文章目录
它只是为了凑个标题
3 Hilbert Spaces
希尔伯特空间是有限维欧拉空间的推广。
3.1Introduction
Definition
Let X X X be a linear space over a field F \mathbb{F} F. An inner product on X X X is a scalar-valued function ⟨ ⋅ , ⋅ ⟩ : X × X → F \langle\cdot, \cdot\rangle: X \times X \rightarrow \mathbb{F} ⟨⋅,⋅⟩:X×X→F such that for all x , y , z ∈ X x, y, z \in X x,y,z∈X and for all α , β ∈ F , \alpha, \beta \in \mathbb{F}, α,β∈F, we have
IP1. ⟨ x , x ⟩ ≥ 0 \langle x, x\rangle \geq 0 ⟨x,x⟩≥0
IP2. ⟨ x , x ⟩ = 0 ⟺ x = 0 \langle x, x\rangle=0 \quad \Longleftrightarrow \quad x=0 ⟨x,x⟩=0⟺x=0
IP3. ⟨ x , y ⟩ = ⟨ y , x ⟩ ‾ \langle x, y\rangle=\overline{\langle y, x\rangle} \quad ⟨x,y⟩=⟨y,x⟩ (The bar denotes complex conjugation.);
IP4. ⟨ α x , y ⟩ = α ⟨ x , y ⟩ \langle\alpha x, y\rangle=\alpha\langle x, y\rangle ⟨αx,y⟩=α⟨x,y⟩
IP5. ⟨ x + y , z ⟩ = ⟨ x , z ⟩ + ⟨ y , z ⟩ \langle x+y, z\rangle=\langle x, z\rangle+\langle y, z\rangle ⟨x+y,z⟩=⟨x,z⟩+⟨y,z⟩
An inner product space ( X , ⟨ ⋅ , ⋅ ⟩ ) (X,\langle\cdot, \cdot\rangle) (X,⟨⋅,⋅⟩) is a linear space X X X together with an inner ⟨ ⋅ , ⋅ ⟩ \langle\cdot, \cdot\rangle ⟨⋅,⋅⟩ product defined on it. A n A n An inner product space is also called pre-Hilbert space.
Example
[1] Fix a positive integer
n
.
n .
n. Let
X
=
F
n
.
X=\mathbb{F}^{n} .
X=Fn. For
x
=
(
x
1
,
x
2
,
…
,
x
n
)
x=\left(x_{1}, x_{2}, \ldots, x_{n}\right)
x=(x1,x2,…,xn) and
y
=
(
y
1
,
y
2
,
…
,
y
n
)
y=\left(y_{1}, y_{2}, \ldots, y_{n}\right)
y=(y1,y2,…,yn) in
X
X
X, define
⟨
x
,
y
⟩
=
∑
i
=
1
n
x
i
y
i
‾
\langle x, y\rangle=\sum_{i=1}^{n} x_{i} \overline{y_{i}}
⟨x,y⟩=i=1∑nxiyi
The space
R
n
\mathbb{R}^{n}
Rn (resp.
C
n
\mathbb{C}^{n}
Cn ) with this inner product is called the Euclidean
n
n
n -space (resp. unitary
n
n
n -space) and will be denoted by
ℓ
2
(
n
)
\ell_{2}(n)
ℓ2(n).
Theorem
[1] (Cauchy-Bunyakowsky-Schwarz Inequality). Let
(
X
,
⟨
⋅
,
⋅
⟩
)
(X,\langle\cdot, \cdot\rangle)
(X,⟨⋅,⋅⟩) be an inner product space over a field
F
\mathbb{F}
F. Then for all
x
,
y
∈
X
x, y \in X
x,y∈X,
∣
⟨
x
,
y
⟩
∣
≤
⟨
x
,
x
⟩
⟨
y
,
y
⟩
|\langle x, y\rangle| \leq \sqrt{\langle x, x\rangle} \sqrt{\langle y, y\rangle}
∣⟨x,y⟩∣≤⟨x,x⟩⟨y,y⟩
Moreover, given any
x
,
y
∈
X
,
x, y \in X,
x,y∈X, the equality
∣
⟨
x
,
y
⟩
∣
=
⟨
x
,
x
⟩
⟨
y
,
y
⟩
|\langle x, y\rangle|=\sqrt{\langle x, x\rangle} \sqrt{\langle y, y\rangle}
∣⟨x,y⟩∣=⟨x,x⟩⟨y,y⟩
holds if and only if
x
x
x and
y
y
y are linearly dependent.
[2] Let
(
X
,
⟨
⋅
,
⋅
⟩
)
(X,\langle\cdot, \cdot\rangle)
(X,⟨⋅,⋅⟩) be an inner product space over a field
F
\mathbb{F}
F. For each
x
∈
X
,
x \in X,
x∈X, define
∥
x
∥
:
=
⟨
x
,
x
⟩
\|x\|:=\sqrt{\langle x, x\rangle}
∥x∥:=⟨x,x⟩
Then
∥
⋅
∥
\|\cdot\|
∥⋅∥ defines a norm on
X
.
X .
X. That is,
(
X
,
∥
⋅
∥
)
(X,\|\cdot\|)
(X,∥⋅∥) is a normed linear space over
F
\mathbb{F}
F.
任何内积空间都可以通过上面的方式定义一个范数来形成赋范空间,该范数又称为内积范数。
以下例子中的范数都是内积范数inner product norm
[3] Theorem (Polarization Identity). Let
(
X
,
⟨
⋅
,
⋅
⟩
)
(X,\langle\cdot, \cdot\rangle)
(X,⟨⋅,⋅⟩) be an inner product space over a field
F
\mathbb{F}
F. Then for all
x
,
y
∈
X
x, y \in X
x,y∈X,
⟨
x
,
y
⟩
=
∥
x
+
y
∥
2
4
−
∥
x
−
y
∥
2
4
if
F
=
R
,
and
\langle x, y\rangle=\frac{\|x+y\|^{2}}{4}-\frac{\|x-y\|^{2}}{4} \text { if } \mathbb{F}=\mathbb{R}, \quad \text { and }
⟨x,y⟩=4∥x+y∥2−4∥x−y∥2 if F=R, and
⟨ x , y ⟩ = ∥ x + y ∥ 2 4 − ∥ x − y ∥ 2 4 + i ( ∥ x + y i ∥ 2 4 − ∥ x − y i ∥ 2 4 ) if F = C \langle x, y\rangle=\frac{\|x+y\|^{2}}{4}-\frac{\|x-y\|^{2}}{4}+i\left(\frac{\|x+y i\|^{2}}{4}-\frac{\|x-y i\|^{2}}{4}\right) \text { if } \mathbb{F}=\mathbb{C} ⟨x,y⟩=4∥x+y∥2−4∥x−y∥2+i(4∥x+yi∥2−4∥x−yi∥2) if F=C
[4] Theorem (Parallelogram Identity). Let
(
X
,
⟨
⋅
,
⋅
⟩
)
(X,\langle\cdot, \cdot\rangle)
(X,⟨⋅,⋅⟩) be an inner product space over a field
F
\mathbb{F}
F. Then for all
x
,
y
∈
X
x, y \in X
x,y∈X,
∥
x
−
y
∥
2
+
∥
x
+
y
∥
2
=
2
∥
x
∥
2
+
2
∥
y
∥
2
\|x-y\|^{2}+\|x+y\|^{2}=2\|x\|^{2}+2\|y\|^{2}
∥x−y∥2+∥x+y∥2=2∥x∥2+2∥y∥2
平行四边形恒等式表示,对角线平方的和等于四条边平方的和。
[5] A normed linear space
X
X
X over a field
F
\mathbb{F}
F is an inner product space if and only if the Parallelogram Identity
∥
x
−
y
∥
2
+
∥
x
+
y
∥
2
=
2
∥
x
∥
2
+
2
∥
y
∥
2
\|x-y\|^{2}+\|x+y\|^{2}=2\|x\|^{2}+2\|y\|^{2}
∥x−y∥2+∥x+y∥2=2∥x∥2+2∥y∥2
holds for all
x
,
y
∈
X
x, y \in X
x,y∈X.
对于一个赋范线性空间,如果其中的所有元素都满足平行四边形恒等式,那么该赋范线性空间就是一个内积空间。
Corollary
Let ( X , ∥ ⋅ ∥ ) (X,\|\cdot\|) (X,∥⋅∥) be a normed linear space over a field F \mathbb{F} F. If every two-dimensional linear subspace of X X X is an inner product space over F , \mathbb{F}, F, then X X X is an inner product space.
范数函数对应关系是 X → F X \rightarrow \mathbb F X→F, 是X中的一个元素对应一个复数
内积函数对应关系是 X × X → F X\times X \rightarrow \mathbb F X×X→F,是X中的一个状态对对应一个复数
3.2 Hilbert Spaces Definition
Let ( X , ⟨ ⋅ , ⋅ ⟩ ) (X,\langle\cdot, \cdot\rangle) (X,⟨⋅,⋅⟩) be an inner product space. If X X X is complete with respect to the norm induced by the inner product ⟨ ⋅ , ⋅ ⟩ , \langle\cdot, \cdot\rangle, ⟨⋅,⋅⟩, then we say that X X X is a Hilbert space.
Hilbert space = Banach space : an complete inner product space with inner product norm
3.3 Orthogonality
Definition
Two elements
x
x
x and
y
y
y in an inner product space
(
X
,
⟨
⋅
,
⋅
⟩
)
(X,\langle\cdot, \cdot\rangle)
(X,⟨⋅,⋅⟩) are said to be orthogonal, denoted by
x
⊥
y
x \perp y
x⊥y, if
⟨
x
,
y
⟩
=
0
\langle x, y\rangle=0
⟨x,y⟩=0
The set
M
⊂
X
M \subset X
M⊂X is called orthogonal if it consists of non-zero pairwise orthogonal elements. If
M
M
M is a subset of
X
X
X such that
⟨
x
,
m
⟩
=
0
\langle x, m\rangle=0
⟨x,m⟩=0 for all
m
∈
M
,
m \in M,
m∈M, then we say that
x
x
x is orthogonal to
M
M
M and write
x
⊥
M
.
x \perp M .
x⊥M. We shall denote by
M
⊥
=
{
x
∈
X
:
⟨
x
,
m
⟩
=
0
∀
m
∈
M
}
M^{\perp}=\{x \in X:\langle x, m\rangle=0 \forall m \in M\}
M⊥={x∈X:⟨x,m⟩=0∀m∈M}
the set of all elements in
X
X
X that are orthogonal to
M
.
M .
M. The set
M
⊥
M^{\perp}
M⊥ is called the orthogonal complement of
M
M
M
M的正交补:X中所有与M正交的元素的集合。
Proposition
Let
M
M
M and
N
N
N be subsets of an inner product space
(
X
,
⟨
⋅
,
⋅
⟩
)
(X,\langle\cdot, \cdot\rangle)
(X,⟨⋅,⋅⟩). Then
[1]
{
0
}
⊥
=
X
\{0\}^{\perp}=X
{0}⊥=X and
X
⊥
=
{
0
}
X^{\perp}=\{0\}
X⊥={0}
[2]
M
⊥
M^{\perp}
M⊥ is a closed linear subspace of
X
X
X;
[
3
]
M
⊂
(
M
⊥
)
⊥
=
M
⊥
⊥
[3] M \subset\left(M^{\perp}\right)^{\perp}=M^{\perp \perp}
[3]M⊂(M⊥)⊥=M⊥⊥
[4] If
M
M
M is a linear subspace, then
M
∩
M
⊥
=
{
0
}
M \cap M^{\perp}=\{0\}
M∩M⊥={0};
[5] If
M
⊂
N
,
M \subset N,
M⊂N, then
N
⊥
⊂
M
⊥
N^{\perp} \subset M^{\perp}
N⊥⊂M⊥
[6]
M
⊥
=
(
lin
M
)
⊥
=
(
lin
‾
M
)
⊥
M^{\perp}=(\operatorname{lin} M)^{\perp}=(\overline{\operatorname{lin}} M)^{\perp}
M⊥=(linM)⊥=(linM)⊥
Theorem
[1] (Pythagoras). Let
(
X
,
⟨
⋅
,
⋅
⟩
)
(X,\langle\cdot, \cdot\rangle)
(X,⟨⋅,⋅⟩) be an inner product space over a field
F
\mathbb{F}
F and let
x
,
y
∈
X
x, y \in X
x,y∈X
[a] If
F
=
R
,
\mathbb{F}=\mathbb{R},
F=R, then
x
⊥
y
x \perp y
x⊥y if and only if
∥
x
+
y
∥
2
=
∥
x
∥
2
+
∥
y
∥
2
\|x+y\|^{2}=\|x\|^{2}+\|y\|^{2}
∥x+y∥2=∥x∥2+∥y∥2
[b] If
F
=
C
,
\mathbb{F}=\mathbb{C},
F=C, then
x
⊥
y
x \perp y
x⊥y if and only if
∥
x
+
y
∥
2
=
∥
x
∥
2
+
∥
y
∥
2
and
∥
x
+
i
y
∥
2
=
∥
x
∥
2
+
∥
y
∥
2
.
\|x+y\|^{2}=\|x\|^{2}+\|y\|^{2} \text { and }\|x+i y\|^{2}=\|x\|^{2}+\|y\|^{2} .
∥x+y∥2=∥x∥2+∥y∥2 and ∥x+iy∥2=∥x∥2+∥y∥2.
Corollary
If
M
=
{
x
1
,
x
2
,
…
,
x
n
}
M=\left\{x_{1}, x_{2}, \ldots, x_{n}\right\}
M={x1,x2,…,xn} is an orthogonal set in an inner product space
(
X
,
⟨
⋅
,
⋅
⟩
)
(X,\langle\cdot, \cdot\rangle)
(X,⟨⋅,⋅⟩) then
∥
∑
i
=
1
n
x
i
∥
2
=
∑
i
=
1
n
∥
x
i
∥
2
\left\|\sum_{i=1}^{n} x_{i}\right\|^{2}=\sum_{i=1}^{n}\left\|x_{i}\right\|^{2}
∥∥∥∥∥i=1∑nxi∥∥∥∥∥2=i=1∑n∥xi∥2
3.4 Best Approximation in Hilbert Spaces
Definition
[1] Let
K
K
K be a closed subset of an inner product space
(
X
,
⟨
⋅
,
⋅
⟩
)
.
(X,\langle\cdot, \cdot\rangle) .
(X,⟨⋅,⋅⟩). For a given
x
∈
X
\
K
,
x \in X \backslash K,
x∈X\K, a best approximation or nearest point to
x
x
x from
K
K
K is any element
y
0
∈
K
y_{0} \in K
y0∈K such that
∥
x
−
y
0
∥
≤
∥
x
−
y
∥
for all
y
∈
K
\left\|x-y_{0}\right\| \leq\|x-y\| \quad \text { for all } \quad y \in K
∥x−y0∥≤∥x−y∥ for all y∈K
Equivalently,
y
0
∈
K
y_{0} \in K
y0∈K is a best approximation to
x
x
x from
K
K
K if
∥
x
−
y
0
∥
=
inf
y
∈
K
∥
x
−
y
∥
=
d
(
x
,
K
)
\left\|x-y_{0}\right\|=\inf _{y \in K}\|x-y\|=d(x, K)
∥x−y0∥=y∈Kinf∥x−y∥=d(x,K)
nearest point:顾名思义,是K中距离K的补集中某点x的最近的一些点。最佳近似点来自于K集。
[2] The (possibly empty) set of all best approximations to
x
x
x from
K
K
K is denoted by
P
K
(
x
)
.
P_{K}(x) .
PK(x). That is,
P
K
(
x
)
=
{
y
∈
K
:
∥
x
−
y
∥
=
d
(
x
,
K
)
}
P_{K}(x)=\{y \in K:\|x-y\|=d(x, K)\}
PK(x)={y∈K:∥x−y∥=d(x,K)}
The (generally set-valued) map
P
K
P_{K}
PK which associates each
x
x
x in
X
X
X with its best approximations in
K
K
K is called the metric projection or the nearest point map. The set
K
K
K is called
P K P_K PK: K中离x最近的点的映射。
[a] proximinal if each x ∈ X x \in X x∈X has a best approximation in K K K; i.e., P K ( x ) ≠ ∅ P_{K}(x) \neq \emptyset PK(x)=∅ for each x ∈ X x \in X x∈X;
存在最近点,什么情况下会不存在呢?K是开集的情况下?
[b] Chebyshev if each x ∈ X x \in X x∈X has a unique best approximation in K K K; i.e., the set P K ( x ) P_{K}(x) PK(x) consists of a single point.
只存在一个最近点。
The following important result asserts that if K K K is a complete convex subset of an inner product space ( X , ⟨ ⋅ , ⋅ ⟩ ) , (X,\langle\cdot, \cdot\rangle), (X,⟨⋅,⋅⟩), then each x ∈ X x \in X x∈X has one and only one element of best approximation in K K K.
Theorem
[1] Every nonempty complete convex subset K K K of an inner product space ( X , ⟨ ⋅ , ⋅ ⟩ ) (X,\langle\cdot, \cdot\rangle) (X,⟨⋅,⋅⟩) is a Chebyshev set.
[2] Let
K
K
K be a nonempty closed convex subset of a Hilbert space
(
H
,
⟨
⋅
,
⋅
⟩
)
,
x
∈
H
\
K
(\mathcal{H},\langle\cdot, \cdot\rangle), x \in \mathcal{H} \backslash K
(H,⟨⋅,⋅⟩),x∈H\K and
y
0
∈
K
.
y_{0} \in K .
y0∈K. Then
y
0
y_{0}
y0 is the best approximation to
x
x
x from
K
K
K if and only if
ℜ
e
⟨
x
−
y
0
,
y
−
y
0
⟩
≤
0
for all
y
∈
K
\Re_{\mathrm{e}}\left\langle x-y_{0}, y-y_{0}\right\rangle \leq 0 \text { for all } y \in K
ℜe⟨x−y0,y−y0⟩≤0 for all y∈K
[3] (Projection Theorem). Let
H
\mathcal{H}
H be a Hilbert space,
M
M
M a closed subspace of
H
.
\mathcal{H} .
H. Then
[
a
]
H
=
M
⊕
M
⊥
.
[a] \mathcal{H}=M \oplus M^{\perp} .
[a]H=M⊕M⊥. That is, each
x
∈
H
x \in \mathcal{H}
x∈H can be uniquely decomposed in the form
x
=
y
+
z
with
y
∈
M
and
z
∈
M
⊥
x=y+z \text { with } y \in M \text { and } z \in M^{\perp}
x=y+z with y∈M and z∈M⊥
[
b
]
M
=
M
⊥
⊥
[b] M=M^{\perp \perp}
[b]M=M⊥⊥
Corollary
[1] Every nonempty closed convex subset of a Hilbert space is Chebyshev.
[2] (Characterization of Best Approximations from closed subspaces). Let M M M be a closed subspace of a Hilbert space H \mathcal{H} H and let x ∈ H \ M x \in \mathcal{H} \backslash M x∈H\M. Then an element y 0 ∈ M y_{0} \in M y0∈M is the best approximation to x x x from M M M if and only if ⟨ x − y 0 , y ⟩ = 0 \left\langle x-y_{0}, y\right\rangle=0 ⟨x−y0,y⟩=0 for all y ∈ M y \in M y∈M (i.e., x − y 0 ∈ M ⊥ ) \left.x-y_{0} \in M^{\perp}\right) x−y0∈M⊥).
so map P M : x → P M ( x ) P_{M}: x \rightarrow P_{M}(x) PM:x→PM(x) is also called the orthogonal projection of H \mathcal{H} H onto M M M
[3] If
M
M
M is a closed subspace of a Hilbert space
H
,
\mathcal{H},
H, and if
M
≠
H
,
M \neq \mathcal{H},
M=H, then there exists
z
∈
H
\
{
0
}
z \in \mathcal{H} \backslash\{0\}
z∈H\{0} such that
z
⊥
M
z \perp M
z⊥M
Proof. Let
x
∈
H
\
M
x \in \mathcal{H} \backslash M
x∈H\M. Then by the Projection Theorem,
x
=
y
+
z
,
where
y
∈
M
and
z
∈
M
⊥
x=y+z, \text { where } y \in M \text { and } z \in M^{\perp}
x=y+z, where y∈M and z∈M⊥
Hence
z
≠
0
z \neq 0
z=0 and
z
⊥
M
z \perp M
z⊥M.
Proposition
Let
S
S
S be a nonempty subset of a Hilbert space
H
\mathcal{H}
H. Then
[
1
]
S
⊥
⊥
=
lin
‾
S
[1] S^{\perp \perp}=\overline{\operatorname{lin}} S
[1]S⊥⊥=linS
[
2
]
S
⊥
=
{
0
}
[2] S^{\perp}=\{0\}
[2]S⊥={0} if and only if
lin
‾
S
=
H
\overline{\operatorname{lin}} S=\mathcal{H}
linS=H
3.5 Orthonormal Sets and Orthonormal Bases
Definition
[1] Let
(
X
,
⟨
⋅
,
⋅
⟩
)
(X,\langle\cdot, \cdot\rangle)
(X,⟨⋅,⋅⟩) be an inner product space over
F
.
A
\mathbb{F} . A
F.A set
S
=
{
x
α
:
α
∈
Λ
}
S=\left\{x_{\alpha}: \alpha \in \Lambda\right\}
S={xα:α∈Λ} of elements of
X
X
X is called an orthonormal set if
(a)
⟨
x
α
,
x
β
⟩
=
0
\left\langle x_{\alpha}, x_{\beta}\right\rangle=0
⟨xα,xβ⟩=0 for all
α
≠
β
\alpha \neq \beta
α=β (i.e.,
S
S
S is an orthogonal set), and
(b)
∥
x
α
∥
=
1
\left\|x_{\alpha}\right\|=1
∥xα∥=1 for all
α
∈
Λ
\alpha \in \Lambda
α∈Λ.
If
S
=
{
x
α
:
α
∈
Λ
}
S=\left\{x_{\alpha}: \alpha \in \Lambda\right\}
S={xα:α∈Λ} is an orthonormal set and
x
∈
X
,
x \in X,
x∈X, then the numbers
⟨
x
,
x
α
⟩
\left\langle x, x_{\alpha}\right\rangle
⟨x,xα⟩ are called the Fourier coefficients of
x
x
x with respect to
S
S
S and the formal series
∑
α
∈
Λ
⟨
x
,
x
α
⟩
x
α
\sum_{\alpha \in \Lambda}\left\langle x, x_{\alpha}\right\rangle x_{\alpha}
∑α∈Λ⟨x,xα⟩xα the Fourier series of
x
x
x.
Λ \Lambda Λ表示S是X的规范正交系
[2] An orthonormal set S S S in an inner product space ( X , ⟨ ⋅ , ⋅ ⟩ ) (X,\langle\cdot, \cdot\rangle) (X,⟨⋅,⋅⟩) is said to be complete in X X X if S ⊂ T S \subset T S⊂T and T T T is an orthonormal set in X , X, X, then S = T S=T S=T.
Simply put, a complete orthonormal set S S S in an inner product space is an orthonormal set that is not properly contained in any other orthonormal set in X ; X ; X; in other words, S S S is complete if it is a maximal orthonormal set in X X X.
完备正交集是X中最大的正交集。
It is easy exercise to show that a set S S S is complete in an inner product ( X , ⟨ ⋅ , ⋅ ⟩ ) (X,\langle\cdot, \cdot\rangle) (X,⟨⋅,⋅⟩) if and only if S ⊥ = { 0 } S^{\perp}=\{0\} S⊥={0}.
[3] Let
(
X
,
⟨
⋅
,
⋅
⟩
)
(X,\langle\cdot, \cdot\rangle)
(X,⟨⋅,⋅⟩) be an inner product space over
F
\mathbb{F}
F. An orthonormal set
{
x
n
}
\left\{x_{n}\right\}
{xn} is called an orthonormal basis for
X
X
X if for each
x
∈
X
x \in X
x∈X
x
=
∑
k
=
1
∞
⟨
x
,
x
k
⟩
x
k
x=\sum_{k=1}^{\infty}\left\langle x, x_{k}\right\rangle x_{k}
x=k=1∑∞⟨x,xk⟩xk
在无限维的空间中定义的。
That is, the sequence of partial sums ( s n ) , \left(s_{n}\right), (sn), where s n = ∑ k = 1 n ⟨ x , x k ⟩ x k , s_{n}=\sum_{k=1}^{n}\left\langle x, x_{k}\right\rangle x_{k}, sn=∑k=1n⟨x,xk⟩xk, converges to x x x.
[4] isomorphic(同构)
Two linear spaces
X
X
X and
Y
Y
Y over the same field
F
\mathbb{F}
F are said to be isomorphic if there is a one-to-one
map
T
\operatorname{map} T
mapT from
X
X
X onto
Y
Y
Y such that for all
x
1
,
x
2
∈
X
x_{1}, x_{2} \in X
x1,x2∈X and all
α
,
β
∈
F
,
\alpha, \beta \in \mathbb{F},
α,β∈F,
T
(
α
x
1
+
β
x
2
)
=
α
T
(
x
1
)
+
β
T
(
x
2
)
T\left(\alpha x_{1}+\beta x_{2}\right)=\alpha T\left(x_{1}\right)+\beta T\left(x_{2}\right)
T(αx1+βx2)=αT(x1)+βT(x2)
the result of that T is applied to x ∈ X x\in X x∈X is y ∈ Y y\in Y y∈Y. 同构的用途是将一些领域的定理能够迁移到另一个同构的领域。
[5] isometry(等距)
Let
(
X
,
∥
⋅
∥
)
(X,\|\cdot\|)
(X,∥⋅∥) and
(
Y
,
∥
⋅
∥
)
(Y,\|\cdot\|)
(Y,∥⋅∥) be two normed linear spaces and
T
:
X
→
Y
T: X \rightarrow Y
T:X→Y. Then
T
T
T is called an isometry if
∥
T
x
∥
=
∥
x
∥
for all
x
∈
X
.
\|T x\|=\|x\| \text { for all } x \in X .
∥Tx∥=∥x∥ for all x∈X.
Simply put, an isometry is a map that preserves lengths.
等距映射,其中左边的范数是Y中的范数,右边的范数是X中的范数。为了简洁,都写成一样的。
Normed linear spaces that are isometrically isomorphic are essentially identical
等距同构的赋范线性空间基本相同。
Theorem
[1] An orthonormal set S S S in a separable inner product space ( X , ⟨ ⋅ , ⋅ ⟩ ) (X,\langle\cdot, \cdot\rangle) (X,⟨⋅,⋅⟩) is at most countable.
可分离的内积空间的正交集合最多。
Let
(
X
,
⟨
⋅
,
⋅
⟩
)
(X,\langle\cdot, \cdot\rangle)
(X,⟨⋅,⋅⟩) be a separable inner product space over
F
\mathbb{F}
F.
[2] (Best Fit). If
{
x
1
,
x
2
,
…
,
x
n
}
\left\{x_{1}, x_{2}, \ldots, x_{n}\right\}
{x1,x2,…,xn} is a finite orthonormal set in
X
X
X and
M
=
lin
{
x
1
,
x
2
,
…
,
x
n
}
,
M=\operatorname{lin}\left\{x_{1}, x_{2}, \ldots, x_{n}\right\},
M=lin{x1,x2,…,xn}, then for each
x
∈
X
x \in X
x∈X there exists
y
0
∈
M
y_{0} \in M
y0∈M such that
$$
\begin{array}{l}
\qquad\left|x-y_{0}\right|=d(x, M) \
\end{array}
$$
In fact,
y
0
=
∑
k
=
1
n
⟨
x
,
x
k
⟩
x
k
\text { In fact, } y_{0}=\sum_{k=1}^{n}\left\langle x, x_{k}\right\rangle x_{k}
In fact, y0=∑k=1n⟨x,xk⟩xk
[3] (Bessel’s Inequality). Let
(
x
n
)
n
=
1
∞
\left(x_{n}\right)_{n=1}^{\infty}
(xn)n=1∞ be an orthonormal sequence in
X
.
X .
X. Then for any
x
∈
X
x \in X
x∈X,
∑
k
=
1
∞
∣
⟨
x
,
x
k
⟩
∣
2
≤
∥
x
∥
2
\sum_{k=1}^{\infty}\left|\left\langle x, x_{k}\right\rangle\right|^{2} \leq\|x\|^{2}
k=1∑∞∣⟨x,xk⟩∣2≤∥x∥2
In particular,
⟨
x
,
x
k
⟩
→
0
\left\langle x, x_{k}\right\rangle \rightarrow 0
⟨x,xk⟩→0 as
k
→
∞
k \rightarrow \infty
k→∞.
[4] (Riesz-Fischer Theorem). Let
(
x
n
)
1
∞
\left(x_{n}\right)_{1}^{\infty}
(xn)1∞ be an orthonormal sequence in a separable Hilbert space
H
\mathcal{H}
H and let
(
c
n
)
1
∞
\left(c_{n}\right)_{1}^{\infty}
(cn)1∞ be a sequence of scalars. Then the series
∑
k
=
1
∞
c
k
x
k
\sum_{k=1}^{\infty} c_{k} x_{k}
∑k=1∞ckxk converges in
H
\mathcal{H}
H if and only if
c
=
(
c
n
)
1
∞
∈
ℓ
2
.
c=\left(c_{n}\right)_{1}^{\infty} \in \ell_{2} .
c=(cn)1∞∈ℓ2. In this case,
∥
∑
k
=
1
∞
c
k
x
k
∥
=
(
∑
k
=
1
∞
∣
c
k
∣
2
)
1
2
\left\|\sum_{k=1}^{\infty} c_{k} x_{k}\right\|=\left(\sum_{k=1}^{\infty}\left|c_{k}\right|^{2}\right)^{\frac{1}{2}}
∥∥∥∥∥k=1∑∞ckxk∥∥∥∥∥=(k=1∑∞∣ck∣2)21
[5] Let
H
\mathcal{H}
H be a separable infinite-dimensional Hilbert space and assume that
S
=
{
x
n
}
S=\left\{x_{n}\right\}
S={xn} is an orthonormal set in
H
\mathcal{H}
H. Then the following statements are equivalent:
[a]
S
S
S is complete in
H
;
\mathcal{H} ;
H; i.e.,
S
⊥
=
{
0
}
S^{\perp}=\{0\}
S⊥={0}.
[b]
lin
‾
S
=
H
;
\overline{\operatorname{lin}} S=\mathcal{H} ;
linS=H; i.e., the linear span of
S
S
S is norm-dense in
H
\mathcal{H}
H.
[c] (Fourier Series Expansion.) For any
x
∈
H
x \in \mathcal{H}
x∈H, we have
x
=
∑
i
=
1
∞
⟨
x
,
x
i
⟩
x
i
x=\sum_{i=1}^{\infty}\left\langle x, x_{i}\right\rangle x_{i}
x=i=1∑∞⟨x,xi⟩xi
That is,
S
S
S is an orthonormal basis for
H
\mathcal{H}
H.
[d] (Parseval’s Identity.) For all
x
,
y
∈
H
x, y \in \mathcal{H}
x,y∈H,
⟨
x
,
y
⟩
=
∑
k
=
1
∞
⟨
x
,
x
k
⟩
⟨
y
,
x
k
⟩
‾
\langle x, y\rangle=\sum_{k=1}^{\infty}\left\langle x, x_{k}\right\rangle \overline{\left\langle y, x_{k}\right\rangle}
⟨x,y⟩=k=1∑∞⟨x,xk⟩⟨y,xk⟩
[e] For any
x
∈
H
x \in \mathcal{H}
x∈H,
∥
x
∥
2
=
∑
k
=
1
∞
∣
⟨
x
,
x
k
⟩
∣
2
\|x\|^{2}=\sum_{k=1}^{\infty}\left|\left\langle x, x_{k}\right\rangle\right|^{2}
∥x∥2=k=1∑∞∣⟨x,xk⟩∣2
[6] (Gram-Schmidt Orthonormalisation Procedure). If
{
x
k
}
1
∞
\left\{x_{k}\right\}_{1}^{\infty}
{xk}1∞ is a linearly independent set in an inner product space
(
X
,
⟨
⋅
,
⋅
⟩
)
(X,\langle\cdot, \cdot\rangle)
(X,⟨⋅,⋅⟩) then there exists an orthonormal set
{
e
k
}
1
∞
\left\{e_{k}\right\}_{1}^{\infty}
{ek}1∞ in
X
X
X such that
lin
{
x
1
,
x
2
,
…
,
x
n
}
=
lin
{
e
1
,
e
2
,
…
,
e
n
}
for all
\operatorname{lin}\left\{x_{1}, x_{2}, \ldots, x_{n}\right\}=\operatorname{lin}\left\{e_{1}, e_{2}, \ldots, e_{n}\right\} \quad \text { for all }
lin{x1,x2,…,xn}=lin{e1,e2,…,en} for all
[7] Every separable Hilbert space
H
\mathcal{H}
H has a countable orthonormal basis.
[8] Every separable infinite-dimensional Hilbert space H \mathcal{H} H is isometrically isomorphic to ℓ 2 \ell_{2} ℓ2.
每个可分的无限维希尔伯特空间和赋范空间 ℓ 2 \ell_{2} ℓ2都是等距同构的。
Lemma
Let
M
=
lin
{
x
n
}
M=\operatorname{lin}\left\{x_{n}\right\}
M=lin{xn} be a linear subspace of
X
.
X .
X. Then there exists a subsequence
{
x
n
k
}
\left\{x_{n_{k}}\right\}
{xnk} of
{
x
n
}
\left\{x_{n}\right\}
{xn} which has the following properties:
(i)
lin
{
x
n
k
}
=
M
\operatorname{lin}\left\{x_{n_{k}}\right\}=M
lin{xnk}=M
(ii)
{
x
n
k
}
\left\{x_{n_{k}}\right\}
{xnk} is linearly independent.