Functional Analysis Lecture 15

[[lecture-data]]

← Lecture 14 | Lecture 16 →
Lecture Notes: Rodriguez, page 77

3. Hilbert Spaces

Orthonormal Bases

Recall from last time, we introduced

Maximal orthonormal set

{eλ}λΛ in a [[Hilbert space]] is maximal if uH and u,eλ=0 for all λΛ implies u=0.

see [[maximal orthonormal set]]

Theorem

Let H be a [[separable]] [[Hilbert space]]. Then H has a countable [[maximal orthonormal set]].

See [[separable hilbert spaces have countable maximal orthonormal sets]]

Orthonormal Basis

Let H be a maximal orthonormal subset {en} of H.

see [[orthonormal basis of a hilbert space]]

Theorem

If {en} is an [[orthonormal basis of a hilbert space]] H, then for all uH we have

limmn=1mu,enen=n=1u,enen=u
Note

ie like in finite-dimensional linear algebra, we can write any element as a linear combination of the basis elements. But in this space, there might be an infinite number of such elements.

)

First, we prove {n=1mu,enen}m is Cauchy. Let ϵ>0. Then by [[Bessel's inequality]], we have

\sum_{n=1}^\infty \lvert \langle u, e_{n} \rangle \rvert { #2} \leq \lvert \lvert u \rvert \rvert { #2} < \infty

Thus, for all MN such that for all NM we have m=N+1|u,en|2<ϵ

Then for all m>M we compute

\begin{align} \left\lvert \left\lvert \sum_{n=1}^m \langle u, e_{n} \rangle e_{n} - \sum_{n=1}^\ell \langle u, e_{n} \rangle e_{n} \right\rvert \right\rvert &= \sum_{n=\ell+1}^m \lvert \langle u, e_{n} \rangle \rvert { #2} \\ &\leq \sum_{\ell+1}^\infty \lvert \langle u, e_{n} \rangle \rvert { #2} \\ &< \epsilon^2 \end{align}

thus the sequence is indeed Cauchy. Since H is complete, there exists some u¯H where

u¯=limmn=1mu,enen

By [[continuity of inner product]], we know that for all N

uu¯,e=limmun=1mu,enen,e=limm[u,en=1mu,enen,e]=u,eu,e1=0

Thus uu¯,e=0 for all if and only if uu¯=0.

see [[all elements of hilbert spaces with orthonormal bases can be written as sums of the basis elements]]

Thus if we have an orthonormal basis, every element can be expanded in this series in terms of the basis elements (Bessel-Fourier series). And thus every separable [[Hilbert space]] has an orthonormal basis.

Theorem

If a orthonormal basis, then H is [[separable]].

Proof

Suppose {en} is an orthonormal basis for H. Then

S=mN{n=1mqnen|q1,,qmQ+iQ}

is a countable. This is because elements in each component (index by m) have a bijection with Q2m, which is countable. Then since the m are countable, the union is countable and therefore yields a countable set.

So by [[all elements of hilbert spaces with orthonormal bases can be written as sums of the basis elements]], S is dense in H

So the sums will be in one of the parts of S above, ie S is a countable dense subset of H, ie H is separable.

See [[Hilbert spaces with orthonormal bases are separable]]

Theorem (Parseval's Identity)

Let H be a orthonormal basis of H. Then for all uH,

\sum_{n} \lvert \langle u, e_{n} \rangle \rvert { #2} = \lvert \lvert u \rvert \rvert { #2}

(in [[Bessel's inequality]], we only had )

[!proof]
We know that

u=nu,enenfrom , we have \begin{align} \lvert \lvert u \rvert \rvert { #2} &= \lim_{ m \to \infty } \left\langle \sum_{n=1}^m \langle u, e_{n} \rangle e_{n}, \sum_{\ell=1}^m \langle u, e_{\ell} \rangle e_{\ell} \right\rangle \\ &= \lim_{ m \to \infty } \sum_{n, \ell = 1}^m\langle u, e_{n} \rangle \overline{\langle u, e_{\ell} \rangle } \langle e_{n, e_{\ell}} \rangle \\ (e_{n} \perp e_{\ell}, n \neq \ell) \implies&= \lim_{ m \to \infty } \sum_{n=1} \langle u, e_{n} \rangle \overline{\langle u, e_{n} \rangle } \\ &= \lim_{ m \to \infty } \sum_{n=1}^m \lvert \langle u, e_{n} \rangle \rvert { #2}

\end{align}$$

see [[Parseval's identity]]

We now have a way to identify every [[separable]] [[Hilbert space]] with the one that was introduced at the beginning of the course.

Theorem

If H is an infinite-dimensional isometrically isomorphic to 2.

ie, there exists a bijective bounded linear operator T:H2 such that for all u,vH we have

||Tu||2=||u||H and Tu,Tv2=u,vH

The finite case is easier to deal with- we can just show that they are isometrically isomorphic to Cn for some n

Proof (sketch)

Since H is Fourier-Bessel series for each element uH:

u=n=1u,enen||u||=(n=1|u,en|2)1/2

By [[Parseval's identity]]. So we can define our map T as

Tu:={u,en}n

ie, Tu is the sequence of coefficients in the expansion. And this sequence is in 2 (ell-2).

see [[separable Hilbert spaces are bijective with ell-2]]

Fourier Series

Proposition

The subset of functions {einx2π}nZ is an orthonormal subset of L2([π,π])

Note

if we don't like working with complex exponentials we can use

eix=cosx+isinx

And work out everything we need.

Proof

Note that

einx,eimx=ππeinxeimxdx=ππei(nm)xdx={2π,m=n1i(nm)ei(nm)x|ππ=0mn

Since the exponential is periodic in 2π. Normalizing by 2π yields

einx2π,eimx2π={1m=n0mn

Giving us our orthonormal set.

see the [[fourier functions form an orthonormal set]]

Fourier coefficient

For a function fL2([π,π]), the Fourier coefficient f^(n) of f is given by

f^(n)=12πππf(t)eintdt

And the Nth partial Fourier sum is

SNf(x)=|n|Nf^(n)einx=|n|Nf,einx2πeinx2π
Note

this follows from [[all elements of hilbert spaces with orthonormal bases can be written as sums of the basis elements]] and [[Hilbert spaces are bijective with ell-2]]

see [[Fourier coefficient]]

Fourier series

The Fourier series of f is the formal series

nZf^(n)einx

see [[Fourier series]]

Question

Does convergence (in L2) n=1f^einxf hold for all fL2([π,π])? ie, does

\lvert \lvert f - S_{N} f\rvert \rvert _{2} = \left( \int _{-\pi}^\pi \lvert f(x) - S_{N}f(x) \rvert { #2} \, dx \right)^{1/2} \to {0} \, \text{ as } N \to \infty
Answer

It turns out, yes, but we will need to build some more framework to show this.

Theorem

For all fL2([π,π]) and all NN{0},

SNf(x)=ππDN(xt)f(t)dxDN(x)={2N+12πx=0sin(N+12)x2πsinx2x0

The function DN is continuous and also smooth and is called the Dirichlet kernel.

Proof (warmup for some calculations)

For any fL2([π,π]), we know that

SNf(x)=|n|N(12πππf(t)eintdt)einx()=ππf(t)(12π|n|Nein(xt))DN(xt)dt

Where () is by linearity of the integral

DN(x)=12π|n|Neinx=12πeiNxn=02Neinx={12πeiNx[1ei(2N1)x1eix],eix1x02N+12π,x=0={12π[ei(N+1/2)xei(N+1/2)xeix/2eix/2],x02N+12π,x=0(sinx=eixeix2i)={sin(N+12)x2πsinx2,x02N+12π,x=0

Which gives us the desired result

See [[fourier partial sums are given by the Dirichlet kernel]]

Cesaro-Fourier mean

Let fL2([π,π]). The Nth Cesaro-Fourier mean of f is

σNf(x)=1N+1k=0NSkf(x)

Where Skf(x) is the kth partial Fourier sum.

see [[Cesaro-Fourier mean]]

Note

Note

We know from real analysis that the Cesaro means of a sequence behave better but do not lose any information.

Example

Sequences like {1,1,1,1,} do not converge, but their Cersaro means do

Next time, we will see more why this convergence works.

const { dateTime } = await cJS()

return function View() {
	const file = dc.useCurrentFile();
	return <p class="dv-modified">Created {dateTime.getCreated(file)}     ֍     Last Modified {dateTime.getLastMod(file)}</p>
}