[go: nahoru, domu]

Jump to content

Approach space: Difference between revisions

From Wikipedia, the free encyclopedia
Content deleted Content added
→‎Definition: Examples added, expanded
 
(28 intermediate revisions by 21 users not shown)
Line 1: Line 1:
In [[topology]], '''approach spaces''' are a generalization of [[metric space]]s, based on point-to-set distances, instead of point-to-point distances. They were introduced by [http://www.math.ua.ac.be/TOP/ Robert Lowen] in [[1989]].
In [[topology]], a branch of [[mathematics]], '''approach spaces''' are a generalization of [[metric space]]s, based on point-to-[[set (mathematics)|set]] distances, instead of point-to-point distances. They were introduced by Robert Lowen in 1989, in a series of papers on approach theory between 1988 and 1995.


==Definition==
==Definition==


Given a metric space (''X'',''d''), or more generally, an [[Metric_space#Related_concepts_and_alternative_axiom_systems|extended pseudoquasimetric]] (which will be abbreviated ''xpq-metric'' here), one can define an induced map '''d''':''X''×P(''X'')→[0,∞] by '''d'''(x,''A'') = inf { ''d''(''x'',''a'' ) : ''a'' ∈ ''A'' }. With this example in mind, a '''distance''' on ''X'' is defined to be a map ''X''×P(''X'')→[0,∞] satisfying for all ''x'' in ''X'' and ''A'', ''B'' ⊆ ''X'',
Given a metric space (''X'', ''d''), or more generally, an [[Metric (mathematics)#Extending_the_range|extended]] [[Pseudometric_space|pseudo]][[quasimetric]] (which will be abbreviated ''∞pq-metric'' here), one can define an induced map '''d''': ''X'' × P(''X'')[0,] by '''d'''(''x'', ''A'') = [[infimum|inf]]{''d''(''x'', ''a'') : ''a'' ''A''}. With this example in mind, a '''distance''' on ''X'' is defined to be a map ''X'' × P(''X'')[0,] satisfying for all ''x'' in ''X'' and ''A'', ''B'' ''X'',
#'''d'''(''x'',{''x''}) = 0 ;
#'''d'''(''x'', {''x''}) = 0,
#'''d'''(''x'',{}) = ∞ ;
#'''d'''(''x'', Ø) = ∞,
#'''d'''(''x'',''A''∪''B'') = min '''d'''(''x'',''A''),'''d'''(''x'',''B'') ;
#'''d'''(''x'', ''A''''B'') = min('''d'''(''x'', ''A''), '''d'''(''x'', ''B'')),
#For all &epsilon;, 0&le;&epsilon;&le;&infin;, '''d'''(''x'',''A'') &le; '''d'''(''x'',''A''<sup>(&epsilon;)</sup>) + &epsilon; ;
#For all 0 ≤ ε ≤ ∞, '''d'''(''x'', ''A'') '''d'''(''x'', ''A''<sup>(ε)</sup>) + ε,
where ''A''<sup>(&epsilon;)</sup> = { ''x'' : '''d'''(''x'',''A'') &le; &epsilon; } by definition.
where we define ''A''<sup>(ε)</sup> = {''x'' : '''d'''(''x'', ''A'') ε}.


(The "empty infinum is positive infinity" convention is like the [[Empty_product#Nullary_intersection|nullary intersection is everything]] convention.)
(The "[[empty set|empty]] infimum is positive infinity" convention is like the [[Empty product#Nullary intersection|nullary intersection is everything]] convention.)


An approach space is defined to be a pair (''X'','''d''') where '''d''' is a distance function on ''X''. Every approach space has a topology, given by treating ''A''&nbsp;&rarr;&nbsp; ''A''<sup>(0)</sup> as a [[Kuratowski closure axioms|Kuratowski closure operator]].
An approach space is defined to be a pair (''X'',&thinsp;'''d''') where '''d''' is a distance function on ''X''. Every approach space has a [[topological space|topology]], given by treating ''A''&nbsp;&nbsp;''A''<sup>(0)</sup> as a [[Kuratowski closure axioms|Kuratowski closure operator]].


The appropriate maps between approach spaces are the ''contractions''. A map ''f'':(''X'','''d''')&rarr;(''Y'','''e''') is a contraction if '''e'''(''f''(''x''),''f''[''A'']) &le; '''d'''(''x'',''A'') for all ''x'' &isin; ''X'', ''A'' &sube; ''X''.
The appropriate maps between approach spaces are the ''contractions''. A map ''f'': (''X'',&thinsp;'''d''')(''Y'',&thinsp;'''e''') is a contraction if '''e'''(''f''(''x''), ''f''[''A'']) '''d'''(''x'', ''A'') for all ''x'' ''X'' and ''A'' ''X''.


==Examples==
==Examples==


Every xpq-metric space (''X'',''d'') can be ''distancized'' to (''X'','''d'''), as described at the beginning of the definition.
Every ∞pq-metric space (''X'', ''d'') can be ''distanced'' to (''X'',&thinsp;'''d'''), as described at the beginning of the definition.


Given a set ''X'', the ''discrete'' distance is given by '''d'''(''x'',''A'') = 0 if ''x'' &isin; ''A'' and = &infin; if ''x'' &notin; ''A''. The induced topology is the discrete topology.
Given a set ''X'', the ''discrete'' distance is given by '''d'''(''x'', ''A'') = 0 if ''x'' ''A'' and '''d'''(''x'', ''A'') = if ''x'' ''A''. The [[induced topology]] is the [[discrete topology]].


Given a set ''X'', the ''indiscrete'' distance is given by '''d'''(''x'',''A'') = 0 if ''A'' is non-empty, and = &infin; if ''A'' is empty. The induced topology is the indiscrete topology.
Given a set ''X'', the ''indiscrete'' distance is given by '''d'''(''x'', ''A'') = 0 if ''A'' is non-empty, and '''d'''(''x'', ''A'') = if ''A'' is empty. The induced topology is the indiscrete topology.


Given a topological space ''X'', a ''topological'' distance is given by '''d'''(''x'',''A'') = 0 if ''x'' &isin; ''A''&#x203e;, and = &infin; if not. The induced topology is the original topology. In fact, the only two-valued distances are the topological distances.
Given a [[topological space]] ''X'', a ''topological'' distance is given by '''d'''(''x'', ''A'') = 0 if ''x'' <span style="text-decoration: overline;">''A''</span>, and '''d'''(''x'', ''A'') = ∞ otherwise. The induced topology is the original topology. In fact, the only two-valued distances are the topological distances.


Let '''P'''=[0,&infin;], the extended positive reals. Let '''d'''<sup>+</sup>(''x'',''A'') = max (''x''&minus;sup ''A'',0) for ''x''&isin;'''P''' and ''A''&sube;'''P'''. Given any approach space (''X'','''d'''), the maps (for each ''A''&sube;''X'') '''d'''(.,''A'')&nbsp;:&nbsp;(''X'','''d''')&nbsp;&rarr;&nbsp;('''P''','''d'''<sup>+</sup>) are contractions.
Let '''P''' = [0,&thinsp;] be the [[extended real numbers|extended]] non-negative [[real number|reals]]. Let '''d'''<sup>+</sup>(''x'', ''A'') = max(''x'' − [[supremum|sup]]&thinsp;''A'', 0) for ''x'''''P''' and ''A'''''P'''. Given any approach space (''X'',&thinsp;'''d'''), the maps (for each ''A''''X'') '''d'''(., ''A'')&nbsp;:&nbsp;(''X'', '''d''')&nbsp;&nbsp;('''P''', '''d'''<sup>+</sup>) are contractions.


On '''P''', let '''e'''(''x'',''A'') = inf { |''x''&minus;''a''| : ''a''&isin;''A'' } for ''x''<&infin;, let '''e'''(&infin;,''A'') = 0 if ''A'' is unbounded, and let '''e'''(&infin;,''A'') = &infin; if ''A'' is bounded. Then ('''P''','''e''') is an approach space. Topologically, '''P''' is the one-point compactification of [0,&infin;). Note that '''e''' extends the ordinary Euclidean distance. This cannot be done with the ordinary Euclidean metric.
On '''P''', let '''e'''(''x'', ''A'') = inf{|''x''''a''| : ''a''''A''} for ''x'' <, let '''e'''(, ''A'') = 0 if ''A'' is unbounded, and let '''e'''(, ''A'') = if ''A'' is bounded. Then ('''P''',&thinsp;'''e''') is an approach space. Topologically, '''P''' is the one-point compactification of <nowiki>[0,&thinsp;)</nowiki>. Note that '''e''' extends the ordinary Euclidean distance. This cannot be done with the ordinary Euclidean metric.


Let &beta;'''N''' be the Stone-&#268;ech compactification of the integers. A point ''U''&isin;&beta;'''N''' is an ultrafilter on '''N'''. A subset ''A''&sube;&beta;'''N''' induces a filter ''F''(''A'')=&cap;{''U'':''U''&isin;''A''}. Let '''b'''(''U'',''A'') = sup { inf { |''n''-''j''| : ''n''&isin;''X'', ''j''&isin;''E'' } : ''X''&isin;''U'', ''E''&isin;''F''(''A'') }. Then (&beta;'''N''','''b''') is an approach space that extends the ordinary Euclidean distance on '''N'''. In contrast, &beta;'''N''' is not metrizable.
Let β'''N''' be the Stone–Čech compactification of the [[integer]]s. A point ''U'' ∈ β'''N''' is an ultrafilter on '''N'''. A subset ''A'' ⊆ β'''N''' induces a filter ''F''(''A'') =&thinsp;{''U'' : ''U''''A''}. Let '''b'''(''U'', ''A'') = sup{ inf{ |''n''''j''| : ''n''''X'', ''j''''E'' } : ''X''''U'', ''E''''F''(''A'') }. Then (β'''N''',&thinsp;'''b''') is an approach space that extends the ordinary Euclidean distance on '''N'''. In contrast, β'''N''' is not metrizable.


==Equivalent definitions==
==Equivalent definitions==
Line 37: Line 37:


Let XPQ(''X'') denote the set of xpq-metrics on ''X''. A subfamily ''G'' of XPQ(''X'') is called a ''gauge'' if
Let XPQ(''X'') denote the set of xpq-metrics on ''X''. A subfamily ''G'' of XPQ(''X'') is called a ''gauge'' if
#0 &isin; ''G'', where 0 is the zero metric, that is, 0(''x'',''y'')=0, all ''x'',''y'' ;
#0 ''G'', where 0 is the zero metric, that is, 0(''x'', ''y'') = 0 for all ''x'', ''y'',
#''e'' &le; ''d'' &isin; ''G'' implies ''e'' &isin; ''G'' ;
#''e'' ''d'' ''G'' implies ''e'' ''G'',
#''d'', ''e'' &isin; ''G'' implies max ''d'',''e'' &isin; ''G'' (the "max" here is the pointwise maximum);
#''d'', ''e'' ''G'' implies max(''d'',''e'') ''G'' (the "max" here is the [[pointwise maximum]]),
#For all ''d'' &isin; XPQ(''X''), if for all ''x'' &isin; ''X'', &epsilon;>0, ''N''<&infin; there is ''e'' &isin; ''G'' such that min(''d''(''x'',''y''),''N'') &le; ''e''(''x'',''y'') + &epsilon; for all ''y'', then ''d'' &isin; ''G'' .
#For all ''d'' XPQ(''X''), if for all ''x'' ''X'', ε > 0, ''N'' < there is ''e'' ''G'' such that min(''d''(''x'',''y''), ''N'') ''e''(''x'', ''y'') + ε for all ''y'', then ''d'' ''G''.


If ''G'' is a gauge on ''X'', then '''d'''(''x'',''A'') = sup { '''e'''(''x'',''a'') } : ''e'' &isin; ''G'' } is a distance function on ''X''. Conversely, given a distance function '''d''' on ''X'', the set of ''e'' &isin; XPQ(''X'') such that '''e''' &le; '''d''' is a gauge on ''X''. The two operations are inverse to each other.
If ''G'' is a gauge on ''X'', then '''d'''(''x'',''A'') = sup {'''e'''(''x'', ''a'') } : ''e'' &thinsp;''G''} is a distance function on ''X''. Conversely, given a distance function '''d''' on ''X'', the set of ''e'' XPQ(''X'') such that '''e''' '''d''' is a gauge on ''X''. The two operations are inverse to each other.


A contraction ''f'':(''X'','''d''')&rarr;(''Y'','''e''') is, in terms of associated gauges ''G'' and ''H'' respectively, a map such that for all ''d''&isin;''H'', ''d''(''f''(.),''f''(.))&isin;''G''.
A contraction ''f'': (''X'',&thinsp;'''d''')(''Y'',&thinsp;'''e''') is, in terms of associated gauges ''G'' and ''H'' respectively, a map such that for all ''d''''H'', ''d''(''f''(.), ''f''(.))''G''.


A ''tower'' on ''X'' is a set of maps ''A''&rarr;''A''<sup>[&epsilon;]</sup> for ''A''&sube;''X'', &epsilon;&ge;0, satisfying for all ''A'', ''B''&sube;''X'', &delta;, &epsilon; &ge; 0
A ''tower'' on ''X'' is a set of maps ''A''''A''<sup>[ε]</sup> for ''A''''X'', ε ≥ 0, satisfying for all ''A'', ''B''''X'' and δ, ε 0
#''A'' &sube; ''A''<sup>[&epsilon;]</sup> ;
#''A'' ''A''<sup>[ε]</sup>,
#{}<sup>[&epsilon;]</sup> = {} ;
#Ø<sup>[ε]</sup> = Ø,
#(''A''&cup;''B'')<sup>[&epsilon;]</sup> = ''A''<sup>[&epsilon;]</sup>&cup;''B''<sup>[&epsilon;]</sup> ;
#(''A''&thinsp;∪&thinsp;''B'')<sup>[ε]</sup> = ''A''<sup>[ε]</sup>&thinsp;∪&thinsp;''B''<sup>[ε]</sup>,
#''A''<sup>[&epsilon;][&delta;]</sup> &sube; ''A''<sup>[&epsilon;+&delta;]</sup> ;
#''A''<sup>[ε][δ]</sup> ''A''<sup>[ε+δ]</sup>,
#''A''<sup>[&epsilon;]</sup> = &cap;<sub>&delta;>&epsilon;</sub>''A''<sup>[&delta;]</sup> .
#''A''<sup>[ε]</sup> = <sub>δ>ε</sub>&thinsp;''A''<sup>[δ]</sup>.


Given a distance '''d''', the associated ''A''&rarr;''A''<sup>(&epsilon;)</sup> is a tower. Conversely, given a tower, the map '''d'''(''x'',''A'') = inf { &epsilon; : ''x'' &isin; ''A''<sup>[&epsilon;]</sup> } is a distance, and these two operations are inverses of each other.
Given a distance '''d''', the associated ''A''''A''<sup>(ε)</sup> is a tower. Conversely, given a tower, the map '''d'''(''x'',''A'') = inf{ε : ''x'' ''A''<sup>[ε]</sup>} is a distance, and these two operations are inverses of each other.


A contraction ''f'':(''X'','''d''')&rarr;(''Y'','''e''') is, in terms of associated towers, a map such that for all &epsilon;&ge;0, ''f''[''A''<sup>[&epsilon;]</sup>] &sube; ''f''[''A'']<sup>[&epsilon;]</sup>.
A contraction ''f'':(''X'',&thinsp;'''d''')(''Y'',&thinsp;'''e''') is, in terms of associated towers, a map such that for all ε ≥ 0, ''f''[''A''<sup>[ε]</sup>] ''f''[''A'']<sup>[ε]</sup>.

==Topological approach spaces==

Given a topological space ''X'', one can define a two-valued distance function by '''d'''(''x'',''A'') = 0 if ''x'' &isin; ''A''&#x203e;, and &infin; otherwise. Equivalently, the tower is identically the closure operator. One can prove that the only two-valued distance functions come from a topology.


==Categorical properties==
==Categorical properties==


The main interest in approach spaces and their contractions is that they form a category with good properties, while still being quantitative like metric spaces. One can take arbitrary products and coproducts and quotients, and the results appropriately generalize the corresponding results for topologies. One can even "distancize" such badly non-metrizable spaces like &beta;'''N''', the Stone-&#268;ech compactification of the integers.
The main interest in approach spaces and their contractions is that they form a [[category (mathematics)|category]] with good properties, while still being quantitative like metric spaces. One can take arbitrary [[Product (category theory)|products]], [[Coproduct|coproducts]], and quotients, and the results appropriately generalize the corresponding results for topologies. One can even "distancize" such badly non-metrizable spaces like β'''N''', the [[Stone–Čech compactification]] of the integers.

Certain hyperspaces, [[measure space|measure spaces]], and [[Probabilistic metric space|probabilistic metric spaces]] turn out to be naturally endowed with a distance. Applications have also been made to [[approximation theory]].


Certain hyperspaces, measure spaces, and probabilistic metric spaces turn out to be naturally endowed with a distance. Applications have also been made to approximation theory.
==References==
==References==
{{reflist}}
* {{cite book | last=Lowen | first=Robert | title=Approach spaces: the missing link in the topology-uniformity-metric triad | series=Oxford Mathematical Monographs | location=Oxford | publisher=[[Clarendon Press]] | year=1997 | isbn=0-19-850030-0 | zbl=0891.54001 }}
* {{cite book | last=Lowen | first=Robert | title=Index Analysis: Approach Theory at Work | publisher=Springer | year=2015}}

==External links==
* [http://www.math.ua.ac.be/TOP/ Robert Lowen]


[[Category:Closure operators]]
R. Lowen ''Approach Spaces: The Missing Link in the Topology-Uniformity-Metric Triad'' ISBN 0198500300.

Latest revision as of 06:31, 23 May 2021

In topology, a branch of mathematics, approach spaces are a generalization of metric spaces, based on point-to-set distances, instead of point-to-point distances. They were introduced by Robert Lowen in 1989, in a series of papers on approach theory between 1988 and 1995.

Definition[edit]

Given a metric space (X, d), or more generally, an extended pseudoquasimetric (which will be abbreviated ∞pq-metric here), one can define an induced map d: X × P(X) → [0,∞] by d(x, A) = inf{d(x, a) : aA}. With this example in mind, a distance on X is defined to be a map X × P(X) → [0,∞] satisfying for all x in X and A, BX,

  1. d(x, {x}) = 0,
  2. d(x, Ø) = ∞,
  3. d(x, AB) = min(d(x, A), d(x, B)),
  4. For all 0 ≤ ε ≤ ∞, d(x, A) ≤ d(x, A(ε)) + ε,

where we define A(ε) = {x : d(x, A) ≤ ε}.

(The "empty infimum is positive infinity" convention is like the nullary intersection is everything convention.)

An approach space is defined to be a pair (X, d) where d is a distance function on X. Every approach space has a topology, given by treating A → A(0) as a Kuratowski closure operator.

The appropriate maps between approach spaces are the contractions. A map f: (X, d) → (Y, e) is a contraction if e(f(x), f[A]) ≤ d(x, A) for all xX and AX.

Examples[edit]

Every ∞pq-metric space (X, d) can be distanced to (X, d), as described at the beginning of the definition.

Given a set X, the discrete distance is given by d(x, A) = 0 if xA and d(x, A) = ∞ if xA. The induced topology is the discrete topology.

Given a set X, the indiscrete distance is given by d(x, A) = 0 if A is non-empty, and d(x, A) = ∞ if A is empty. The induced topology is the indiscrete topology.

Given a topological space X, a topological distance is given by d(x, A) = 0 if xA, and d(x, A) = ∞ otherwise. The induced topology is the original topology. In fact, the only two-valued distances are the topological distances.

Let P = [0, ∞] be the extended non-negative reals. Let d+(x, A) = max(xsupA, 0) for xP and AP. Given any approach space (X, d), the maps (for each AX) d(., A) : (X, d) → (P, d+) are contractions.

On P, let e(x, A) = inf{|xa| : aA} for x < ∞, let e(∞, A) = 0 if A is unbounded, and let e(∞, A) = ∞ if A is bounded. Then (P, e) is an approach space. Topologically, P is the one-point compactification of [0, ∞). Note that e extends the ordinary Euclidean distance. This cannot be done with the ordinary Euclidean metric.

Let βN be the Stone–Čech compactification of the integers. A point U ∈ βN is an ultrafilter on N. A subset A ⊆ βN induces a filter F(A) = ∩ {U : UA}. Let b(U, A) = sup{ inf{ |nj| : nX, jE } : XU, EF(A) }. Then (βN, b) is an approach space that extends the ordinary Euclidean distance on N. In contrast, βN is not metrizable.

Equivalent definitions[edit]

Lowen has offered at least seven equivalent formulations. Two of them are below.

Let XPQ(X) denote the set of xpq-metrics on X. A subfamily G of XPQ(X) is called a gauge if

  1. 0 ∈ G, where 0 is the zero metric, that is, 0(x, y) = 0 for all x, y,
  2. edG implies eG,
  3. d, eG implies max(d,e) ∈ G (the "max" here is the pointwise maximum),
  4. For all d ∈ XPQ(X), if for all xX, ε > 0, N < ∞ there is eG such that min(d(x,y), N) ≤ e(x, y) + ε for all y, then dG.

If G is a gauge on X, then d(x,A) = sup {e(x, a) } : e ∈ G} is a distance function on X. Conversely, given a distance function d on X, the set of e ∈ XPQ(X) such that ed is a gauge on X. The two operations are inverse to each other.

A contraction f: (X, d) → (Y, e) is, in terms of associated gauges G and H respectively, a map such that for all dH, d(f(.), f(.)) ∈ G.

A tower on X is a set of maps AA[ε] for AX, ε ≥ 0, satisfying for all A, BX and δ, ε ≥ 0

  1. AA[ε],
  2. Ø[ε] = Ø,
  3. (A ∪ B)[ε] = A[ε] ∪ B[ε],
  4. A[ε][δ]A[ε+δ],
  5. A[ε] = ∩δ>εA[δ].

Given a distance d, the associated AA(ε) is a tower. Conversely, given a tower, the map d(x,A) = inf{ε : xA[ε]} is a distance, and these two operations are inverses of each other.

A contraction f:(X, d)→(Y, e) is, in terms of associated towers, a map such that for all ε ≥ 0, f[A[ε]] ⊆ f[A][ε].

Categorical properties[edit]

The main interest in approach spaces and their contractions is that they form a category with good properties, while still being quantitative like metric spaces. One can take arbitrary products, coproducts, and quotients, and the results appropriately generalize the corresponding results for topologies. One can even "distancize" such badly non-metrizable spaces like βN, the Stone–Čech compactification of the integers.

Certain hyperspaces, measure spaces, and probabilistic metric spaces turn out to be naturally endowed with a distance. Applications have also been made to approximation theory.

References[edit]

  • Lowen, Robert (1997). Approach spaces: the missing link in the topology-uniformity-metric triad. Oxford Mathematical Monographs. Oxford: Clarendon Press. ISBN 0-19-850030-0. Zbl 0891.54001.
  • Lowen, Robert (2015). Index Analysis: Approach Theory at Work. Springer.

External links[edit]