequence which implies that (wn ) is a Cauchy sequence,
then z = z1 .

You might also like