Movatterモバイル変換


[0]ホーム

URL:


Jump to content
WikipediaThe Free Encyclopedia
Search

Surjection of Fréchet spaces

From Wikipedia, the free encyclopedia
Characterization of surjectivity
This article has multiple issues. Please helpimprove it or discuss these issues on thetalk page.(Learn how and when to remove these messages)
This article includes a list ofgeneral references, butit lacks sufficient correspondinginline citations. Please help toimprove this article byintroducing more precise citations.(May 2020) (Learn how and when to remove this message)
icon
This articleneeds additional citations forverification. Please helpimprove this article byadding citations to reliable sources. Unsourced material may be challenged and removed.
Find sources: "Surjection of Fréchet spaces" – news ·newspapers ·books ·scholar ·JSTOR
(May 2020) (Learn how and when to remove this message)
(Learn how and when to remove this message)

The theorem on thesurjection ofFréchet spaces is an important theorem, due toStefan Banach,[1] that characterizes when acontinuous linear operator between Fréchet spaces is surjective.

The importance of this theorem is related to theopen mapping theorem, which states that a continuous linear surjection between Fréchet spaces is anopen map. Often in practice, one knows that they have a continuous linear map between Fréchet spaces and wishes to show that it is surjective in order to use the open mapping theorem to deduce that it is also an open mapping. This theorem may help reach that goal.

Preliminaries, definitions, and notation

[edit]

LetL:XY{\displaystyle L:X\to Y} be a continuous linear map betweentopological vector spaces.

Thecontinuous dual space ofX{\displaystyle X} is denoted byX.{\displaystyle X^{\prime }.}

Thetranspose ofL{\displaystyle L} is the maptL:YX{\displaystyle {}^{t}L:Y^{\prime }\to X^{\prime }} defined byL(y):=yL.{\displaystyle L\left(y^{\prime }\right):=y^{\prime }\circ L.} IfL:XY{\displaystyle L:X\to Y} is surjective thentL:YX{\displaystyle {}^{t}L:Y^{\prime }\to X^{\prime }} will beinjective, but the converse is not true in general.

Theweak topology onX{\displaystyle X} (resp.X{\displaystyle X^{\prime }}) is denoted byσ(X,X){\displaystyle \sigma \left(X,X^{\prime }\right)} (resp.σ(X,X){\displaystyle \sigma \left(X^{\prime },X\right)}). The setX{\displaystyle X} endowed with this topology is denoted by(X,σ(X,X)).{\displaystyle \left(X,\sigma \left(X,X^{\prime }\right)\right).} The topologyσ(X,X){\displaystyle \sigma \left(X,X^{\prime }\right)} is the weakest topology onX{\displaystyle X} making all linear functionals inX{\displaystyle X^{\prime }} continuous.

IfSY{\displaystyle S\subseteq Y} then thepolar ofS{\displaystyle S} inY{\displaystyle Y} is denoted byS.{\displaystyle S^{\circ }.}

Ifp:XR{\displaystyle p:X\to \mathbb {R} } is aseminorm onX{\displaystyle X}, thenXp{\displaystyle X_{p}} will denoted the vector spaceX{\displaystyle X} endowed with the weakestTVS topology makingp{\displaystyle p} continuous.[1] A neighborhood basis ofXp{\displaystyle X_{p}} at the origin consists of the sets{xX:p(x)<r}{\displaystyle \left\{x\in X:p(x)<r\right\}} asr{\displaystyle r} ranges over the positive reals. Ifp{\displaystyle p} is not a norm thenXp{\displaystyle X_{p}} is notHausdorff andkerp:={xX:p(x)=0}{\displaystyle \ker p:=\left\{x\in X:p(x)=0\right\}} is a linear subspace ofX{\displaystyle X}. Ifp{\displaystyle p} is continuous then the identity mapId:XXp{\displaystyle \operatorname {Id} :X\to X_{p}} is continuous so we may identify the continuous dual spaceXp{\displaystyle X_{p}^{\prime }} ofXp{\displaystyle X_{p}} as a subset ofX{\displaystyle X^{\prime }} via the transpose of the identity maptId:XpX,{\displaystyle {}^{t}\operatorname {Id} :X_{p}^{\prime }\to X^{\prime },} which isinjective.

Surjection of Fréchet spaces

[edit]

Theorem[1] (Banach)IfL:XY{\displaystyle L:X\to Y} is a continuous linear map between two Fréchet spaces, thenL:XY{\displaystyle L:X\to Y} is surjective if and only if the following two conditions both hold:

  1. tL:YX{\displaystyle {}^{t}L:Y^{\prime }\to X^{\prime }} isinjective, and
  2. theimage oftL,{\displaystyle {}^{t}L,} denoted byImtL,{\displaystyle \operatorname {Im} {}^{t}L,} is weakly closed inX{\displaystyle X^{\prime }} (i.e. closed whenX{\displaystyle X^{\prime }} is endowed with the weak-* topology).

Extensions of the theorem

[edit]

Theorem[1]IfL:XY{\displaystyle L:X\to Y} is a continuous linear map between two Fréchet spaces then the following are equivalent:

  1. L:XY{\displaystyle L:X\to Y} is surjective.
  2. The following two conditions hold:
    1. tL:YX{\displaystyle {}^{t}L:Y^{\prime }\to X^{\prime }} isinjective;
    2. theimageImtL{\displaystyle \operatorname {Im} {}^{t}L} oftL{\displaystyle {}^{t}L} is weakly closed inX.{\displaystyle X^{\prime }.}
  3. For every continuous seminormp{\displaystyle p} onX{\displaystyle X} there exists a continuous seminormq{\displaystyle q} onY{\displaystyle Y} such that the following are true:
    1. for everyyY{\displaystyle y\in Y} there exists somexX{\displaystyle x\in X} such thatq(L(x)y)=0{\displaystyle q(L(x)-y)=0};
    2. for everyyY,{\displaystyle y^{\prime }\in Y,} iftL(y)Xp{\displaystyle {}^{t}L\left(y^{\prime }\right)\in X_{p}^{\prime }} thenyYq.{\displaystyle y^{\prime }\in Y_{q}^{\prime }.}
  4. For every continuous seminormp{\displaystyle p} onX{\displaystyle X} there exists a linear subspaceN{\displaystyle N} ofY{\displaystyle Y} such that the following are true:
    1. for everyyY{\displaystyle y\in Y} there exists somexX{\displaystyle x\in X} such thatL(x)yN{\displaystyle L(x)-y\in N};
    2. for everyyY,{\displaystyle y^{\prime }\in Y^{\prime },} iftL(y)Xp{\displaystyle {}^{t}L\left(y^{\prime }\right)\in X_{p}^{\prime }} thenyN.{\displaystyle y^{\prime }\in N^{\circ }.}
  5. There is anon-increasing sequenceN1N2N3{\displaystyle N_{1}\supseteq N_{2}\supseteq N_{3}\supseteq \cdots } of closed linear subspaces ofY{\displaystyle Y} whose intersection is equal to{0}{\displaystyle \{0\}} and such that the following are true:
    1. for everyyY{\displaystyle y\in Y} and every positive integerk{\displaystyle k}, there exists somexX{\displaystyle x\in X} such thatL(x)yNk{\displaystyle L(x)-y\in N_{k}};
    2. for every continuous seminormp{\displaystyle p} onX{\displaystyle X} there exists an integerk{\displaystyle k} such that anyxX{\displaystyle x\in X} that satisfiesL(x)Nk{\displaystyle L(x)\in N_{k}} is the limit, in the sense of the seminormp{\displaystyle p}, of a sequencex1,x2,{\displaystyle x_{1},x_{2},\ldots } in elements ofX{\displaystyle X} such thatL(xi)=0{\displaystyle L\left(x_{i}\right)=0} for alli.{\displaystyle i.}

Lemmas

[edit]

The following lemmas are used to prove the theorems on the surjectivity of Fréchet spaces. They are useful even on their own.

Theorem[1]LetX{\displaystyle X} be a Fréchet space andZ{\displaystyle Z} be a linear subspace ofX.{\displaystyle X^{\prime }.} The following are equivalent:

  1. Z{\displaystyle Z} is weakly closed inX{\displaystyle X^{\prime }};
  2. There exists a basisB{\displaystyle {\mathcal {B}}} of neighborhoods of the origin ofX{\displaystyle X} such that for everyBB,{\displaystyle B\in {\mathcal {B}},}BZ{\displaystyle B^{\circ }\cap Z} is weakly closed;
  3. The intersection ofZ{\displaystyle Z} with every equicontinuous subsetE{\displaystyle E} ofX{\displaystyle X^{\prime }} is relatively closed inE{\displaystyle E} (whereX{\displaystyle X^{\prime }} is given the weak topology induced byX{\displaystyle X} andE{\displaystyle E} is given the subspace topology induced byX{\displaystyle X^{\prime }}).

Theorem[1]On the dualX{\displaystyle X^{\prime }} of a Fréchet spaceX{\displaystyle X}, the topology of uniform convergence on compact convex subsets ofX{\displaystyle X} is identical to the topology ofuniform convergence on compact subsets ofX{\displaystyle X}.

Theorem[1]LetL:XY{\displaystyle L:X\to Y} be a linear map between Hausdorfflocally convex TVSs, withX{\displaystyle X} alsometrizable. If the mapL:(X,σ(X,X))(Y,σ(Y,Y)){\displaystyle L:\left(X,\sigma \left(X,X^{\prime }\right)\right)\to \left(Y,\sigma \left(Y,Y^{\prime }\right)\right)} is continuous thenL:XY{\displaystyle L:X\to Y} is continuous (whereX{\displaystyle X} andY{\displaystyle Y} carry their original topologies).

Applications

[edit]

Borel's theorem on power series expansions

[edit]

Theorem[2] (E. Borel)Fix a positive integern{\displaystyle n}. IfP{\displaystyle P} is an arbitraryformal power series inn{\displaystyle n} indeterminates with complex coefficients then there exists aC{\displaystyle {\mathcal {C}}^{\infty }} functionf:RnC{\displaystyle f:\mathbb {R} ^{n}\to \mathbb {C} } whose Taylor expansion at the origin is identical toP{\displaystyle P}.

That is, suppose that for everyn{\displaystyle n}-tuple of non-negative integersp=(p1,,pn){\displaystyle p=\left(p_{1},\ldots ,p_{n}\right)} we are given a complex numberap{\displaystyle a_{p}} (with no restrictions). Then there exists aC{\displaystyle {\mathcal {C}}^{\infty }} functionf:RnC{\displaystyle f:\mathbb {R} ^{n}\to \mathbb {C} } such thatap=(/x)pf|x=0{\displaystyle a_{p}=\left(\partial /\partial x\right)^{p}f{\bigg \vert }_{x=0}} for everyn{\displaystyle n}-tuplep.{\displaystyle p.}

Linear partial differential operators

[edit]
See also:Distribution (mathematics)

Theorem[3]LetD{\displaystyle D} be a linear partial differential operator withC{\displaystyle {\mathcal {C}}^{\infty }} coefficients in an open subsetURn.{\displaystyle U\subseteq \mathbb {R} ^{n}.} The following are equivalent:

  1. For everyfC(U){\displaystyle f\in {\mathcal {C}}^{\infty }(U)} there exists someuC(U){\displaystyle u\in {\mathcal {C}}^{\infty }(U)} such thatDu=f.{\displaystyle Du=f.}
  2. U{\displaystyle U} isD{\displaystyle D}-convex andD{\displaystyle D} is semiglobally solvable.

D{\displaystyle D} beingsemiglobally solvable inU{\displaystyle U} means that for everyrelatively compact open subsetV{\displaystyle V} ofU{\displaystyle U}, the following condition holds:

to everyfC(U){\displaystyle f\in {\mathcal {C}}^{\infty }(U)} there is somegC(U){\displaystyle g\in {\mathcal {C}}^{\infty }(U)} such thatDg=f{\displaystyle Dg=f} inV{\displaystyle V}.

U{\displaystyle U} beingD{\displaystyle D}-convex means that for every compact subsetKU{\displaystyle K\subseteq U} and every integern0,{\displaystyle n\geq 0,} there is a compact subsetCn{\displaystyle C_{n}} ofU{\displaystyle U} such that for everydistributiond{\displaystyle d} with compact support inU{\displaystyle U}, the following condition holds:

iftDd{\displaystyle {}^{t}Dd} is of ordern{\displaystyle \leq n} and ifsupptDdK,{\displaystyle \operatorname {supp} {}^{t}Dd\subseteq K,} thensuppdCn.{\displaystyle \operatorname {supp} d\subseteq C_{n}.}

See also

[edit]

References

[edit]
  1. ^abcdefgTrèves 2006, pp. 378–384.
  2. ^Trèves 2006, p. 390.
  3. ^Trèves 2006, p. 392.

Bibliography

[edit]
Spaces
Properties
Theorems
Operators
Algebras
Open problems
Applications
Advanced topics
Retrieved from "https://en.wikipedia.org/w/index.php?title=Surjection_of_Fréchet_spaces&oldid=1184504087"
Category:
Hidden categories:

[8]ページ先頭

©2009-2025 Movatter.jp