표준 벡터 공간의 측정
기능 분석 에서 이중 표준 은 정규 벡터 공간 에 정의된 연속 선형 함수 에 대한 크기 측도다.
정의 X {\displaystyle X} 을(를) 표준 벡터 공간 으로 하고, X {\ displaystyle \cdot \ \cdot \} 을(를) 이중 공간 으로 한다.X ∗ {\ displaystyle X^{*} 에 속하는 연속 선형 함수 f {\displaystyle f} 의 이중 표준 은 다음 등가 공식 중 하나로 정의된 비음수[1] 실수다.
‖ f ‖ = up { f ( x ) : ‖ x ‖ ≤ 1 그리고 x ∈ X } = up { f ( x ) : ‖ x ‖ < 1 그리고 x ∈ X } = 바 조로 { c ∈ [ 0 , ∞ ) : f ( x ) ≤ c ‖ x ‖ 대체적으로 x ∈ X } = up { f ( x ) : ‖ x ‖ = 1 또는 0 그리고 x ∈ X } = up { f ( x ) : ‖ x ‖ = 1 그리고 x ∈ X } 이 평등은 만약의 경우에 한해서만 유지된다. X ≠ { 0 } = up { f ( x ) ‖ x ‖ : x ≠ 0 그리고 x ∈ X } 이 평등은 만약의 경우에 한해서만 유지된다. X ≠ { 0 } {\displaystyle{\begin{alignedat}{5}\ f\&=\sup&&\{f())&&일:~\x\ \leq 1~&,&~{\text{과}}~&,&x\in X\}\\&,=\sup&&\{f())&&일:~\ x\<>1~&,&~{\text{과}}~&,&x\in X\}\\&, =\inf&&\{모든{에 c\in는 경우에는 0,\infty)&&일:~ f())\leqc\ x\ ~&,&~{\text}}~&,&.X\in X\}\\&,=\sup&&\{f())&&일:~\x\ =1{\text{또는}}0~&,&~{\text{과}}~&,&x\in X\}\\&,=\sup&&){f())&&일:~\ x\. =1~&&~{\text{ and }}~&&x\in X\}\;\;\;{\text{ this equality holds if and only if }}X\neq \{0\}\\&=\sup &&{\bigg \{}{\frac { f(x) }{\ x\ }}~&&~:~x\neq 0&&~{\text{ and }}~&&x\in X{\bigg \}}\;\;\;{\text{ this equality holds if and only if }}X\neq \{0\}\\\end{alignedat}}}
여기서 supp {\displaystyle \sup } 과 inf {\displaystyle \inf } 은 각각 우월감과 최소치 를 나타낸다 . 상수 0 지도는 항상 표준이 0 이고 그것은 벡터 공간 X ∗ . {\displaystyle X^{*}}} 만약 X = { 0 } {\displaystyle X=\{ 0\}}}}}} 의 유일한 선형 기능 은 상수 0 지도일 뿐이며, 더욱이 마지막 두 행의 집합은 모두 비어 있고 결과적으로 그 수가 제곱 일 것이다. 리무진 은 정확한 값 인 0 대신 ∞ 과 같을 것이다.
지도 f ↦ ‖ f ‖ {\displaystyle f\mapsto \f\}은( 는) X ∗. {\displaystyle X^{*}}} 에 규범을 정의한다 ( 아래 정리 1과 2 참조).
이중 표준은 규범 벡터 공간 사이의 각 (경계) 선형 지도에 대해 정의된 연산자 표준의 특별한 경우다.
‖{\ displaystyle X^{*}} 가 유도 한 X ∗{\ displaystyle \cdot \}} 의 위상은 X ∗ . {\displaysty X^{*} 의 약한-* 위상 만큼 강한 것으로 나타났다 .}
X {\displaystyle X} 의 접지 필드 가 완료 되면 X ∗ {\ displaystyle X^{*}} 는 Banach 공간 이다 .
표준 선형 공간의 이중 이중화 The double dual (or second dual) X ∗ ∗ {\displaystyle X^{**}} of X {\displaystyle X} is the dual of the normed vector space X ∗ {\displaystyle X^{*}} . There is a natural map φ : X → X ∗ ∗ {\displaystyle \varphi :X\to X^{**}} . Indeed, for each w ∗ {\displaystyle w^{*}} in X ∗ {\displa ystyle X^{*}} 정의
φ ( v ) ( w ∗ ) := w ∗ ( v ) . {\displaystyle \varphi (v)(w^{*}):=w^{*}(v). } 지도 φ {\displaystyle \varphi } 은(는) 선형 , 주입형 , 거리 보존형 이다 .[2] 특히 X {\displaystyle X} 이(가) 완료되면 (즉, Banach 공간) φ {\displaystyle \varphi } 은(는) X ∗ {\ displaystyle X^{*} 의 닫힌 하위 공간에 대한 등각도법이다. [3]
일반적으로 지도 φ {\displaystyle \varphi } 은 (는) 추월적이지 않다. 예를 들어, X {\ displaystyle X }이 (가) Banach 공간 L ∞{\ displaystyle L^{\inforty } 이(가) 최상규범과 함께 실제 라인의 경계 함수로 구성되어 있다면, 지도 φ {\ displaysty \varphi } 은(가) 을 ( 참조 ). 만일 { {\displaystyle \varphi }이 (가) 허탈적이라면, X {\displaystyle X} 은 반사적인 Banach 공간 이라고 한다 .만약 1< p < ∞ , {\displaystyle 1<\p<\ p<\p>}}) 이면 공간 L p {\ displaystyle L^{p}} 는 반사적인 바나흐 공간이다 .
수학적 최적화 R n . {\displaystyle \cdot \} 을(를) R n . {\displaystyle \mathb {R} ^{n} 에서 표준이 되게 두십시오. ‖ ⋅ ‖ ∗ , , {\displaystyle \cdot \ \ \ _{*} 로 표시된 관련 이중 표준 은 다음과 같이 정의된다 .
‖ z ‖ ∗ = up { z ⊺ x ∣ ‖ x ‖ ≤ 1 } . {\displaystyle \\\{*}=\sup\{z^{\intercal }x\mid \x\\\leq 1\}} (이것은 표준으로 보일 수 있다.) The dual norm can be interpreted as the operator norm of z ⊺ {\displaystyle z^{\intercal }} , interpreted as a 1 × n {\displaystyle 1\times n} matrix, with the norm ‖ ⋅ ‖ {\displaystyle \ \cdot \ } on R n {\displaystyle \mathbb {R} ^{n}} , and the absolute value on R {\displaystyle \mathbb {R} } :
‖ z ‖ ∗ = up { z ⊺ x ∣ ‖ x ‖ ≤ 1 } . \\\{*}=\sup\{z^{\intercal }x \mid \x\\\leq 1\}} 이중규범의 정의로부터 우리는 불평등을 가진다.
z ⊺ x = ‖ x ‖ ( z ⊺ x ‖ x ‖ ) ≤ ‖ x ‖ ‖ z ‖ ∗ {\displaystyle z^{\intercal }x=\x\\x\\x\\intercal }{\frac {x}{\x\}\leq \x\\\\ _{*}}} 모든 x 와 z를 지탱하는.[4] 이중 표준의 이중은 원래 표준이다: 우리는 모든 x에 대해 for x = = = x x \\displaystyle \ x\ _{**}=\ x\ } 을(를) 가지고 있다. (이는 무한 차원 벡터 공간에서 유지할 필요가 없다.)
유클리드 규범 의 이중성은 유클리드 규범인데, 그 이후부터이다.
up { z ⊺ x ∣ ‖ x ‖ 2 ≤ 1 } = ‖ z ‖ 2 . \displaystyle \sup\{z^{\intercal }x\mid \ x\ _{2}\leq 1\}=\z\ _{2}. } (This follows from the Cauchy–Schwarz inequality ; for nonzero z , the value of x that maximises z ⊺ x {\displaystyle z^{\intercal }x} over ‖ x ‖ 2 ≤ 1 {\displaystyle \ x\ _{2}\leq 1} is z ‖ z ‖ 2 {\displaystyle {\tfrac {z}{\ z\ _{2}}}} .)
ℓ{\ displaystyle \ell _{\infit }- norm의 이중은 ℓ 1 {\ displaystyle \ell _{1} -norm :
up { z ⊺ x ∣ ‖ x ‖ ∞ ≤ 1 } = ∑ i = 1 n z i = ‖ z ‖ 1 , {\displaystyle \sup\{z^{}\intercal }x\mid \x\x\mid \_{\\\\\n}\sum _{i=1}^{n}z} =\\{1},} and 1 {\ displaystyle \ell _{1}- norm의 이중은 ℓ ∞ {\ displaystyle \ell _{\inflt }}- norm이다 .
More generally, Hölder's inequality shows that the dual of the ℓ p {\displaystyle \ell _{p}} -norm is the ℓ q {\displaystyle \ell _{q}} -norm, where, q satisfies 1 p + 1 q = 1 {\displaystyle {\tfrac {1}{p}}+{\tfrac {1}{q}}=1} , i.e., q = p p − 1 . {\displaystyle q={\tfrac {p}{p-1}}. }
또 다른 예로, R m × n {\ displaystyle \ell _{2 }} - 또는 스펙트럼 노먼(R mathb {R} ^{m\time n}) 을 들 수 있다. 관련 이중 표준은
‖ Z ‖ 2 ∗ = up { t r ( Z ⊺ X ) ∣ ‖ X ‖ 2 ≤ 1 } , {\displaystyle \ Z\ _{2*}=\supp\{\mathbf {tr}(Z^{\intercal }X)\mid \ X\ _{2}\leq 1\}}} 단수 값의 합으로 밝혀지고
‖ Z ‖ 2 ∗ = σ 1 ( Z ) + ⋯ + σ r ( Z ) = t r ( Z ⊺ Z ) , {\displaystyle \ Z\ _{2*}=\sigma _{1}(Z)+\cdots +\sigma _{r}=\mathbf {tr}({Z^{\intercal })}),} 여기 서 r = r n k Z . {\displaystyle r=\mathbf {rank} Z. }} 이 규범을 핵 규범이 라고도 한다.[5]
예 행렬에 대한 이중 표준 에 의해 정의 된 프로베니우스 규범
‖ A ‖ F = ∑ i = 1 m ∑ j = 1 n a i j 2 = 자취를 감추다 ( A ∗ A ) = ∑ i = 1 분 { m , n } σ i 2 \\displaystyle \ A\ _{\text{ F}}={\sqrt {\sum _{i=1}^{m}\sum _{j=1}^{n}\left a_{ij}\right ^{2}}}={\sqrt {\operatorname {trace} (A^{*}A)}}={\sqrt {\sum _{i=1}^{\min\{m,n\}}\sigma _{i}^{2}}}} 자체 이중화, 즉 이중규격 은 ‖ ⋅ F = = ‖ F . {\displaystyle \cdot \'_{\text {\text{ F}}=\ \cdot \ _{\text{ F}}.}
p = 2 {\displaystyle p=2} 일 때 유도된 규범 의 특별한 경우인 스펙트럼 노먼 은 행렬의 최대 단수 값 으로 정의된다.
‖ A ‖ 2 = σ 맥스. ( A ) , {\displaystyle \ A\ _{2}=\sigma _{\max }(A),} 핵 규범을 그 이중 규범(dual norm)으로 하고 있는데, 이 규범은 다음과 같이 정의된다.
‖ B ‖ 2 ′ = ∑ i σ i ( B ) , {\displaystyle \ B\ '_{2}=\sum _{i}\sigma _{i}(B),} σ i ( B ) {\displaystyle \sigma _{i}(B)} 이(가) 단수 값을[citation needed ] 나타내는 모든 행렬 B {\displaystyle B } 에 대해
운영자 표준에 대한 몇 가지 기본 결과 더 일반적으로, X{X\displaystyle}, Y{Y\displaystyle} 위상 벡터 공간과 L(X, Y) X{X\displaystyle}의 Y.{Y\displaystyle}에 의해 선형 매핑(운전자나)의 경우{L(X,Y)\displaystyle}[6] 컬렉션자 여기서 X{X\displaystyle}와 y { \displaystyle Y} 은 (는) 표준 벡터 공간이며, L( X , Y ) {\displaystyle L(X,Y)} 은(는) 표준 규격을 부여할 수 있다 .
정리 1 — X {\displaystyle X} 및 Y {\displaystyle Y} 을 (를) 정규화된 공간으로 두십시오.각 연속 선형 연산자 f ∈ ( X , Y ) {\displaystyle f\in L(X,Y)} 에 할당
‖ f ‖ = up { ‖ f ( x ) ‖ : x ∈ X , ‖ x ‖ ≤ 1 } {\displaystyle \f\ =\suppleft\{\f(x)\ :x\in X,\\\leq 1\오른쪽\}} defines a norm ‖ ⋅ ‖ : L ( X , Y ) → R {\displaystyle \ \cdot \ ~:~L(X,Y)\to \mathbb {R} } on L ( X , Y ) {\displaystyle L(X,Y)} that makes L ( X , Y ) {\displaystyle L(X,Y)} into a normed space. 더욱이 Y {\displaystyle Y} 이 (가 ) Banach 공간 이라면 L ( X , Y ) . {\displaystyle L(X,Y )도 마찬가지다. } [7]
증명
한normed 공간의 하위 집합 만일 제품 구의 일부 배수로에 있다, 따라서‖ f‖<>마다 f에 ∞{\displaystyle)f\<>\infty}L(X, Y){\displaystyle f\in L(X,Y)}만약α{\displaystyle \alpha}은 scalar, 그때(fα)()))α ⋅ f){\displaystyle(f\alpha)())=\alpha \cdot fx}∈ 경계를 이루고 있다. 그렇게 티 모자를 씌우다
‖ α f ‖ = α ‖ f ‖ \displaystyle \\ \f\ = \cHB \ f\ } Y {\displaystyle Y} 의 삼각형 불평등 은 다음을 나타낸다.
‖ ( f 1 + f 2 ) x ‖ = ‖ f 1 x + f 2 x ‖ ≤ ‖ f 1 x ‖ + ‖ f 2 x ‖ ≤ ( ‖ f 1 ‖ + ‖ f 2 ‖ ) ‖ x ‖ ≤ ‖ f 1 ‖ + ‖ f 2 ‖ {\displaystyle {\signified}\ \left(f_{1}+f_{2}\right)x\ ~&=\ f_{1}x+f_{2} }x\\\&\leq ~\ f_{1}x\ +\ f_{2}x\\\\\\\좌측(\ f_{1}\\\f_{2}\오른쪽)\x\\\\\\\\\\\\\\\\\\\\\\\\\\\\{1}\i1}\\{1}}}}}}}}}}}}}}}}}}}}}}}}}정렬정렬}}}} for every x ∈ X {\displaystyle x\in X} satisfying ‖ x ‖ ≤ 1. {\displaystyle \ x\ \leq 1.} This fact together with the definition of ‖ ⋅ ‖ : L ( X , Y ) → R {\displaystyle \ \cdot \ ~:~L(X,Y)\to \mathbb {R} } implies the triangle inequality:
‖ f 1 + f 2 ‖ ≤ ‖ f 1 ‖ + ‖ f 2 ‖ {\displaystyle \left\f_{1}+f_{2}\right\ \leq \left\f_{1}\{1}\right\ +\\ft\{2}\right\}}} Since { f ( x ) : x ∈ X , ‖ x ‖ ≤ 1 } {\displaystyle \{ f(x) :x\in X,\ x\ \leq 1\}} is a non-empty set of non-negative real numbers, ‖ f ‖ = sup { f ( x ) : x ∈ X , ‖ x ‖ ≤ 1 } {\displaystyle \ f\ =\sup \left\{ f(x) :x\in X,\ x\ \leq 1\right\}} is a non-negative real number. 어떤 x0∈ X를 만약 f≠ 0{\displaystyle f\neq 0} 다음 f)0≠ 0{\displaystyle fx_{0}\neq 0}일 경우(X,}그것은 ‖ f)0‖하는 것을 의미하기;0{\displaystyle\left\ fx_{0}\right\<>를 사용하여 0}일 경우, 결과적으로‖ f‖<>를 사용하여 0입니다.{\displaystyle)f\>0.}.이것은(L(X, Y),‖ ⋅을 보여 준다. ‖ ) {\ displaystyle \left(L(X,Y),\\cdot \right) 은 표준 공간이다.[8]
이제 Y {\displaystyle Y} 이 (가) 완료되었다고 가정하고 (L (X , Y ) , ‖ ⋅ ‖ ) {\ displaystyle \left(L(X,Y),\cdot \right)} 이 (가) 완료되었다고 가정해 보십시오.Let f ∙ = ( f n ) n = 1 ∞ {\displaystyle f_{\bullet }=\left(f_{n}\right)_{n=1}^{\infty }} be a Cauchy sequence in L ( X , Y ) , {\displaystyle L(X,Y),} so by definition ‖ f n − f m ‖ → 0 {\displaystyle \left\ f_{n}-f_{m}\right\ \to 0} as n , m → ∞ . {\displaystyle n,m\to \infty .} This fa 관계와 함께 ct를 찍다.
‖ f n x − f m x ‖ = ‖ ( f n − f m ) x ‖ ≤ ‖ f n − f m ‖ ‖ x ‖ {\displaystyle \left\f_{n}x-f_{m}x\right\\\\reft(f_{n}-f_{m}\right\\ \leq\ \leq\ \ft\{n}-f_{m}\rig\x\ \} implies that ( f n x ) n = 1 ∞ {\displaystyle \left(f_{n}x\right)_{n=1}^{\infty }} is a Cauchy sequence in Y {\displaystyle Y} for every x ∈ X . {\displaystyle x\in X.} It follows that for every x ∈ X , {\displaystyle x\in X,} the limit lim n → ∞ f n x {\displaystyle \lim _{n\to \infty }f_ {n}x} 이(가) Y {\displaystyle Y} 에 있으므로 이 (필요하게 고유한) 한도 를 f x , {\displaystyle fx,}( 으)로 나타내겠다. 즉 ,
f x = 임이 있는 n → ∞ f n x . {\displaystyle fx~=~\lim _{n\to \ft \f_{n}x.}
f : X → Y {\displaystyle f:X\to Y} 이(가) 선형임을 알 수 있다.만약 ε > 0 {\displaystyle \varepsilon >0} 이면, 모든 큰 정수 n 과 m 에 대해 ‖ f - f ‖ x ‖ x x x x { { \ \\f_{n}-f_{m}\right\ \ ~\\\\\leq ~\leps. 그 뒤를 잇는다.
‖ f x − f m x ‖ ≤ ε ‖ x ‖ \displaystyle \fx-f_{m}x\right\ ~\leq ~\varepsilon \x\ } for sufficiently all large m . {\displaystyle m.} Hence ‖ f x ‖ ≤ ( ‖ f m ‖ + ε ) ‖ x ‖ , {\displaystyle \ fx\ \leq \left(\left\ f_{m}\right\ +\varepsilon \right)\ x\ ,} so that f ∈ L ( X , Y ) {\displaystyle f\in L(X,Y)} and ‖ f − f m ‖ ≤ ε . {\displaystyle \left\ f-f_{m}\right\ \leq \varepsilon .} 이 는 L ( X , Y )의 표준 위상에서의 f m → f {\displaystyle f_{m}\to f} 을(를) 보여준다. {\displaystyle L( X , Y ) . {\displaysty L(X,Y )의 완전성을 확립한다. } [9]
When Y {\displaystyle Y} is a scalar field (i.e. Y = C {\displaystyle Y=\mathbb {C} } or Y = R {\displaystyle Y=\mathbb {R} } ) so that L ( X , Y ) {\displaystyle L(X,Y)} is the dual space X ∗ {\displaystyle X^{*}} of X {\displaystyle X} .
정리 2 — X {\displaystyle X} 을(를) 표준 공간으로 하고 모든 x ∗ X ∗ ∗ { display {\displaystyle x^{*}\in X^{*}}}}}}에 대해 표준 공간으로 한다.
‖ x ∗ ‖ := up { ⟨ x , x ∗ ⟩ : x ∈ X 와 함께 ‖ x ‖ ≤ 1 } {\displaystyle \reft\x^{*}\right\\\cHB\{\langle x,x^{*}\rangele ~:~x\in X{\}\in X{\}\\leq 1\right\}}}}}}표시 형식 여기서 정의 x x , x ∗ := x ( ( x ) {\displaystyle \langle x,x^{*}\angle ~:=~x^{*(x)} 는 스칼라이다. 그러면.
‖ ⋅ : X ∗ → R {\ displaystyle \,\cdot \,\ :X^{*}\to \mathb {R}}} 은 X ∗ {\ displaysty X^{*} 를 바나흐 공간으로 만드는 규범이 다 .[10] B ∗{\ displaystyle B^{*}} 이 X ∗ {\displaystyle X^{*} 의 닫힌 단위 볼이라면, 모든 x ∈ X , {\displaystyle x\in X,} ‖ x ‖ = up { ⟨ x , x ∗ ⟩ : x ∗ ∈ B ∗ } = up { x ∗ ( x ) : ‖ x ∗ ‖ ≤ 1 와 함께 x ∗ ∈ X ∗ } . {\displaystyle {\begin{alignedat}{4}\ x\ ~&=~\sup \left\{ \langle x,x^{*}\rangle ~:~x^{*}\in B^{*}\right\}\\&=~\sup \left\{\left x^{*}(x)\right ~:~\left\ x^{*}\right\ \leq 1{\text{ with }}x^{*}\in X^{*}\right\}. \\end{aignatedat}}} 따라서 x ∗ ↦ x x , x ∗ { { { { { { { \displaystyle x^{*}\mapsto \langle x, x ^{*}\langle }}}} 은 (는) X ∗ {\displaystystystyle X^{*} 에 대한 경계 선형 기능 이다. B ∗{\ displaystyle B^{*}} 는 약*-compact이다. 참고 항목
메모들 ^ 루딘 1991 , 페이지 87 ^ 루딘 1991년 , 섹션 4.5, 페이지 95 ^ 루딘 1991 , 페이지 95 ^ 이 불평등은 다음과 같은 의미에서 팽팽하다: 어떤 x 에 대해서도 불평등이 평등하게 유지 되는 z가 있다. (비슷하게, 어떤 z 에도 평등을 주는 x 가 있다.) ^ 보이드 & 반덴버그 2004 , 페이지 637 ^ 각 L ( X , Y ) {\displaystyle L(X,Y)} 은 기능의 추가 및 스칼라 곱에 대한 일반적인 정의를 가진 벡터 공간이다. 이는 X {\displaystyle X} 이 아니라 Y {\displaysty Y} 의 벡터 공간 구조에만 의존한다. ^ 루딘 1991 페이지 92 ^ 루딘 1991 , 페이지 93 ^ 루딘 1991 , 페이지 93 ^ 알리프란티스 2006 , 페이지 230 harvnb 오류: 대상 없음: CITREFAliprantis2006(도움말 ) ^ 루딘 1991 , 정리 3.3 코롤라리, 페이지 59 ^ 루딘 1991 , 정리 3.15 바나흐-알라오글루 정리 알고리즘, 페이지 68 ^ 루딘 1991 , 페이지 94 참조 Aliprantis, Charalambos D.; Border, Kim C. (2006). Infinite Dimensional Analysis: A Hitchhiker's Guide (3rd ed.). Springer. ISBN 9783540326960 . Boyd, Stephen ; Vandenberghe, Lieven (2004). Convex Optimization . Cambridge University Press . ISBN 9780521833783 . Kolmogorov, A.N. ; Fomin, S.V. (1957). Elements of the Theory of Functions and Functional Analysis, Volume 1: Metric and Normed Spaces . Rochester: Graylock Press. Narici, Lawrence; Beckenstein, Edward (2011). Topological Vector Spaces . Pure and applied mathematics (Second ed.). Boca Raton, FL: CRC Press. ISBN 978-1584888666 . OCLC 144216834 . Rudin, Walter (1991). Functional Analysis . International Series in Pure and Applied Mathematics. Vol. 8 (Second ed.). New York, NY: McGraw-Hill Science/Engineering/Math . ISBN 978-0-07-054236-5 . OCLC 21163277 . Schaefer, Helmut H. ; Wolff, Manfred P. (1999). Topological Vector Spaces . GTM . Vol. 8 (Second ed.). New York, NY: Springer New York Imprint Springer. ISBN 978-1-4612-7155-0 . OCLC 840278135 . Trèves, François (2006) [1967]. Topological Vector Spaces, Distributions and Kernels . Mineola, N.Y.: Dover Publications. ISBN 978-0-486-45352-1 . OCLC 853623322 .
외부 링크
공간
정리 연산자 알헤브라스 문제 열기 적용들 고급 주제