The concept of uniform space isweil (Weil, A.) was introduced in 1938.Bulbaki(Bourbaki, N.) first gave a systematic exposition in 1940.In 1940, Tukey (J.W.) defined and studied the concept of equivalence of uniform spaces by covering families.There are three equivalent definitions of uniform space, namely, perimeter definition, pseudo metric definition and uniform coverage definition.[1]
In consistent structure andtopological structure The conceptual difference between them is that specific concepts about relative proximity and inter point proximity can be formalized in a uniform space.In other words, ideas like“xAdjacent toaThanyAdjacent tob”It makes sense in uniform space.Relatively, in generaltopologyIn space, given setA, BI can only say something meaningfulxAny AdjacentA(That is to say, in A'sclosureChinese), orAIs more thanBSmallerxOf“Neighborhood”However, inter point proximity and relative proximity cannot be described by topology alone.
A consistent structure is a structure on a set.Let X be a set and U be non empty of X × XSubset family。U is said to be a consistent structure on X if U meets the following conditions:
1. Each element of U contains diagonal Δ
2. If U ∈ U, then U ^ - 1 ∈ U, where:
U^-1={(x,y)|(y,x)∈U}.
3. If U ∈ U, then there is V ∈ U such that V ° VU,Including:
4. If U, V ∈ U, then U ∨ V ∈ U
5. If U ∈ U and UVX × X, then V ∈ U
The set X with uniform structure U is called uniform space, which is marked as (X, U). The concept of uniform space isweil (Weil, A.) was introduced in 1938.Bulbaki(Bourbaki, N.) first gave a systematic exposition in 1940.In 1940, Tukey (J.W.) defined and studied the concept of equivalence of uniform spaces by covering families.In his book published in 1964, Isbell, J.R., included the important development of the theory of uniform spaces described by covering. Uniform spaces can also be described by pseudo metric families, which were given by Bulbaki in 1948.
definition
Announce
edit
Uniform space has three equivalent definitions.
Surrounding definition
Uniform space(X, Φ) YesaggregateXEquipped withCartesian productX×XIs not emptySubset family(Φ is calledXOfConsistent structureoruniformityIts elements are calledaround(FrenchEntourage (neighbors or surroundings)) meet the following axioms:
IfUIn Φ, there isVIn Φ, so that as long as(x,y)And(y,z)OnVMedium, then(x,z)OnUMedium.
IfUIn Φ, thenU= { (y,x) : (x,y) ∈U}Also in Φ.
If the last property is omitted, the space is calledQuasi consistencyOf.Notice that two, three, two areFilterDefinition of.
Usually writeU[x]={y: (x,y)∈U}。On the graph, the typical surroundings are drawn around“y=x”Diagonal spots;U[x]They are longitudinal sections.If(x,y) ∈U, it can be said thatxandyIt is "U-adjacent".Similarly, ifXSubset ofAAll pairs of points in areU-Adjacent (that is, ifA×AIncluded inUMedium), thenAIt is called "U-Small".aroundUIs symmetrical if(y,x) ∈UJust in(x,y) ∈UWhen.The first axiom claims thatUEvery point isU-Adjacent to itself.The third axiom guarantees that "simultaneouslyU-Proximity andV-Proximity "is also a proximity relationship in consistency.The fourth axiom claims thatUThere is a surroundingVIt's "half big".The final axiom claims the essential symmetry of "proximity" with respect to uniform structures.
Consistent ΦSystem around foundationAny set around ΦB, so that the surrounding area of all ФBA collection of.Therefore, in general, the above property 2, the system around the foundationBSufficient unambiguous specification Φ: Φ is inclusiveBOf a collection ofX×XA collection of subsets of.All uniform spaces use a foundation surrounding system consisting of symmetrical surroundings.
The correct intuition about consistency can be derived frommetric spaceAn instance of provides: If(X,d)Is the metric space, set
there
FormedXThe standard consistent structure of the foundation surrounding the system.bexandyyesUa-AdjacentxAndyMaximum distance betweenaWhen.
Consistency Φ "fine" to another consistency Ψ on the same set, if Φ ⊇ Ψ;At this time Ψ is called "coarse" than Φ.
Pseudometric definition
Consistent space availablePseudo metricThe system is defined equivalently, which is the functional analysis (withSeminormPseudometrics provided) are particularly useful.More precisely, setf:X×X→RIs in the collectionXPseudo metric on.Inverse imageUa=f([0, a]) Fora>0 can be confirmed to form a consistent foundation surrounding system.fromUaThe generated consistency is determined by a single pseudo metricfThe definition is consistent.
For onXPseudometric Families on(fi), the consistent structure defined by this family is a separate pseudo metricfThe "minimum upper bound" of the uniform structure defined by i.This consistency is based on the surrounding system being measured by separate pseudo metricsfiIt is provided by the set of finite intersections defined all the time around.If the family of pseudo metrics is finite, it can be seen that the same consistent structure can be defined from a single pseudo metric, which is the "upper envelope" sup of the familyfi。
Less trivial, the surrounding system that can verify the basis of the allowable number (and therefore is specifically consistent with the definition of countable pseudo metric families) can be defined from a single pseudo metric.The conclusion is that any consistent structure can be defined from (possibly uncountable) pseudo metric families as above (see Bourbaki: Chapter IX § 1 no. 4 of General Topology).
Consistent coverage definition
Uniform space(X,Θ)Is a collectionXEquipped with a prominent "consistent coverage" familyΘ, it comes fromXThe set of coverage ofFilter。You can call it overlayPYes OverwriteQOfAsterisk Delicate(refinement) written asP<*Q, if for allA∈P, YesU∈QMake ifA∩B≠∅,B∈P, thenB⊆U。Axiomatization can be simplified as:
{10} Is consistent coverage.
IfP<*QalsoPIs consistent coverage, thenQIt is also consistent coverage.
IfPalsoQIf it is consistent coverage, there will be consistent coverageRdelicatePandQBoth.
Given a pointxAnd consistent coverageP, you can includexOfPOf members ofUnionThink isxSize ofPTypical ofNeighborhoodAnd this intuitive metric is consistently applied to this space.
Given a consistent space in the surrounding sense, define coveragePIs consistent, if there is a surroundingUSo that for eachx∈X, there is oneA∈PbringU[x]⊆A。These consistent covers form the consistent space of the second definition.Conversely, given a uniform space in the sense of uniform coverage{A×A:A∈P}BecausePThe value is taken from the uniform coverage, which is around the first defined uniform space.In addition, these two transformations are mutually inverse.
topology
Announce
edit
definition
All consistent spacesXCan becometopological space , by definitionXSubset ofOIs an open set, if and only ifOInxPresence aroundVbringV[x]YesOA subset of.In thistopologyMiddle, pointxOfNeighborhoodFilterYes{V[x]:V∈Φ}。This can be done byrecursionThe use of the "half big" around the existence of proof.Compared with general topological space, the existence of uniform structure makes it possible to compare neighborhood sizes:V[x]AndV[y]It is considered as "the same size".[3]
The topology defined by the consistent structure is calledThrow self consistency。In the topological space, the consistent structure is compatible with this topology. If the topology defined by the consistent structure is consistent with the original topology.Generally speaking, there are several different consistent structures that can be compatible withXThe given topology on the.
Unifiable space
topologySpace is calledUnifiableIf the consistent structure is compatible with this topology.
All unifiable spaces arecompletely regular Topology space.In addition, for a uniform spaceXThe following equivalents:
For any compatible consistent structure, all surrounding intersections are diagonal{(x,x) :x∈X}。
The topology of a unifiable space is alwaysSymmetric topology;That means this space isR0 space。
Conversely, every fully regular space is uniform.Compatible with fully regular spacesXOftopologyA consistency of can be defined as the coarsest consistency, which makes allXThe continuous real valued function on isUniform continuity。This consistency is based on the fact that the surrounding system provides(f×f)(V)All finite intersections offyesXContinuous real valued function on andVIs a consistent spaceRAround.This consistency defines a topology that is obviously coarser thanXInitial topology of;And it is also more refined than the original topology (so consistent with it)RegularityThe simple inference of: for anyx∈XandxOfNeighborhoodV, with continuous real valued functionfWithf(x)=0 and forVOfComplementThe point in is equal to 1.
In particular, compactHausdorff spaceIt is consistent.In fact, for compact Hausdorff spaceXstayX×XThe set of all neighborhoods of the middle diagonal forms a unique compatibility with thistopologyConsistency of.
HausdorfConsistent space isMeasurable spaceIf its consistency can be defined as countable pseudo metric family.In fact, as discussed in the definition of pseudo metric above, this consistency can be defined from a single pseudo metric. If this space is Hausdorff's, it must be a metric.In particular, ifVector spaceThe topology of is hausdorff and definable self countableSeminormFamily, it is measurable.
A uniformly continuous function is defined as the inverse image of its surroundings or the surrounding function, or equivalently, the inverse image of uniform coverage or the function of uniform coverage.
All uniformly continuous functions are related to thetopologyIt is continuous.
In consistent spaceXOnCauchyFilterFIt's a filterFSo that for all aroundU, existsA∈FWithA×A⊆U。In other words, a filter is a Cauchy filter if it contains an "arbitrarily small" set.It can be concluded from the definition thattopology)The convergent filters are all Cauchy filters.The Cauchy filter is called "minimal". If it does not contain smaller (or thicker) Cauchy filters (except for itself).It can be proved that all Cauchy filters contain a unique "very small CauchyFilter”。Of each pointNeighborhoodThe filter (consisting of all the neighbors of this point) is a very small Cauchy filter.
Conversely, the uniform space is calledcompleteIf all Cauchy filters converge.Any tightnessHausdorff spaceThey are all about complete uniform spaces compatible with the uniform structure of this topology.
Complete uniform spaces have the following important properties: iff:A→YIs from consistent spaceXDense subset ofATo complete uniform spaceYOfUniform continuityFunction, thenfCan be expanded (unique) into a wholeXUniformly continuous function on.
Hausdorff completeness of uniform spaces
likemetric space, all consistent spacesXallHausdorfcompletelyThat is to say, there is a complete Hausdorff uniform spaceYandUniform continuitymappingi:X→YIt has the following properties:
For anyXTo complete Hausdorff uniform spaceZUniform continuous mapping off, there is a uniqueUniformly continuous mappingg:Y→Zbringf=gi。
Hausdorf CompleteYIs the only top toisomorphism。As a setYCan be selected asXVery small Cauchy onFilterform.As eachXmidpointxOfNeighborhoodFilterB(x), mappingiIt can be defined asxMap toB(x)。Mapping so definediGenerally notMonomorphism;In fact, equivalence relationi(x) =i(x')'s image isXAll surrounding intersections of the, soiIt's a single shotXyesHausdorff spaceWhen.
stayYThe consistent structure on is defined as follows:V(That is to say, make(x,y)OnVHappens to be in(y,x)OnV), setC(V)All very small Cauchy with "at least one V-small set"FilterOf(F,G)Collection of.aggregateC(V)It can be confirmed that the foundation surrounding system is formed;This defines theY。
aggregatei(X)So it'sYA dense subset of.IfXyesHausdorff space, theniYes toi(X)Ofisomorphism, soXIt can be identified by its complete dense subset.In addition,i(X)AlwaysHausdorfOf;It is called associated withXOfHausdorff uniform space。IfRIndicate equivalence relationshipi(x) =i(x'), thenQuotient spaceX/RHomeomorphismtoi(X)。[4]
example
Announce
edit
Allmetric space(M,d)Can be regarded as a uniform space.In fact, because the metric is of course a pseudo metric, the definition of pseudo metric above givesMThe consistent structure of.This consistency is based on the surrounding system providing self aggregation.thisMThe consistent structure ofMNormal metric space ontopology。However, different metric spaces can have the same consistent structure (ordinary examples can be provided through metric constants).This consistent structure also generatesUniform continuityandCompleteness of metric spaceEquivalent definition of.
Using metrics, you can construct simple examples of different consistent structures with corresponding topologies.For example, setd1(x,y) = |x − y|Is onRNormal measurement on, and setd2(x,y) = |e− e|。Both of these metrics are raised in theRNormal ontopology, but the consistent structure is different, because {(x, y): | x − y |<1} isdAround the consistent structure of 1, but notd2.Informal, this example can be seen as selecting the normal consistency andUniform continuityThe function distorts it.
AllTopological groupG(Especially allTopological vector space)Become a consistent space if we defineG×GSubset ofVIs surrounding if and only if it contains the set {(x,y) :x⋅y∈U}ForGOfUnit yuanOne ofNeighborhoodU。thisGThe consistent structure on is calledGRight consistency on the because for allGIna, right multiplicationx→x⋅aIt's about this consistent structureUniform continuityOf.You can also defineGLeft consistency on the;They do not need to match, but they are generated inGGiven ontopology。
history
Announce
edit
stayAndre Veyto1937Before giving a clear definition of consistent structure for the first time, the concept of consistency is as followsCompletenessUsedmetric spaceDiscussion.Nicolas Bourbaki The book Topologie G é n é ral provides a definition based on the surrounding consistent structure, whileJohn TukeyThe definition of consistent coverage is given.weil The uniform spaces are also characterized by pseudo metric families.[5]