This paper was converted on www.awesomepapers.org from LaTeX by an anonymous user.
Want to know more? Visit the Converter page.

SOME PROPERTIES OF PROXIMAL HOMOTOPY THEORY

MELİH İS and İSMET KARACA Melih Is, Ege University
Faculty of Sciences
Department of Mathematics
Izmir, Türkiye
[email protected] Ismet Karaca1, Ege University
Faculty of Science
Department of Mathematics
Izmir, Türkiye
Ismet Karaca2, Azerbaijan State Agrarian University
Faculty Agricultural Economics
Department of Agrarian Economics
Gence, Azerbaijan
[email protected]
(Date: February 23, 2025)
Abstract.

Nearness theory comes into play in homotopy theory because the notion of closeness between points is essential in determining whether two spaces are homotopy equivalent. While nearness theory and homotopy theory have different focuses and tools, they are intimately connected through the concept of a metric space and the notion of proximity between points, which plays a central role in both areas of mathematics. This manuscript investigates some concepts of homotopy theory in proximity spaces. Moreover, these concepts are taken into account in descriptive proximity spaces.

Key words and phrases:
Proximity, descriptive proximity, homotopy, fibration, cofibration
2010 Mathematics Subject Classification:
54E05, 54E17, 14D06, 55P05

1. Introduction

Topological perspective first appears in the scientific works of Riemann and Poincare in the 19th century[19, 18]. The concept reveals that the definitions of topological space emerge either through Kuratowski’s closure operator[2] or through the use of open sets. Given the Kuratowski’s closure operator, there are many strategies and approaches that seem useful in different situations and are worth developing, as in nearness theory. Proximity spaces are created by reflection of the concept of being near/far on sets. Given a nonempty set XX, and any subsets EE, FXF\subset X, we say that EE is near FF if EFE\cap F\neq\emptyset. A method based on the idea of near sets is first proposed by Riesz, is revived by Wallace, and is axiomatically elaborated it by Efremovic[20, 22, 1]. Let XX be a nonempty set. A proximity is a binary relation (actually a nearness relation) defined on subsets of XX and generally denoted by δ\delta. One can construct a topology on XX induced by the pair (X,δ)(X,\delta) using the closure operator (named a proximity space). Indeed, for any point xXx\in X, if {x}\{x\} is near EE, then xE¯x\in\overline{E}. In symbols, if {x}δE\{x\}\delta E, then xδE¯x\delta\overline{E}. It should be noted that the notation xδEx\delta E is sometimes used instead of {x}δE\{x\}\delta E for abbreviation (in particular in a metric space). It appears that several proximities may correspond in this way to the same topology on XX. Moreover, several topological conclusions can be inferred from claims made about proximity spaces.

The near set theory is reasonably improved by Smirnov’s compactification, Leader’s non-symmetric proximity, and Lodato’s symmetric proximity[21, 3, 4]. Peters also contributes to the theory of nearness by introducing the concept of spatial nearness and descriptive nearness[9, 10]. In addition, the strong structure of proximity spaces stands out in the variety of application areas: In[6], it is possible to see the construction of proximity spaces in numerous areas such as cell biology, the topology of digital images, visual marketing, and so on. In a broader context, the application areas of near spaces are listed along with the history of the subject in [11]. According to this, some near set theory-related topics are certain engineering problems, image analysis, and human perception. The main subject of this article, algebraic topology approaches in proximity spaces, is a work in progress in the literature. Mapping spaces, one of the fundamental concepts in homotopy theory, is examined in proximity spaces in [8]. The proximal setting of the notion fibration is first defined in [17]. Peters and Vergili have recently published interesting research on descriptive proximal homotopy, homotopy cycles, path cycles, and Lusternik-Schnirelmann theory of proximity spaces[14, 15, 16, 17].

This paper is primarily concerned with the theory of proximal homotopy and is organized as follows. In Section 2, we discuss the general properties of proximity and descriptive proximity spaces. Section 3 covered 4 main topics in proximity spaces: Mapping spaces, covering spaces, fibrations, and cofibrations. They provide different types of examples and frequently used algebraic topology results in proximal homotopy cases. Next, the descriptive proximal homotopy theory, which is handled in Section 5, discusses the ideas from Section 3.3 and illustrates them with examples by using feature vectors as color scales. Finally, the last section establishes the direction for future works by clearly emphasizing the application areas of homotopy theory.

2. Preliminaries

Before proceeding with the main results, it is critical to remember the fundamental characteristics of proximity and descriptive proximity spaces.

2.1. On Proximity Spaces

Consider a pseudo-metric space (X,d)(X,d). A binary relation δ\delta defined by

E′′δFD(E,F)=0′′{}^{\prime\prime}E\delta F\ \ \Leftrightarrow\ \ D(E,F)=0^{\prime\prime}

satisfies

(a)EδFFδE,\displaystyle\textbf{(a)}\hskip 28.45274ptE\delta F\ \ \Rightarrow\ \ F\delta E,
(b)(EF)δGEδGFδG,\displaystyle\textbf{(b)}\hskip 28.45274pt(E\cup F)\delta G\ \ \Leftrightarrow\ \ E\delta G\ \vee\ F\delta G,
(c)EδFEF,\displaystyle\textbf{(c)}\hskip 28.45274ptE\delta F\ \ \Rightarrow\ \ E\neq\emptyset\ \land\ F\neq\emptyset,
(d)Eδ¯FGX:Eδ¯G(XG)δ¯F,\displaystyle\textbf{(d)}\hskip 28.45274ptE\underline{\delta}F\ \ \Rightarrow\ \ \exists G\subset X:E\underline{\delta}G\ \land\ (X-G)\underline{\delta}F,
(e)EFEδF\displaystyle\textbf{(e)}\hskip 28.45274ptE\cap F\neq\emptyset\ \ \Rightarrow\ \ E\delta F

for D(E,F)=inf{d(x1,x2):x1E,x2F}D(E,F)=\inf\{d(x_{1},x_{2}):x_{1}\in E,x_{2}\in F\}[1, 21, 5].

δ\delta is a nearness relation and EδFE\delta F is read as “EE is near FF”. Otherwise, the notation Eδ¯FE\underline{\delta}F means that “EE is far from FF”.

Definition 2.1.

[1, 21, 5] The nearness relation δ\delta for the subsets of XX is said to be an Efremovic proximity (simply denoted by EF-proximity or proximity) provided that δ\delta satisfies (a)-(e). (X,δ)(X,\delta) is said to be an EF-proximity (or proximity) space.

As an example of a proximity space, the discrete proximity δ\delta on a (nonempty) set XX is defined by “EδFEFE\delta F\ \Leftrightarrow\ E\cap F\neq\emptyset” for EE, FXF\subset X. Also, the indiscrete proximity δ\delta^{{}^{\prime}} on a (nonempty) set XX is given by EδFE\delta^{{}^{\prime}}F for any nonempty subsets EE and FF in XX. A subset EE of XX with a proximity δ\delta is a closed set if “xδExEx\delta E\ \Rightarrow\ x\in E”. The converse is also valid. Therefore, given a proximity δ\delta on XX, a topology τ(δ)\tau(\delta) is defined by the family of complements of all closed sets via Kuratowski closure operator[5].

Theorem 2.2.

[5] For a proximity δ\delta and a topology τ(δ)\tau(\delta) on a set XX, we have that the closure E¯\overline{E} coincides with {x:xδE}\{x:x\delta E\}.

Given any proximities δ\delta and δ\delta^{{}^{\prime}} on respective sets XX and XX^{{}^{\prime}}, a map hh from XX to XX^{{}^{\prime}} is called proximally continuous if “EδFh(E)δh(F)E\delta F\ \Rightarrow\ h(E)\delta^{{}^{\prime}}h(F)” for EE, FXF\subset X[1, 21]. We denote a proximally continuous map by “pc-map”. Given a proximity δ\delta on XX and a subset EXE\subset X, a subspace proximity δE\delta_{E} is defined on the subsets of EE as follows[5]: “E1δE2E1δEE2E_{1}\delta E_{2}\ \Leftrightarrow\ E_{1}\delta_{E}E_{2}” for E1E_{1}, E2EE_{2}\subset E. Let (X,δ)(X,\delta) be a proximity space and (E,δE)(E,\delta_{E}) a subspace proximity. A pc-map k:(X,δ)(E,δE)k:(X,\delta)\rightarrow(E,\delta_{E}) is a proximal retraction provided that kjk\circ j is an identity map on 1E1_{E}, where j:(E,δE)(X,δ)j:(E,\delta_{E})\rightarrow(X,\delta) is an inclusion map.

Lemma 2.3.

[14](Gluing Lemma) Assume that f1:(X,δ1)(Y,δ2)f_{1}:(X^{{}^{\prime}},\delta_{1}^{{}^{\prime}})\rightarrow(Y^{{}^{\prime}},\delta_{2}^{{}^{\prime}}) and f2:(X′′,δ1′′)(Y,δ2)f_{2}:(X^{{}^{\prime\prime}},\delta_{1}^{{}^{\prime\prime}})\rightarrow(Y^{{}^{\prime}},\delta_{2}^{{}^{\prime}}) are pc-maps with the property that they agree on the intersection of XX and X′′X^{{}^{\prime\prime}}. Then the map f1f2:(XX′′,δ)(Y,δ2)f_{1}\cup f_{2}:(X^{{}^{\prime}}\cup X^{{}^{\prime\prime}},\delta)\rightarrow(Y^{{}^{\prime}},\delta_{2}^{{}^{\prime}}), defined by f1f2(s)={f1(s),sXf2(s),sX′′f_{1}\cup f_{2}(s)=\begin{cases}f_{1}(s),&s\in X^{{}^{\prime}}\\ f_{2}(s),&s\in X^{{}^{\prime\prime}}\end{cases} for any sXX′′s\in X^{{}^{\prime}}\cup X^{{}^{\prime\prime}}, is a pc-map.

We say that hh is a proximity isomorphism provided that hh is a bijection and each of hh and h1h^{-1} is pc-map[5]. According to this, (X,δ)(X,\delta) and (X,δ)(X^{{}^{\prime}},\delta^{{}^{\prime}}) are said to be proximally isomorphic spaces. Another important proximity relation is given on the subsets of the cartesian product of two proximity spaces as follows[3]: Let δ\delta and δ\delta^{{}^{\prime}} be any proximities on respective sets XX and XX^{{}^{\prime}}. For any subsets E1×E2E_{1}\times E_{2} and F1×F2F_{1}\times F_{2} of X×XX\times X^{{}^{\prime}}, E1×E2E_{1}\times E_{2} is near F1×F2F_{1}\times F_{2} if E1δF1E_{1}\delta F_{1} and E2δF2E_{2}\delta^{{}^{\prime}}F_{2}.

Definition 2.4.

[14] Given two pc-maps h1h_{1} and h2h_{2} from XX to XX^{{}^{\prime}}, if there is a pc-map FF from X×IX\times I to XX^{{}^{\prime}} with the properties F(x,0)=h1(x)F(x,0)=h_{1}(x) and F(x,1)=h2(x)F(x,1)=h_{2}(x), then h1h_{1} and h2h_{2} are called proximally homotopic maps.

The map FF in Definition 2.4 is said to be a proximal homotopy between hh and hh^{{}^{\prime}}. We simply denote a proximal homotopy by “prox-hom”. Similar to topological spaces, prox-hom is an equivalence relation on proximity spaces. Let δ\delta be a proximity on XX and EXE\subset X. EE is called a δ\delta-neighborhood of FF, denoted by FδEF\ll_{\delta}E, provided that Fδ¯(XE)F\underline{\delta}(X-E)[5]. The proximal continuity of any function h:(X,δ)(X,δ)h:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) can also be expressed as

E′′δFh1(E)δh1(F)′′{}^{\prime\prime}E\ll_{\delta^{{}^{\prime}}}F\ \Rightarrow\ h^{-1}(E)\ll_{\delta}h^{-1}(F)^{\prime\prime}

for any EE, FXF\subset X^{\prime}.

Theorem 2.5.

[5] Let EkδFkE_{k}\ll_{\delta}F_{k} for k=1,,rk=1,\cdots,r. Then

k=1rEkδk=1rFkandk=1rEkδk=1rFk.\displaystyle\displaystyle\bigcap_{k=1}^{r}E_{k}\ll_{\delta}\displaystyle\bigcap_{k=1}^{r}F_{k}\ \ \ \text{and}\ \ \ \displaystyle\bigcup_{k=1}^{r}E_{k}\ll_{\delta}\displaystyle\bigcup_{k=1}^{r}F_{k}.
Definition 2.6.

[14] For any two elements x1x_{1} and x2x_{2} in XX with a proximity δ\delta, a proximal path from x1x_{1} to x2x_{2} in XX is a pc-map hh from I=[0,1]I=[0,1] to XX for which h(0)=x1h(0)=x_{1} and h(1)=x2h(1)=x_{2}.

The proximal continuity of the proximal path h:IXh:I\rightarrow X in Definition 2.6 means that “D(E,F)=0h(E)δh(F)D(E,F)=0\ \Rightarrow\ h(E)\delta h(F)” for EE, FIF\in I. Recall that XX is a connected proximity space if and only if for all nonempty EE, F𝒫(X)F\in\mathcal{P}(X), EF=XE\cup F=X implies that EδFE\delta F[7]. Let δ\delta be a proximity on XX. Then XX is called a path-connected proximity space if, for any points x1x_{1} and x2x_{2} in XX, there exists a proximal path from x1x_{1} to x2x_{2} in XX.

Lemma 2.7.

Proximal path-connectedness implies proximal connectedness as in the same as topological spaces.

Proof.

Let δ\delta be a path-connected proximity on XX. Suppose that (X,δ)(X,\delta) is not proximally connected. Then there exists two nonempty subsets EE, FF in XX such that EF=XE\cup F=X and Eδ¯FE\underline{\delta}F. Since XX is proximally path-connected, there is a pc-map h:[0,1]Xh:[0,1]\rightarrow X with h(0)=Eh(0)=E and h(1)=Fh(1)=F. Consider the subsets h1(E)h^{-1}(E) and h1(F)Ih^{-1}(F)\in I. They are nonempty sets because 0h1(E)0\in h^{-1}(E) and 1h1(F)1\in h^{-1}(F). Their union is [0,1][0,1], and by the proximal continuity of hh, h1(E)δ¯h1(F)h^{-1}(E)\underline{\delta}h^{-1}(F). This contradicts with the fact that [0,1][0,1] is proximally connected. Finally, XX is proximally connected. ∎

Theorem 2.8.

Proximal path-connectedness coincides with proximal connectedness.

Proof.

Given a proximity δ\delta on XX, by Lemma 2.7, it is enough to prove that any connected proximity space is a path-connected proximity space. Suppose that XX is not a path-connected proximity space. Then any map h:([0,1],δ)(X,δ)h:([0,1],\delta^{{}^{\prime}})\rightarrow(X,\delta) with h(0)=xh(0)=x and h(1)=yh(1)=y is not proximally continuous, i.e., if EδFE\delta^{{}^{\prime}}F for all EE, F[0,1]F\in[0,1], then h(E)δ¯h(F)h(E)\underline{\delta}h(F). Take E={0}IE=\{0\}\subset I and F=(0,1]IF=(0,1]\subset I. Since D(E,F)=inf{d(0,z):zF}=0D(E,F)=\inf\{d(0,z):z\in F\}=0, we have that EδFE\delta F. It follows that h(E)={x}h(E)=\{x\} is not near h(F)=X{x}h(F)=X\setminus\{x\}. On the other hand,

h(E)h(F)={x}X{x}=X.\displaystyle h(E)\cup h(F)=\{x\}\cup X\setminus\{x\}=X.

Thus, XX is not proximally connected and this is a contradiction. ∎

2.2. On Descriptive Proximity Spaces

Assume that XX is a nonempty set and xXx\in X. Consider the set Φ={ϕ1,,ϕm}\Phi=\{\phi_{1},\cdots,\phi_{m}\} of maps (generally named as probe functions) ϕj:X\phi_{j}:X\rightarrow\mathbb{R}, j=1,,mj=1,\cdots,m, such that ϕj(x)\phi_{j}(x) denotes a feature value of xx. Let EXE\subset X. Then the set of descriptions of a point ee in EE, denoted by 𝒬(E)\mathcal{Q}(E), is given by the set {Φ(e):eE}\{\Phi(e):e\in E\}, where Φ(e)\Phi(e) (generally called a feature vector for ee) equals (ϕ1(e),,ϕm(e))(\phi_{1}(e),\cdots,\phi_{m}(e)). For EE, FXF\subset X, the binary relation δΦ\delta_{\Phi} is defined by

(1) E′′δΦF𝒬(E)𝒬(F)′′,{}^{\prime\prime}E\delta_{\Phi}F\ \Leftrightarrow\ \mathcal{Q}(E)\cap\mathcal{Q}(F)\neq\emptyset^{\prime\prime},

and EδΦFE\delta_{\Phi}F is read as “EE is descriptively near FF[9, 10, 12]. Also, EδΦ¯FE\underline{\delta_{\Phi}}F is often used to state “EE is descriptively far from FF”. The descriptive intersection of EE and FF and the descriptive union of EE and FF are defined by

EΦF={xEF:Φ(x)𝒬(E)Φ(x)𝒬(F)},\displaystyle E\displaystyle\bigcap_{\Phi}F=\{x\in E\cup F:\Phi(x)\in\mathcal{Q}(E)\ \land\ \Phi(x)\in\mathcal{Q}(F)\},

and

EΦF={xEF:Φ(x)𝒬(E)Φ(x)𝒬(F)},\displaystyle E\displaystyle\bigcup_{\Phi}F=\{x\in E\cup F:\Phi(x)\in\mathcal{Q}(E)\ \vee\ \Phi(x)\in\mathcal{Q}(F)\},

respectively[12].

A binary relation δΦ\delta_{\Phi} defined by (1) [6] satisfies

(f)EδΦFEF,\displaystyle\textbf{(f)}\hskip 28.45274ptE\delta_{\Phi}F\ \ \Rightarrow\ \ E\neq\emptyset\ \land F\neq\emptyset,
(g)EΦFEδΦF,\displaystyle\textbf{(g)}\hskip 28.45274ptE\displaystyle\bigcap_{\Phi}F\neq\emptyset\ \ \Rightarrow\ \ E\delta_{\Phi}F,
(h)EΦFFΦE,\displaystyle\textbf{(h)}\hskip 28.45274ptE\displaystyle\bigcap_{\Phi}F\neq\emptyset\ \ \Rightarrow\ \ F\displaystyle\bigcap_{\Phi}E,
(i)EδΦ(FG)EδΦFEδΦG,\displaystyle\textbf{(i)}\hskip 28.45274ptE\delta_{\Phi}(F\cup G)\ \ \Leftrightarrow\ \ E\delta_{\Phi}F\ \vee\ E\delta_{\Phi}G,
(k)EδΦ¯FGX:EδΦ¯G(XG)δΦ¯F.\displaystyle\textbf{(k)}\hskip 28.45274ptE\underline{\delta_{\Phi}}F\ \ \Rightarrow\ \ \exists G\subset X:E\underline{\delta_{\Phi}}G\ \land\ (X-G)\underline{\delta_{\Phi}}F.

δΦ\delta_{\Phi} is a descriptive nearness relation.

Definition 2.9.

[6] The nearness relation δΦ\delta_{\Phi} for the subsets of XX is said to be an descriptive Efremovic proximity (simply denoted by descriptive EF-proximity or descriptive proximity) if δΦ\delta_{\Phi} satisfies (f)-(k). (X,δΦ)(X,\delta_{\Phi}) is said to be a descriptive EF-proximity (or descriptive proximity) space.

A map h:(X,δΦ)(X,δΦ)h:(X,\delta_{\Phi})\rightarrow(X,\delta_{\Phi}^{{}^{\prime}}) is called descriptive proximally continuous provided that “EδΦFh(E)δΦh(F)E\delta_{\Phi}F\ \Rightarrow\ h(E)\delta_{\Phi}^{{}^{\prime}}h(F)” for EE, FXF\subset X[13, 14]. We denote a descriptive proximally continuous map by “dpc-map”. Let δΦ\delta_{\Phi} be a descriptive proximity on XX, and EXE\subset X a subset. Then a descriptive subspace proximity δΦE\delta_{\Phi}^{E} is defined on the subsets of EE as follows:

E1′′δΦE2E1δΦEE2′′{}^{\prime\prime}E_{1}\delta_{\Phi}E_{2}\ \Leftrightarrow\ E_{1}\delta_{\Phi}^{E}E_{2}^{\prime\prime}

for E1E_{1}, E2EE_{2}\subset E. Given a descriptive proximity δΦ\delta_{\Phi} on XX, a descriptive subspace proximity (E,δΦE)(E,\delta_{\Phi}^{E}), and the inclusion j:(E,δΦE)(X,δΦ)j:(E,\delta_{\Phi}^{E})\rightarrow(X,\delta_{\Phi}), a dpc-map k:(X,δΦ)(E,δΦE)k:(X,\delta_{\Phi})\rightarrow(E,\delta_{\Phi}^{E}) is called a descriptive proximal retraction if kj=1Ek\circ j=1_{E}.

Lemma 2.10.

[14](Gluing Lemma) Assume that f1:(X,δΦ1)(Y,δΦ2)f_{1}:(X^{{}^{\prime}},\delta_{\Phi_{1}}^{{}^{\prime}})\rightarrow(Y^{{}^{\prime}},\delta_{\Phi_{2}}^{{}^{\prime}}) and f2:(X′′,δΦ1′′)(Y,δΦ2)f_{2}:(X^{{}^{\prime\prime}},\delta_{\Phi_{1}}^{{}^{\prime\prime}})\rightarrow(Y^{{}^{\prime}},\delta_{\Phi_{2}}^{{}^{\prime}}) are two dpc-maps with the property that they agree on the intersection of XX^{{}^{\prime}} and X′′X^{{}^{\prime\prime}}. Then the map f1f2f_{1}\cup f_{2} from (XX′′,δΦ)(X^{{}^{\prime}}\cup X^{{}^{\prime\prime}},\delta_{\Phi}) to (Y,δΦ2)(Y^{{}^{\prime}},\delta_{\Phi_{2}}^{{}^{\prime}}), defined by f1f2(s)={f1(s),sXf2(s),sX′′f_{1}\cup f_{2}(s)=\begin{cases}f_{1}(s),&s\in X^{{}^{\prime}}\\ f_{2}(s),&s\in X^{{}^{\prime\prime}}\end{cases} for any sXX′′s\in X^{{}^{\prime}}\cup X^{{}^{\prime\prime}}, is a dpc-map.

hh is a descriptive proximity isomorphism if hh is a bijection and each of hh and h1h^{-1} is dpc-map[5]. Hence, (X,δΦ)(X,\delta_{\Phi}) and (X,δΦ)(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}}) are called descriptive proximally isomorphic spaces. A descriptive proximity relation on the cartesian product of descriptive proximity spaces is defined as follows[3]: Assume that δΦ\delta_{\Phi} and δΦ\delta_{\Phi}^{{}^{\prime}} are any descriptive proximities on XX and XX^{{}^{\prime}}, respectively. EδΦFE\delta_{\Phi}F and EδΦFE^{{}^{\prime}}\delta_{\Phi}^{{}^{\prime}}F^{{}^{\prime}} implies that E×EE\times E^{{}^{\prime}} is descriptively near F×FF\times F^{{}^{\prime}}, where E×EE\times E^{{}^{\prime}} and F×FF\times F^{{}^{\prime}} are any subsets of X×XX\times X^{{}^{\prime}}.

Definition 2.11.

[14] Let h1h_{1}, h2:(X,δΦ)(X,δΦ)h_{2}:(X,\delta_{\Phi})\rightarrow(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}}) be any map. Then h1h_{1} and h2h_{2} are said to be descriptive proximally homotopic maps provided that there exists a dpc-map G:X×IXG:X\times I\rightarrow X^{{}^{\prime}} with G(x,0)=h1(x)G(x,0)=h_{1}(x) and G(x,1)=h2(x)G(x,1)=h_{2}(x).

In Definition 2.11, GG is a descriptive proximal homotopy between h1h_{1} and h2h_{2}. We simply denote a descriptive proximal homotopy by “dprox-hom”. Given a descriptive proximity δΦ\delta_{\Phi} on XX and a subset FXF\subset X, FF is said to be a δΦ\delta_{\Phi}-neighborhood of EE, denoted by EδΦFE\ll_{\delta_{\Phi}}F, if EδΦ¯(XF)E\underline{\delta_{\Phi}}(X-F)[13].

Theorem 2.12.

[5] Let EjδΦFjE_{j}\ll_{\delta_{\Phi}}F_{j} for j=1,,mj=1,\cdots,m. Then

j=1mEjδΦj=1mFjandj=1mEjδΦj=1mFj.\displaystyle\displaystyle\bigcap_{j=1}^{m}E_{j}\ll_{\delta_{\Phi}}\displaystyle\bigcap_{j=1}^{m}F_{j}\ \ \ \text{and}\ \ \ \displaystyle\bigcup_{j=1}^{m}E_{j}\ll_{\delta_{\Phi}}\displaystyle\bigcup_{j=1}^{m}F_{j}.
Definition 2.13.

[14] Let x1x_{1} and x2x_{2} be any two elements in XX with a descriptive proximity δΦ\delta_{\Phi}. Then a descriptive proximal path from x1x_{1} to x2x_{2} in XX is a dpc-map hh from I=[0,1]I=[0,1] to XX for which h(0)=x1h(0)=x_{1} and h(1)=x2h(1)=x_{2}.

In Definition 2.13, the fact h:IXh:I\rightarrow X is descriptive proximally continuous means that “D(E,F)=0h(E)δΦh(F)D(E,F)=0\ \Rightarrow\ h(E)\delta_{\Phi}h(F)” for EE, FIF\in I. A descriptive proximity space (X,δΦ)(X,\delta_{\Phi}) is connected if and only if for all nonempty EE, F𝒫(X)F\in\mathcal{P}(X), EF=XE\cup F=X implies that EδΦFE\delta_{\Phi}F[7]. A descriptive proximity space (X,δΦ)(X,\delta_{\Phi}) is path-connected if, for any points x1x_{1} and x2x_{2} in XX, there exists a descriptive proximal path from x1x_{1} to x2x_{2} in XX.

Theorem 2.14.

In a descriptive proximity space, path-connectedness coincides with connectedness.

Proof.

Follow the method in the proof of Theorem 2.8. ∎

3. Homotopy Theory on Proximity Spaces

This section, one of the main parts (Section 3 and Section 4) of the paper, examines the projection of the homotopy theory elements in parallel with the proximity spaces. First, we start with the notion of proximal mapping spaces. Then we have proximal covering spaces. The last two parts are related to proximal fibrations and its dual notion of proximal cofibrations. Results on these four topics that we believe will be relevant to future proximity space research are presented.

3.1. Proximal Mapping Spaces

The work of mapping spaces in nearness theory starts with [8] and is still open to improvement. Note that the study of discrete invariants of function spaces is essentially homotopy theory in algebraic topology, and recall that depending on the nature of the spaces, it may be useful to attempt to impose a topology on the space of continuous functions from one topological space to another. One of the best-known examples of this is the compact-open topology.

Definition 3.1.

Let δ1\delta_{1} and δ2\delta_{2} be two proximities on XX and YY, respectively. The proximal mapping space YXY^{X} is defined as {α:XY|αis a pc-map}\{\alpha:X\rightarrow Y\ |\ \alpha\ \text{is a pc-map}\} having the following proximity relation δ\delta on itself: Let EE, FXF\subset X and {αi}iI\{\alpha_{i}\}_{i\in I} and {βj}jJ\{\beta_{j}\}_{j\in J} be any subsets of pc-maps in YXY^{X}. We say that {αi}iIδ{βj}jJ\{\alpha_{i}\}_{i\in I}\delta\{\beta_{j}\}_{j\in J} if the fact Eδ1FE\delta_{1}F implies that αi(E)δ2βj(F)\alpha_{i}(E)\delta_{2}\beta_{j}(F).

Refer to caption
Figure 3.1. The picture represented by X={a,b,c,d,e,f,g,h}X=\{a,b,c,d,e,f,g,h\}.
Example 3.2.

Consider the set X={a,b,c,d,e,f,g,h}X=\{a,b,c,d,e,f,g,h\} of cells in Figure 3.1 with the proximity δ\delta. Define three proximal paths α1\alpha_{1}, α2\alpha_{2}, and α3XI\alpha_{3}\in X^{I} by

α1:abcdefgh,\displaystyle\alpha_{1}:a\mapsto b\mapsto c\mapsto d\mapsto e\mapsto f\mapsto g\mapsto h,
α2:habcdefg,\displaystyle\alpha_{2}:h\mapsto a\mapsto b\mapsto c\mapsto d\mapsto e\mapsto f\mapsto g,
α3:ahgfedcb.\displaystyle\alpha_{3}:a\mapsto h\mapsto g\mapsto f\mapsto e\mapsto d\mapsto c\mapsto b.

For all tIt\in I, α1(t)δα2(t)\alpha_{1}(t)\delta\alpha_{2}(t). This means that α1\alpha_{1} is near α2\alpha_{2}. On the other hand, for t[2/8,3/8]t\in[2/8,3/8], we have that α1(t)=c\alpha_{1}(t)=c and α3(t)=g\alpha_{3}(t)=g, that is, α1\alpha_{1} and α3\alpha_{3} are not near in XX.

Definition 3.3.

For the proximal continuity of a map H:(X,δ1)(ZY,δ)H:(X,\delta_{1})\rightarrow(Z^{Y},\delta^{{}^{\prime}}), we say that the fact Eδ1FE\delta_{1}F implies that H(E)δH(F)H(E)\delta^{{}^{\prime}}H(F) for any subsets EE, FXF\subset X.

Proposition 3.4.

Let δ1\delta_{1}, δ2\delta_{2}, and δ3\delta_{3} be any proximities on XX, YY, and ZZ, respectively. Then the map G:(X×Y,δ′′)(Z,δ3)G:(X\times Y,\delta^{{}^{\prime\prime}})\rightarrow(Z,\delta_{3}) is pc-map if and only if the map H:(X,δ1)(ZY,δ)H:(X,\delta_{1})\rightarrow(Z^{Y},\delta^{{}^{\prime}}) defined by H(E)(F):=G(E×F)H(E)(F):=G(E\times F) is pc-map for EXE\subset X and FYF\subset Y.

Proof.

Assume that E1δ1F1E_{1}\delta_{1}F_{1} for E1E_{1}, F1XF_{1}\subset X. If E2δ2F2E_{2}\delta_{2}F_{2} for E2E_{2}, F2YF_{2}\subset Y, then we find (E1×E2)δ′′(F1×F2)(E_{1}\times E_{2})\delta^{{}^{\prime\prime}}(F_{1}\times F_{2}). Since GG is a pc-map, we get G(E1×E2)δ3G(F1×F2)G(E_{1}\times E_{2})\delta_{3}G(F_{1}\times F_{2}). It follows that H(E1)(E2)δ3H(F1)(F2)H(E_{1})(E_{2})\delta_{3}H(F_{1})(F_{2}). This shows that H(E1)δH(F1)H(E_{1})\delta^{{}^{\prime}}H(F_{1}), i.e., HH is a pc-map. Conversely, assume that (E1×E2)δ′′(F1×F2)(E_{1}\times E_{2})\delta^{{}^{\prime\prime}}(F_{1}\times F_{2}). Then we get E1δ1F1E_{1}\delta_{1}F_{1} in XX and E2δ2F2E_{2}\delta_{2}F_{2} in YY. Since HH is a pc-map, we get H(E1)δH(F1)H(E_{1})\delta^{{}^{\prime}}H(F_{1}). So, we have that H(E1)(E2)δ3H(F1)(F2)H(E_{1})(E_{2})\delta_{3}H(F_{1})(F_{2}). This leads to the fact that G(E1×E2)δ′′G(F1×F2)G(E_{1}\times E_{2})\delta^{{}^{\prime\prime}}G(F_{1}\times F_{2}), namely that, GG is a pc-map. ∎

Theorem 3.5.

Let δ1\delta_{1}, δ2\delta_{2}, and δ3\delta_{3} be any proximities on XX, YY, and ZZ, respectively. Then (ZX×Y,δ4)(Z^{X\times Y},\delta_{4}) and ((ZY)X,δ5)((Z^{Y})^{X},\delta_{5}) are proximally isomorphic spaces.

Proof.

Define a bijective map f:ZX×Y(ZY)Xf:Z^{X\times Y}\rightarrow(Z^{Y})^{X} by f(G)=Hf(G)=H. For any pc-maps GG, GZX×YG^{{}^{\prime}}\subset Z^{X\times Y} such that Gδ4GG\delta_{4}G^{{}^{\prime}}, we have that f(G)δ5f(G)f(G)\delta_{5}f(G^{{}^{\prime}}). Indeed, for E1×E2E_{1}\times E_{2}, F1×F2X×YF_{1}\times F_{2}\subset X\times Y, we have that G(E1×E2)δ3G(F1×F2)G(E_{1}\times E_{2})\delta_{3}G(F_{1}\times F_{2}). This means that H(E1)(E2)δ3H(F1)(F2)H(E_{1})(E_{2})\delta_{3}H(F_{1})(F_{2}). Another saying, we find Hδ5HH\delta_{5}H^{{}^{\prime}}. Therefore, ff is a pc-map. For the proximal continuity of f1f^{-1}, assume that Hδ5HH\delta_{5}H^{{}^{\prime}}. Then we have that H(E1)H(E_{1}) and H(F1)H^{{}^{\prime}}(F_{1}) are near in ZYZ^{Y} for E1E_{1}, F1XF_{1}\subset X. If E2δ2F2E_{2}\delta_{2}F_{2} in YY, then we have that H(E1)(E2)δ3H(F1)(F2)H(E_{1})(E_{2})\delta_{3}H^{{}^{\prime}}(F_{1})(F_{2}). It follows that G(E1×E2)δ3G(F1×F2)G(E_{1}\times E_{2})\delta_{3}G^{{}^{\prime}}(F_{1}\times F_{2}). Thus, we obtain that Gδ4GG\delta_{4}G^{{}^{\prime}}, which means that f1(H)δ4f1(H)f^{-1}(H)\delta_{4}f^{-1}(H^{{}^{\prime}}). Finally, ff is a proximity isomorphism. ∎

Theorem 3.6.

Let δ1\delta_{1}, δ2\delta_{2}, and δ3\delta_{3} be any proximities on XX, YY, and ZZ, respectively. Then ((Y×Z)X,δ4)((Y\times Z)^{X},\delta_{4}) and (YX×ZX,δ5)(Y^{X}\times Z^{X},\delta_{5}) are proximally isomorphic spaces.

Proof.

The proximal isomorphism is given by the map

f:((Y×Z)X,δ4)(YX×ZX,δ5)f:((Y\times Z)^{X},\delta_{4})\rightarrow(Y^{X}\times Z^{X},\delta_{5})

with f(α)=(π1α,π2α)f(\alpha)=(\pi_{1}\circ\alpha,\pi_{2}\circ\alpha), where π1\pi_{1} and π2\pi_{2} are the projection maps from Y×ZY\times Z to the respective spaces. For any {αi}iI\{\alpha_{i}\}_{i\in I}, {βj}jJ(Y×Z)X\{\beta_{j}\}_{j\in J}\subset(Y\times Z)^{X} such that {αi}iI\{\alpha_{i}\}_{i\in I} is near {βj}jJ\{\beta_{j}\}_{j\in J}, we obtain that πk{αi}iI\pi_{k}\circ\{\alpha_{i}\}_{i\in I} is near πk{βi}jJ\pi_{k}\circ\{\beta_{i}\}_{j\in J} for each k{1,2}k\in\{1,2\}. Therefore, we have that (π1{αi}iI,π2{αi}iI)(\pi_{1}\circ\{\alpha_{i}\}_{i\in I},\pi_{2}\circ\{\alpha_{i}\}_{i\in I}) is near (π1{βj}jJ,π2{βj}jJ)(\pi_{1}\circ\{\beta_{j}\}_{j\in J},\pi_{2}\circ\{\beta_{j}\}_{j\in J}). Thus, f({αi}iI)f(\{\alpha_{i}\}_{i\in I}) is near f({βj}jJ)f(\{\beta_{j}\}_{j\in J}), i.e., ff is a pc-map. For the pc-map

g:(YX×ZX,δ5)((Y×Z)X,δ4)g:(Y^{X}\times Z^{X},\delta_{5})\rightarrow((Y\times Z)^{X},\delta_{4})

with g(β,γ)=(β×γ)ΔXg(\beta,\gamma)=(\beta\times\gamma)\circ\Delta_{X}, where ΔX:(X,δ1)(X2,δ1)\Delta_{X}:(X,\delta_{1})\rightarrow(X^{2},\delta_{1}^{{}^{\prime}}) is a diagonal map of proximity spaces on XX, we have that gfg\circ f and fgf\circ g are identity maps on respective proximity spaces (Y×Z)X(Y\times Z)^{X} and YX×ZXY^{X}\times Z^{X}. Consequently, ((Y×Z)X,δ4)((Y\times Z)^{X},\delta_{4}) and (YX×ZX,δ5)(Y^{X}\times Z^{X},\delta_{5}) are proximally isomorphic spaces. ∎

Definition 3.7.

Let δ1\delta_{1} and δ2\delta_{2} be any proximities on XX and YY, respectively. Then the proximal evaluation map

eX,Y:(YX×X,δ)(Y,δ2)e_{X,Y}:(Y^{X}\times X,\delta)\rightarrow(Y,\delta_{2})

is defined by e(α,x)=α(x)e(\alpha,x)=\alpha(x).

To show that the evaluation map eX,Ye_{X,Y} is a pc-map, we first assume that ({αi}iI×E)δ({βj}jJ×F)(\{\alpha_{i}\}_{i\in I}\times E)\delta(\{\beta_{j}\}_{j\in J}\times F) in YX×XY^{X}\times X. This means that {αi}iIδ{βj}jJ\{\alpha_{i}\}_{i\in I}\delta^{{}^{\prime}}\{\beta_{j}\}_{j\in J} for a proximity relation δ\delta^{{}^{\prime}} on YXY^{X} and Eδ1FE\delta_{1}F in XX. It follows that αi(E)δ2βj(F)\alpha_{i}(E)\delta_{2}\beta_{j}(F) in YY for any iIi\in I and jJj\in J. Finally, we conclude that

eX,Y({αi}iI×E)δ2eX,Y({βj}jJ×F).\displaystyle e_{X,Y}(\{\alpha_{i}\}_{i\in I}\times E)\delta_{2}e_{X,Y}(\{\beta_{j}\}_{j\in J}\times F).
Example 3.8.

Consider the proximal evaluation map eI,X:(XI×I,δ)(X,δ1)e_{I,X}:(X^{I}\times I,\delta)\rightarrow(X,\delta_{1}). Since XI×{0}X^{I}\times\{0\} is proximally isomorphic to XIX^{I} by the map (α,0)α(0)(\alpha,0)\mapsto\alpha(0), the restriction

eI,X0=eI,X|(XI×{0}):(XI,δ)(X,δ1),e^{0}_{I,X}=e_{I,X}|_{(X^{I}\times\{0\})}:(X^{I},\delta^{{}^{\prime}})\rightarrow(X,\delta_{1}),

defined by eI,X0(α)=α(0)e^{0}_{I,X}(\alpha)=\alpha(0), is a pc-map.

Example 3.9.

Let eI,X×X:((X×X)I×I,δ)(X,δ1)e_{I,X\times X}:((X\times X)^{I}\times I,\delta)\rightarrow(X,\delta_{1}) be the proximal evaluation map. By Theorem 3.6, the restriction

eI,X×X0=eI,X×X|(XI×{0}):(XI,δ)(X×X,δ),e^{0}_{I,X\times X}=e_{I,X\times X}|_{(X^{I}\times\{0\})}:(X^{I},\delta^{{}^{\prime}})\rightarrow(X\times X,\delta^{{}^{\prime}}),

defined by eI,X×X0(α)=(α(0),α(1))e^{0}_{I,X\times X}(\alpha)=(\alpha(0),\alpha(1)), is a pc-map.

Note that, in topological spaces, the map XIX×XX^{I}\rightarrow X\times X, α(α(0),α(1))\alpha\mapsto(\alpha(0),\alpha(1)), is the path fibration. Similarly, the map XIXX^{I}\rightarrow X, αα(0)\alpha\mapsto\alpha(0), is the path fibration with a fixed initial point at t=0t=0.

3.2. Proximal Covering Spaces

A covering space of a topological space and the fundamental group are tightly related. One can categorize all the covering spaces of a topological space using the subgroups of its fundamental group. Covering spaces are not only useful in algebraic topology, but also in complex dynamics, geometric group theory, and the theory of Lie groups.

Definition 3.10.

A surjective and pc-map p:(X,δ)(X,δ)p:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) is a proximal covering map if the following hold:

  • Let {x}X\{x^{{}^{\prime}}\}\subseteq X^{{}^{\prime}} be any subset with {x}δY\{x^{{}^{\prime}}\}\ll_{\delta^{{}^{\prime}}}Y^{{}^{\prime}}. Then there is an index set II satisfying that

    p1(Y)=iIYi\displaystyle p^{-1}(Y^{{}^{\prime}})=\displaystyle\bigcup_{i\in I}Y_{i}

    with ViδYiV_{i}\ll_{\delta}Y_{i}, where Vip1({x})V_{i}\in p^{-1}(\{x^{{}^{\prime}}\}) for each iIi\in I.

  • YiYjY_{i}\neq Y_{j} when iji\neq j for ii, jIj\in I.

  • p|Yi:YiYp|_{Y_{i}}:Y_{i}\rightarrow Y^{{}^{\prime}} is a proximal isomorphism for every iIi\in I.

In Definition 3.10, (X,δ)(X,\delta) is called a proximal covering space of (X,δ)(X^{{}^{\prime}},\delta^{{}^{\prime}}). For iIi\in I, YiY_{i} is said to be a proximal sheet. For any xXx^{{}^{\prime}}\in X^{{}^{\prime}}, p1({x})p^{-1}(\{x^{{}^{\prime}}\}) is called a proximal fiber of xx^{{}^{\prime}}. The map p|Yi:YiYp|_{Y_{i}}:Y_{i}\rightarrow Y^{{}^{\prime}} is a proximal isomorphism if the map p:(X,δ)(X,δ)p:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) is a proximal isomorphism. However, the converse is not generally true. Given any proximity δ\delta on XX, it is obvious that the identity map on XX is always a proximal covering map.

Refer to caption
Figure 3.2. A map pp from {a1,a2,a3,a4}{b1,b2,b3,b4}{c1,c2,c3,c4}\{a_{1},a_{2},a_{3},a_{4}\}\cup\{b_{1},b_{2},b_{3},b_{4}\}\cup\{c_{1},c_{2},c_{3},c_{4}\} to {d1,d2,d3,d4}\{d_{1},d_{2},d_{3},d_{4}\} defined by p(ai)=p(bi)=p(ci)=dip(a_{i})=p(b_{i})=p(c_{i})=d_{i} for any i=1,2,3,4i=1,2,3,4.
Example 3.11.

Assume that X={a1,a2,a3,a4}{b1,b2,b3,b4}{c1,c2,c3,c4}X=\{a_{1},a_{2},a_{3},a_{4}\}\cup\{b_{1},b_{2},b_{3},b_{4}\}\cup\{c_{1},c_{2},c_{3},c_{4}\} and X={d1,d2,d3,d4}X^{{}^{\prime}}=\{d_{1},d_{2},d_{3},d_{4}\} are two proximity spaces such that p:(X,δ)(X,δ)p:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) is a surjective and pc-map defined by p(ai)=p(bi)=p(ci)=dip(a_{i})=p(b_{i})=p(c_{i})=d_{i} for each i=1,2,3,4i=1,2,3,4 (see Figure 3.2). Let {d1}X\{d_{1}\}\subset X^{{}^{\prime}} and Y={d1,d2,d4}Y^{{}^{\prime}}=\{d_{1},d_{2},d_{4}\} a proximal δ\delta^{{}^{\prime}}-neighborhood of {d1}\{d_{1}\}. For V1={a1}V_{1}=\{a_{1}\}, V2={b1}V_{2}=\{b_{1}\}, and V3={c1}V_{3}=\{c_{1}\}, we have p1(Y)=i=13Yip^{-1}(Y^{{}^{\prime}})=\displaystyle\bigcup_{i=1}^{3}Y_{i}, where Y1={a1,a2,a4}Y_{1}=\{a_{1},a_{2},a_{4}\}, V2={b1,b2,b4}V_{2}=\{b_{1},b_{2},b_{4}\}, and V3={c1,c2,c4}V_{3}=\{c_{1},c_{2},c_{4}\}. Note that V1δ¯(XY1)V_{1}\underline{\delta}(X-Y_{1}), V2δ¯(XY2)V_{2}\underline{\delta}(X-Y_{2}), and V3δ¯(XY3)V_{3}\underline{\delta}(X-Y_{3}), i.e., YiY_{i} is a proximal δ\delta-neighborhood of ViV_{i} for each i{1,2,3}i\in\{1,2,3\}. Moreover, for ii, j{1,2,3}j\in\{1,2,3\} with iji\neq j, we have that YiY_{i} is not YjY_{j}, and p|Y1:{a1,a2,a4}{d1,d2,d4}p|_{Y_{1}}:\{a_{1},a_{2},a_{4}\}\rightarrow\{d_{1},d_{2},d_{4}\}, p|Y2:{b1,b2,b4}{d1,d2,d4}p|_{Y_{2}}:\{b_{1},b_{2},b_{4}\}\rightarrow\{d_{1},d_{2},d_{4}\}, and p|Y3:{c1,c2,c4}{d1,d2,d4}p|_{Y_{3}}:\{c_{1},c_{2},c_{4}\}\rightarrow\{d_{1},d_{2},d_{4}\} are proximal isomorphisms. For other points d2d_{2}, d3d_{3}, and d4d_{4}, a similar process is done. This shows that pp is a proximal covering map.

Example 3.12.

Given a proximity δ\delta on XX, consider the surjective and pc-map p:(X×{0,1,2},δ)(X,δ)p:(X\times\{0,1,2\cdots\},\delta^{{}^{\prime}})\rightarrow(X,\delta) with p(x,t)=xp(x,t)=x. For a proximal δ\delta-neighborhood YY of any subset {x}X\{x\}\subset X, we have that

p1(Y)=Y×+X×{0,1,2}p^{-1}(Y)=Y\times\mathbb{Z}^{+}\subset X\times\{0,1,2\cdots\}

for a proximal δ\delta^{{}^{\prime}}-neighborhood YY^{{}^{\prime}} of Y×+Y\times\mathbb{Z}^{+}. Moreover, p|Y×+:Y×+Yp|_{Y\times\mathbb{Z}^{+}}:Y\times\mathbb{Z}^{+}\rightarrow Y, p|Y×+(x,t)=xp|_{Y\times\mathbb{Z}^{+}}(x,t)=x, is a proximal isomorphism. Thus, pp is a proximal covering map.

Proposition 3.13.

Any proximal isomorphism is a proximal covering map.

Proof.

Let p:(X,δ)(X,δ)p:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) be a proximal isomorphism. Then pp is a pc-map. Lemma 4.1 of [6] says that pp is continuous with respect to compatible topologies. Therefore, we get p1(Y)=YXp^{-1}(Y^{{}^{\prime}})=Y\subset X for an open neighborhood YY^{{}^{\prime}} of any subset {x}X\{x^{{}^{\prime}}\}\subseteq X^{{}^{\prime}}. Combining with the fact that a proximal neighborhood of a set is also a neighborhood, we conclude that p1(Y)=Yp^{-1}(Y^{{}^{\prime}})=Y for a proximal δ\delta^{{}^{\prime}}-neighborhood YY^{{}^{\prime}} of {x}\{x^{{}^{\prime}}\} in XX^{{}^{\prime}} and a proximal δ\delta-neighborhood YY of VV, where Vp1({x})V\in p^{-1}(\{x^{{}^{\prime}}\}) in XX. Furthermore, p|Y:YYp|_{Y}:Y\rightarrow Y^{{}^{\prime}} is an isomorphism of proximity spaces because pp is an isomorphism of proximity spaces. Finally, pp is a proximal covering map. ∎

The following diagram illustrates two ways to prove that any proximal isomorphism p:(X,δ)(X,δ)p:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) is a covering map between respective compatible topologies on both (X,δ)(X,\delta) and (X,δ)(X^{{}^{\prime}},\delta^{{}^{\prime}}):

proximal isomorphismproximal covering maphomeomorphismcovering map.\textstyle{\text{covering map}.}
Theorem 3.14.

The cartesian product of two proximal covering maps is a proximal covering map.

Proof.

Let p:(X,δ1)(X,δ1)p:(X,\delta_{1})\rightarrow(X^{{}^{\prime}},\delta_{1}^{{}^{\prime}}) and q:(Y,δ2)(Y,δ2)q:(Y,\delta_{2})\rightarrow(Y^{{}^{\prime}},\delta_{2}^{{}^{\prime}}) be two proximal covering maps. Then for a proximal δ1\delta_{1}^{{}^{\prime}}-neighborhood M1M_{1}^{{}^{\prime}} of {x1}X\{x_{1}^{{}^{\prime}}\}\subset X^{{}^{\prime}}, we have that

p1(M1)=iIMi\displaystyle p^{-1}(M_{1}^{{}^{\prime}})=\displaystyle\bigcup_{i\in I}M_{i}

for a proximal δ1\delta_{1}-neighborhood M1M_{1} of ViV_{i}, where Vip1({x1})V_{i}\in p^{-1}(\{x_{1}^{{}^{\prime}}\}). We also have that MiMkM_{i}\neq M_{k} with any kIk\in I when iki\neq k. Similarly, for a proximal δ2\delta_{2}^{{}^{\prime}}-neighborhood N2N_{2}^{{}^{\prime}} of {x2}Y\{x_{2}^{{}^{\prime}}\}\subset Y^{{}^{\prime}}, we have that

q1(N2)=jJNj\displaystyle q^{-1}(N_{2}^{{}^{\prime}})=\displaystyle\bigcup_{j\in J}N_{j}

for a proximal δ2\delta_{2}-neighborhood NjN_{j} of WjW_{j}, where Wjq1({x2})W_{j}\in q^{-1}(\{x_{2}^{{}^{\prime}}\}). Also, we have that NiNlN_{i}\neq N_{l} with any lJl\in J when jlj\neq l. For a proximal neighborhood M1×N2M_{1}^{{}^{\prime}}\times N_{2}^{{}^{\prime}} of {x1}×{x2}X×Y\{x_{1}^{{}^{\prime}}\}\times\{x_{2}^{{}^{\prime}}\}\subset X^{{}^{\prime}}\times Y^{{}^{\prime}}, we get

(p×q)1(M1×N2)=p1(M1)×q1(N2)=iIMi×jJNj=iIjJ(Mi×Nj).(p\times q)^{-1}(M_{1}^{{}^{\prime}}\times N_{2}^{{}^{\prime}})=p^{-1}(M_{1}^{{}^{\prime}})\times q^{-1}(N_{2}^{{}^{\prime}})=\displaystyle\bigcup_{i\in I}M_{i}\times\displaystyle\bigcup_{j\in J}N_{j}=\displaystyle\bigcup_{\begin{subarray}{c}{i\in I}\\ {j\in J}\end{subarray}}(M_{i}\times N_{j}).

It is clear that Mi×NjMk×NlM_{i}\times N_{j}\neq M_{k}\times N_{l} when (i,j)(k,l)(i,j)\neq(k,l) for any ii, kIk\in I and jj, lJl\in J. Moreover, since p|Mi:MiM1p|_{M_{i}}:M_{i}\rightarrow M_{1}^{{}^{\prime}} and q|Nj:NjN2q|_{N_{j}}:N_{j}\rightarrow N_{2}^{{}^{\prime}} are proximal isomorphisms, the map (p×q)|Mi×Nj:Mi×NjM1×N2(p\times q)|_{M_{i}\times N_{j}}:M_{i}\times N_{j}\rightarrow M_{1}^{{}^{\prime}}\times N_{2}^{{}^{\prime}} is a proximal isomorphism. Consequently, p×q:X×YX×Yp\times q:X\times Y\rightarrow X^{{}^{\prime}}\times Y^{{}^{\prime}} is a proximal covering map. ∎

3.3. Proximal Fibrations

Some topological problems can be conceptualized as lifting or extension problems. In the homotopy-theoretic viewpoint, fibrations and cofibrations deal with them, respectively (see Section 3.4 for the detail of cofibrations). Postnikov systems, spectral sequences, and obstruction theory, which are important tools constructed on homotopy theory, involve fibrations. On the other hand, the notion of proximal fibration of proximity spaces is first mentioned in [17], and we extend this with useful properties in proximity cases.

Definition 3.15.

A pc-map p:(X,δ)(X,δ)p:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) is said to have the proximal homotopy lifting property (PHLP) with respect to a proximity space (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) if for an inclusion map i0:(X′′,δ′′)(X′′×I,δ1)i_{0}:(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})\rightarrow(X^{{}^{\prime\prime}}\times I,\delta_{1}), i0(x′′)=(x′′,0)i_{0}(x^{{}^{\prime\prime}})=(x^{{}^{\prime\prime}},0), for every pc-map k:(X′′,δ′′)(X,δ)k:(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})\rightarrow(X,\delta), and prox-hom G:(X′′×I,δ1)(X,δ)G:(X^{{}^{\prime\prime}}\times I,\delta_{1})\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) with pk=Gi0p\circ k=G\circ i_{0}, then there exists a prox-hom G:(X′′×I,δ1)(X,δ)G^{{}^{\prime}}:(X^{{}^{\prime\prime}}\times I,\delta_{1})\rightarrow(X,\delta) for which G(x′′,0)=k(x′′)G^{{}^{\prime}}(x^{{}^{\prime\prime}},0)=k(x^{{}^{\prime\prime}}) and pG(x′′,t)=G(x′′,t)p\circ G^{{}^{\prime}}(x^{{}^{\prime\prime}},t)=G(x^{{}^{\prime\prime}},t).

X′′\textstyle{X^{{}^{\prime\prime}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}k\scriptstyle{k}i0\scriptstyle{i_{0}}X\textstyle{X\ignorespaces\ignorespaces\ignorespaces\ignorespaces}p\scriptstyle{p}X′′×I\textstyle{X^{\prime\prime}\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}G\scriptstyle{G}G\scriptstyle{G^{{}^{\prime}}}X.\textstyle{X^{{}^{\prime}}.}
Definition 3.16.

A pc-map p:(X,δ)(X,δ)p:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) is said to be a proximal fibration if it has the PHLP for any proximity space (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}).

Example 3.17.

For any proximity spaces (X,δ)(X,\delta) and (X,δ)(X^{{}^{\prime}},\delta^{{}^{\prime}}), we shall show that the projection map π1:(X×X,δ2)(X,δ)\pi_{1}:(X\times X^{{}^{\prime}},\delta_{2})\rightarrow(X,\delta) onto the first factor is a proximal fibration. Consider the diagram

X′′\textstyle{X^{{}^{\prime\prime}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}(kX,kX)\scriptstyle{(k_{X},k_{X^{{}^{\prime}}})}i0\scriptstyle{i_{0}}X×X\textstyle{X\times X^{{}^{\prime}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces}π1\scriptstyle{\pi_{1}}X′′×I\textstyle{X^{\prime\prime}\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}G\scriptstyle{G}G\scriptstyle{G^{{}^{\prime}}}X\textstyle{X^{{}^{\prime}}}

with π1(kX,kX)=Gi0\pi_{1}\circ(k_{X},k_{X^{{}^{\prime}}})=G\circ i_{0}. Then there is a map G:(X′′×I,δ1)(X×X,δ2)G^{{}^{\prime}}:(X^{{}^{\prime\prime}}\times I,\delta_{1})\rightarrow(X\times X^{{}^{\prime}},\delta_{2}) defined by G=(G,F)G^{{}^{\prime}}=(G,F), where F:(X′′×I,δ1)(X,δ)F:(X^{{}^{\prime\prime}}\times I,\delta_{1})\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) is the composition of the first projection map (X′′×I,δ1)(X′′,δ′′)(X^{{}^{\prime\prime}}\times I,\delta_{1})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) and kXk_{X^{{}^{\prime}}}. Since kXk_{X^{{}^{\prime}}} and the first projection map are pc-maps, it follows that FF is a pc-map. Moreover, we get F(x′′,0)=F(x′′,1)=kX(x′′)F(x^{{}^{\prime\prime}},0)=F(x^{{}^{\prime\prime}},1)=k_{X^{{}^{\prime}}}(x^{{}^{\prime\prime}}), which means that FF is a (constant) prox-hom. Combining this result with the fact that HH is a prox-hom, we have that GG^{{}^{\prime}} is a prox-hom. Moreover, we get

Gi(x′′)\displaystyle G^{{}^{\prime}}\circ i(x^{{}^{\prime\prime}}) =\displaystyle= G(x′′,0)=(G(x′′,0),F(x′′,0))=(kX(x′′),kX(x′′))\displaystyle G^{{}^{\prime}}(x^{{}^{\prime\prime}},0)=(G(x^{{}^{\prime\prime}},0),F(x^{{}^{\prime\prime}},0))=(k_{X}(x^{{}^{\prime\prime}}),k_{X^{{}^{\prime}}}(x^{{}^{\prime\prime}}))
=\displaystyle= (kX,kX)(x′′),\displaystyle(k_{X},k_{X^{{}^{\prime}}})(x^{{}^{\prime\prime}}),

and

π1G(x′′,t)=π1(G(x′′,t),F(x′′,t))=G(x′′,t).\displaystyle\pi_{1}\circ G^{{}^{\prime}}(x^{{}^{\prime\prime}},t)=\pi_{1}(G(x^{{}^{\prime\prime}},t),F(x^{{}^{\prime\prime}},t))=G(x^{{}^{\prime\prime}},t).

This shows that π1\pi_{1} is a proximal fibration.

Example 3.18.

Let c:(X,δ)({x0},δ0)c:(X,\delta)\rightarrow(\{x_{0}\},\delta_{0}) be the constant map of proximity spaces. Given the diagram

X′′\textstyle{X^{{}^{\prime\prime}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}k\scriptstyle{k}i0\scriptstyle{i_{0}}X\textstyle{X\ignorespaces\ignorespaces\ignorespaces\ignorespaces}p\scriptstyle{p}X′′×I\textstyle{X^{\prime\prime}\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}G\scriptstyle{G}G\scriptstyle{G^{{}^{\prime}}}{x0}\textstyle{\{x_{0}\}}

with the condition pk(x′′)=Gi0(x′′)={x0}p\circ k(x^{{}^{\prime\prime}})=G\circ i_{0}(x^{{}^{\prime\prime}})=\{x_{0}\}. Then there exists a (constant) prox-hom G:(X′′×I,δ1)(X,δ)G^{{}^{\prime}}:(X^{{}^{\prime\prime}}\times I,\delta_{1})\rightarrow(X,\delta) defined by G(x′′,t)=k(x′′)G^{{}^{\prime}}(x^{{}^{\prime\prime}},t)=k(x^{{}^{\prime\prime}}) satisfying that

pG(x′′,t)=p(G(x′′,t))={x0}=G(x′′,t),\displaystyle p\circ G^{{}^{\prime}}(x^{{}^{\prime\prime}},t)=p(G^{{}^{\prime}}(x^{{}^{\prime\prime}},t))=\{x_{0}\}=G(x^{{}^{\prime\prime}},t),
Gi0(x′′)=G(x′′,0)=k(x′′).\displaystyle G^{{}^{\prime}}\circ i_{0}(x^{{}^{\prime\prime}})=G^{{}^{\prime}}(x^{{}^{\prime\prime}},0)=k(x^{{}^{\prime\prime}}).

This proves that pp is a proximal fibration.

Proposition 3.19.

i) The composition of two proximal fibrations is also a proximal fibration.

ii) The cartesian product of two proximal fibrations is also a proximal fibration.

Proof.

i) Let p1:(X1,δ1)(Y1,δ1)p_{1}:(X_{1},\delta_{1})\rightarrow(Y_{1},\delta_{1}^{{}^{\prime}}) and p2:(Y1,δ1)(Y2,δ2)p_{2}:(Y_{1},\delta_{1}^{{}^{\prime}})\rightarrow(Y_{2},\delta_{2}^{{}^{\prime}}) be any proximal fibrations. Then for the inclusion map i0:(X′′,δ′′)(X′′×I,δ3)i_{0}:(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})\rightarrow(X^{{}^{\prime\prime}}\times I,\delta_{3}), pc-maps k1:(X′′,δ′′)(X1,δ1)k_{1}:(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})\rightarrow(X_{1},\delta_{1}), k2:(X′′,δ′′)(Y1,δ1)k_{2}:(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})\rightarrow(Y_{1},\delta_{1}^{{}^{\prime}}), and proximal homotopies G1:(X′′×I,δ3)(Y1,δ1)G_{1}:(X^{{}^{\prime\prime}}\times I,\delta_{3})\rightarrow(Y_{1},\delta_{1}^{{}^{\prime}}), G2:(X′′×I,δ3)(Y2,δ2)G_{2}:(X^{{}^{\prime\prime}}\times I,\delta_{3})\rightarrow(Y_{2},\delta_{2}^{{}^{\prime}}) with the property p1k1=G1i0p_{1}\circ k_{1}=G_{1}\circ i_{0} and p2k2=G2i0p_{2}\circ k_{2}=G_{2}\circ i_{0}, there exist two proximal homotopies G1:(X′′×I,δ3)(X1,δ1)G_{1}^{{}^{\prime}}:(X^{{}^{\prime\prime}}\times I,\delta_{3})\rightarrow(X_{1},\delta_{1}) and G2:(X′′×I,δ3)(Y1,δ1)G_{2}^{{}^{\prime}}:(X^{{}^{\prime\prime}}\times I,\delta_{3})\rightarrow(Y_{1},\delta_{1}^{{}^{\prime}}) satisfying that

G1i0=k1,p1G1=G1,\displaystyle G_{1}^{{}^{\prime}}\circ i_{0}=k_{1},p_{1}\circ G_{1}^{{}^{\prime}}=G_{1},
G2i0=k2,p2G2=G2.\displaystyle G_{2}^{{}^{\prime}}\circ i_{0}=k_{2},p_{2}\circ G_{2}^{{}^{\prime}}=G_{2}.

If we take G2=G1G_{2}^{{}^{\prime}}=G_{1}, then we have the following commutative diagram:

X′′\textstyle{X^{{}^{\prime\prime}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}k1\scriptstyle{k_{1}}i0\scriptstyle{i_{0}}X1\textstyle{X_{1}\ignorespaces\ignorespaces\ignorespaces\ignorespaces}p2p1\scriptstyle{p_{2}\circ p_{1}}X′′×I\textstyle{X^{\prime\prime}\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}G2\scriptstyle{G_{2}}G1\scriptstyle{G_{1}^{{}^{\prime}}}Y2.\textstyle{Y_{2}.}

Thus, we get

G1i0=k1,\displaystyle G_{1}^{{}^{\prime}}\circ i_{0}=k_{1},
(p2p1)G1=G2.\displaystyle(p_{2}\circ p_{1})\circ G_{1}^{{}^{\prime}}=G_{2}.

This show that the composition p2p1p_{2}\circ p_{1} is a proximal fibration.

ii) Let p1:(X1,δ1)(Y1,δ1)p_{1}:(X_{1},\delta_{1})\rightarrow(Y_{1},\delta_{1}^{{}^{\prime}}) and p2:(X2,δ2)(Y2,δ2)p_{2}:(X_{2},\delta_{2})\rightarrow(Y_{2},\delta_{2}^{{}^{\prime}}) be any proximal fibrations. Then for the inclusion map i0:(X′′,δ′′)(X′′×I,δ3)i_{0}:(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})\rightarrow(X^{{}^{\prime\prime}}\times I,\delta_{3}), pc-maps k1:(X′′,δ′′)(X1,δ1)k_{1}:(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})\rightarrow(X_{1},\delta_{1}), k2:(X′′,δ′′)(X2,δ2)k_{2}:(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})\rightarrow(X_{2},\delta_{2}), and proximal homotopies G1:(X′′×I,δ3)(Y1,δ1)G_{1}:(X^{{}^{\prime\prime}}\times I,\delta_{3})\rightarrow(Y_{1},\delta_{1}^{{}^{\prime}}), G2:(X′′×I,δ3)(Y2,δ2)G_{2}:(X^{{}^{\prime\prime}}\times I,\delta_{3})\rightarrow(Y_{2},\delta_{2}^{{}^{\prime}}) with the property p1k1=G1i0p_{1}\circ k_{1}=G_{1}\circ i_{0} and p2k2=G2i0p_{2}\circ k_{2}=G_{2}\circ i_{0}, there exist two proximal homotopies G1:(X′′×I,δ3)(X1,δ1)G_{1}^{{}^{\prime}}:(X^{{}^{\prime\prime}}\times I,\delta_{3})\rightarrow(X_{1},\delta_{1}) and G2:(X′′×I,δ3)(X2,δ2)G_{2}^{{}^{\prime}}:(X^{{}^{\prime\prime}}\times I,\delta_{3})\rightarrow(X_{2},\delta_{2}) satisfying that

G1i0=k1,p1G1=G1,\displaystyle G_{1}^{{}^{\prime}}\circ i_{0}=k_{1},p_{1}\circ G_{1}^{{}^{\prime}}=G_{1},
G2i0=k2,p2G2=G2.\displaystyle G_{2}^{{}^{\prime}}\circ i_{0}=k_{2},p_{2}\circ G_{2}^{{}^{\prime}}=G_{2}.

Consider the map G3=(G1,G2)G_{3}^{{}^{\prime}}=(G_{1}^{{}^{\prime}},G_{2}^{{}^{\prime}}). Then G3G_{3}^{{}^{\prime}} is clearly a prox-hom and we have the following commutative diagram:

X′′\textstyle{X^{{}^{\prime\prime}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}(k1,k2)\scriptstyle{(k_{1},k_{2})}i0\scriptstyle{i_{0}}X1×X2\textstyle{X_{1}\times X_{2}\ignorespaces\ignorespaces\ignorespaces\ignorespaces}p1×p2\scriptstyle{p_{1}\times p_{2}}X′′×I\textstyle{X^{\prime\prime}\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}(G1,G2)\scriptstyle{(G_{1},G_{2})}G3\scriptstyle{G_{3}^{{}^{\prime}}}Y1×Y2.\textstyle{Y_{1}\times Y_{2}.}

Thus, we get

G3i0=(k1,k2),\displaystyle G_{3}^{{}^{\prime}}\circ i_{0}=(k_{1},k_{2}),
(p1×p2)G3=(G1,G2).\displaystyle(p_{1}\times p_{2})\circ G_{3}^{{}^{\prime}}=(G_{1},G_{2}).

This proves that the cartesian product p1×p2p_{1}\times p_{2} is a proximal fibration. ∎

Let f:(X,δ1)(Y,δ2)f:(X,\delta_{1})\rightarrow(Y,\delta_{2}) be a pc-map. Then for any pc-map g:(Z,δ3)(Y,δ2)g:(Z,\delta_{3})\rightarrow(Y,\delta_{2}), a proximal lifting of ff is a pc-map h:(X,δ1)(Z,δ3)h:(X,\delta_{1})\rightarrow(Z,\delta_{3}) satisfying that f=ghf=g\circ h.

Proposition 3.20.

Let p:(X,δ1)(Y,δ2)p:(X,\delta_{1})\rightarrow(Y,\delta_{2}) be a proximal fibration. Then

i) The pullback gp:(P,δ)(Y,δ2)g^{\ast}p:(P,\delta)\rightarrow(Y^{{}^{\prime}},\delta_{2}^{{}^{\prime}}) is a proximal fibration for any pc-map g:(Y,δ2)(Y,δ2)g:(Y^{{}^{\prime}},\delta_{2}^{{}^{\prime}})\rightarrow(Y,\delta_{2}).

P\textstyle{P\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}π1\scriptstyle{\pi_{1}}gp\scriptstyle{g^{\ast}p}X\textstyle{X\ignorespaces\ignorespaces\ignorespaces\ignorespaces}p\scriptstyle{p}Y\textstyle{Y^{{}^{\prime}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces}g\scriptstyle{g}Y.\textstyle{Y.}

ii) For any proximity space (Z,δ3)(Z,\delta_{3}), the map p:(XZ,δ3)(YZ,δ3′′)p_{\ast}:(X^{Z},\delta_{3}^{{}^{\prime}})\rightarrow(Y^{Z},\delta_{3}^{{}^{\prime\prime}}) is a proximal fibration.

Proof.

i) Let

P={(x,y)|g(y)=p(e)}X×YP=\{(x,y^{{}^{\prime}})\ |\ g(y^{{}^{\prime}})=p(e)\}\subseteq X\times Y^{{}^{\prime}}

be a proximity space with the proximity δ0\delta_{0} on itself. Since pp is a proximal fibration, for an inclusion map i0:(X′′,δ′′)(X′′×I,δ3)i_{0}:(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})\rightarrow(X^{{}^{\prime\prime}}\times I,\delta_{3}), for any pc-map k1k_{1} from (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) to (X,δ1)(X,\delta_{1}), and prox-hom G1:(X′′×I,δ3)(X,δ)G_{1}:(X^{{}^{\prime\prime}}\times I,\delta_{3})\rightarrow(X,\delta) with pk1=G1i0p\circ k_{1}=G_{1}\circ i_{0}, there exists a prox-hom

G1:(X′′×I,δ3)(X,δ1)G_{1}^{{}^{\prime}}:(X^{{}^{\prime\prime}}\times I,\delta_{3})\rightarrow(X,\delta_{1})

for which G1(x′′,0)=k1(x′′)G_{1}^{{}^{\prime}}(x^{{}^{\prime\prime}},0)=k_{1}(x^{{}^{\prime\prime}}) and pG1(x′′,t)=G1(x′′,t)p\circ G_{1}^{{}^{\prime}}(x^{{}^{\prime\prime}},t)=G_{1}(x^{{}^{\prime\prime}},t). Assume that a map k2k_{2} from (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) to (P,δ0)(P,\delta_{0}) is a pc-map and G2:(X′′×I,δ3)(Y,δ2)G_{2}:(X^{{}^{\prime\prime}}\times I,\delta_{3})\rightarrow(Y^{{}^{\prime}},\delta_{2}^{{}^{\prime}}) is a prox-hom with gpk2=G2i0g^{\ast}p\circ k_{2}=G_{2}\circ i_{0}. If we define G2:(X′′×I,δ3)(P,δ0)G_{2}^{{}^{\prime}}:(X^{{}^{\prime\prime}}\times I,\delta_{3})\rightarrow(P,\delta_{0}) by G2=(G1,G2)G_{2}^{{}^{\prime}}=(G_{1}^{{}^{\prime}},G_{2}), then we observe that

G2i0=k2,\displaystyle G_{2}^{{}^{\prime}}\circ i_{0}=k_{2},
gpG2=G2.\displaystyle g^{\ast}p\circ G_{2}^{{}^{\prime}}=G_{2}.

This gives the desired result.

ii) Consider the following diagrams:

Z×X′′\textstyle{Z\times X^{{}^{\prime\prime}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}k1\scriptstyle{k_{1}}i0\scriptstyle{i_{0}}X\textstyle{X\ignorespaces\ignorespaces\ignorespaces\ignorespaces}p\scriptstyle{p}Z×X′′×I\textstyle{Z\times X^{\prime\prime}\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}G1\scriptstyle{G_{1}}G1\scriptstyle{G_{1}^{{}^{\prime}}}Y\textstyle{Y}

and

X′′\textstyle{X^{{}^{\prime\prime}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}k2\scriptstyle{k_{2}}i0\scriptstyle{i_{0}}XZ\textstyle{X^{Z}\ignorespaces\ignorespaces\ignorespaces\ignorespaces}p\scriptstyle{p_{\ast}}X′′×I\textstyle{X^{\prime\prime}\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}G2\scriptstyle{G_{2}}G2\scriptstyle{G_{2}^{{}^{\prime}}}YZ.\textstyle{Y^{Z}.}

Since pp is a proximal fibration, we have H1:(Z×X′′×I,δ′′)(X,δ1)H_{1}^{{}^{\prime}}:(Z\times X^{\prime\prime}\times I,\delta^{{}^{\prime\prime}})\rightarrow(X,\delta_{1}) as the prox-hom in the upper diagram. Z×X′′×IZ\times X^{\prime\prime}\times I is proximally isomorphic to X′′×I×ZX^{{}^{\prime\prime}}\times I\times Z and we can think of G1G_{1}^{{}^{\prime}} as the prox-hom (X′′×I×Z,δ′′)(X,δ1)(X^{{}^{\prime\prime}}\times I\times Z,\delta^{{}^{\prime\prime}})\rightarrow(X,\delta_{1}). By Proposition 3.4, we have the prox-hom G2:(X′′×I,δ)(XZ,δ3)G_{2}^{{}^{\prime}}:(X^{{}^{\prime\prime}}\times I,\delta^{{}^{\prime}})\rightarrow(X^{Z},\delta_{3}^{{}^{\prime}}) in the lower diagram. This map satisfies the desired conditions, and thus, we conclude that pp_{\ast} is a proximal fibration. ∎

3.4. Proximal Cofibrations

Similar to the proximal fibration, we currently deal with the notion of proximal cofibration of proximity spaces. We first study the problem of extension in homotopy theory, and then present the definition of proximal cofibration with its basic results.

Definition 3.21.

Given two proximity spaces (X,δ)(X,\delta) and (X,δ)(X^{{}^{\prime}},\delta^{{}^{\prime}}), a pc-map h:XXh:X\rightarrow X^{{}^{\prime}} is said to have a proximal homotopy extension property (PHEP) with respect to a proximity space (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) if for inclusion maps i0X:(X,δ)(X×I,δ1)i_{0}^{X}:(X,\delta)\rightarrow(X\times I,\delta_{1}) and i0X:(X,δ)(X×I,δ1)i_{0}^{X^{{}^{\prime}}}:(X^{{}^{\prime}},\delta^{{}^{\prime}})\rightarrow(X^{{}^{\prime}}\times I,\delta_{1}^{{}^{\prime}}), for every pc-map k:(X,δ)(X′′,δ′′)k:(X^{{}^{\prime}},\delta^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}), and prox-hom F:(X×I,δ1)(X′′,δ′′)F:(X\times I,\delta_{1})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) with k(h×10)=Fi0Xk\circ(h\times 1_{0})=F\circ i_{0}^{X}, then there exists a prox-hom F:(X×I,δ1)(X′′,δ′′)F^{{}^{\prime}}:(X^{{}^{\prime}}\times I,\delta_{1}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) satisfying Fi0X=kF^{{}^{\prime}}\circ i_{0}^{X^{{}^{\prime}}}=k and F(h×1I)=FF^{{}^{\prime}}\circ(h\times 1_{I})=F.

X×0\textstyle{X\times 0\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}i0X\scriptstyle{i_{0}^{X}}h×10\scriptstyle{h\times 1_{0}}X×I\textstyle{X\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}h×1I\scriptstyle{h\times 1_{I}}F\scriptstyle{F}X′′\textstyle{X^{{}^{\prime\prime}}}X×0\textstyle{X^{{}^{\prime}}\times 0\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}i0X\scriptstyle{i_{0}^{X^{{}^{\prime}}}}k\scriptstyle{k}X×I.\textstyle{X^{{}^{\prime}}\times I.\ignorespaces\ignorespaces\ignorespaces\ignorespaces}F\scriptstyle{F^{{}^{\prime}}}
Refer to caption
Figure 3.3. The picture is represented by X′′={a,b,c,d}X^{{}^{\prime\prime}}=\{a,b,c,d\}.
Example 3.22.

Let X′′={a,b,c,d}X^{{}^{\prime\prime}}=\{a,b,c,d\} be a set with the proximity δ′′\delta^{{}^{\prime\prime}} on itself as in Figure 3.3. Let γ1\gamma_{1} and γ2\gamma_{2} be proximal paths on X′′X^{{}^{\prime\prime}} such that γ1(0)=b\gamma_{1}(0)=b, γ1(1)=a\gamma_{1}(1)=a, γ2(0)=b\gamma_{2}(0)=b, and γ2(1)=c\gamma_{2}(1)=c. Consider the following diagram for an inclusion map h:({0},δ)(I,δ)h:(\{0\},\delta)\rightarrow(I,\delta^{{}^{\prime}}), where k:(I,δ)(X′′,δ′′)k:(I,\delta^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) is the map γ2\gamma_{2} and F:({0}×I,δ1)(X′′,δ′′)F:(\{0\}\times I,\delta_{1})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) is defined by F(0,t)=γ1(t)F(0,t)=\gamma_{1}(t) for all tIt\in I:

{0}×0\textstyle{\{0\}\times 0\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}i0{0}\scriptstyle{i_{0}^{\{0\}}}h×10\scriptstyle{h\times 1_{0}}{0}×I\textstyle{\{0\}\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}h×1I\scriptstyle{h\times 1_{I}}F\scriptstyle{F}X′′\textstyle{X^{{}^{\prime\prime}}}I×0\textstyle{I\times 0\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}i0I\scriptstyle{i_{0}^{I}}k\scriptstyle{k}I×I,\textstyle{I\times I,}

i.e., the equality k(h×10)=Fi0{0}k\circ(h\times 1_{0})=F\circ i_{0}^{\{0\}} holds. Then, by Gluing Lemma, there exists a prox-hom F:(I×I,δ1)(X′′,δ′′)F^{{}^{\prime}}:(I\times I,\delta_{1}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) defined by F(0,t1)=F(0,t1)F^{{}^{\prime}}(0,t_{1})=F(0,t_{1}) and F(t2,0)=k(t2)F^{{}^{\prime}}(t_{2},0)=k(t_{2}) for all (t1,t2)I×I(t_{1},t_{2})\in I\times I which satisfy

F(h×1I)=F,\displaystyle F^{{}^{\prime}}\circ(h\times 1_{I})=F,
Fi0I=k.\displaystyle F^{{}^{\prime}}\circ i_{0}^{I}=k.

Schematically, we have the diagram

{0}×0\textstyle{\{0\}\times 0\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}i0{0}\scriptstyle{i_{0}^{\{0\}}}h×10\scriptstyle{h\times 1_{0}}{0}×I\textstyle{\{0\}\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}h×1I\scriptstyle{h\times 1_{I}}F\scriptstyle{F}X′′\textstyle{X^{{}^{\prime\prime}}}I×0\textstyle{I\times 0\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}i0I\scriptstyle{i_{0}^{I}}k\scriptstyle{k}I×I.\textstyle{I\times I.\ignorespaces\ignorespaces\ignorespaces\ignorespaces}F\scriptstyle{F^{{}^{\prime}}}
Definition 3.23.

A pc-map h:(X,δ)(X,δ)h:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) is said to be a proximal cofibration if it has the PHEP with respect to any proximity space (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}).

Example 3.24.

Let h:(X,δ)(X,δ)h:(X,\delta)\hookrightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) be an inclusion map such that XXX\subset X^{{}^{\prime}}. Then hh is a natural proximal cofibration since there exists a prox-hom

F=F|X:(X×I,δ1)(X′′,δ′′)F^{{}^{\prime}}=F|_{X^{{}^{\prime}}}:(X^{{}^{\prime}}\times I,\delta_{1}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})

satisfying the conditions of PHEP with respect to any proximity space (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}).

Proposition 3.25.

i) Let h:(X,δ)(X,δ)h:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) and h:(Y,δ1)(Y,δ1)h^{{}^{\prime}}:(Y,\delta_{1})\rightarrow(Y^{{}^{\prime}},\delta_{1}^{{}^{\prime}}) be two maps such that XX and XX^{{}^{\prime}} are proximally isomorphic to YY and YY^{{}^{\prime}}, respectively, and the following diagram commutes:

X\textstyle{X\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}δ\scriptstyle{\approx_{\delta}}h\scriptstyle{h}Y\textstyle{Y\ignorespaces\ignorespaces\ignorespaces\ignorespaces}h\scriptstyle{h^{{}^{\prime}}}X\textstyle{X^{{}^{\prime}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces}δ\scriptstyle{\approx_{\delta}}Y.\textstyle{Y^{{}^{\prime}}.}

Then hh is a proximal cofibration if and only if hh^{{}^{\prime}} is a proximal cofibration.

ii) The composition of two proximal cofibrations is also a proximal cofibration.

iii) The coproduct of two proximal cofibrations is also a proximal cofibration.

iv) Let h:(X,δ)(X,δ)h:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) be a proximal cofibration and the following is a pushout diagram.

X\textstyle{X\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}l\scriptstyle{l}h\scriptstyle{h}Y\textstyle{Y\ignorespaces\ignorespaces\ignorespaces\ignorespaces}h\scriptstyle{h^{{}^{\prime}}}X\textstyle{X^{{}^{\prime}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces}l\scriptstyle{l^{{}^{\prime}}}Y.\textstyle{Y^{{}^{\prime}}.}

Then hh^{{}^{\prime}} is a proximal cofibration.

Proof.

i) Let hh be a proximal cofibration. By Definition 3.21, there is a prox-hom F:(X×I,δ2)(X′′,δ′′)F^{{}^{\prime}}:(X^{{}^{\prime}}\times I,\delta_{2}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) such that

Fi0X=kandF(h×1I)=FF^{{}^{\prime}}\circ i_{0}^{X^{{}^{\prime}}}=k\ \ \text{and}\ \ F^{{}^{\prime}}\circ(h\times 1_{I})=F

for any pc-map k:(X,δ)(X′′,δ′′)k:(X^{{}^{\prime}},\delta^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) and prox-hom FF from (X×I,δ2)(X\times I,\delta_{2}) to (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) with k(h×10)=Fi0Xk\circ(h\times 1_{0})=F\circ i_{0}^{X}. Assume that β1:XY\beta_{1}:X\rightarrow Y and β2:XY\beta_{2}:X^{{}^{\prime}}\rightarrow Y^{{}^{\prime}} are two proximal isomorphisms. Since the diagram commutes, we know that

hβ1=β2h.h^{{}^{\prime}}\circ\beta_{1}=\beta_{2}\circ h.

Let i0Y:(Y,δ1)(Y×I,δ3)i_{0}^{Y}:(Y,\delta_{1})\rightarrow(Y\times I,\delta_{3}) and i0Y:(Y,δ1)(Y×I,δ3)i_{0}^{Y^{{}^{\prime}}}:(Y^{{}^{\prime}},\delta_{1}^{{}^{\prime}})\rightarrow(Y^{{}^{\prime}}\times I,\delta_{3}^{{}^{\prime}}) be two inclusion maps, k:=k(β2)1:(Y,δ1)(X′′,δ′′)k^{{}^{\prime}}:=k\circ(\beta_{2})^{-1}:(Y^{{}^{\prime}},\delta_{1}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) a pc-map, and F′′:=F(β11×1I)F^{{}^{\prime\prime}}:=F^{{}^{\prime}}\circ(\beta_{1}^{-1}\times 1_{I}) from (Y×I,δ3)(Y\times I,\delta_{3}) to (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) a prox-hom for which

k(h×10)=Fi0Y.k^{{}^{\prime}}\circ(h^{{}^{\prime}}\times 1_{0})=F^{{}^{\prime}}\circ i_{0}^{Y}.

Then there exists a prox-hom

F′′′:=F((β2)1×1I):(Y×I,δ3)(X′′,δ′′)F^{{}^{\prime\prime\prime}}:=F^{{}^{\prime}}\circ((\beta_{2})^{-1}\times 1_{I}):(Y^{{}^{\prime}}\times I,\delta_{3}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})

such that F′′i0Y=kF^{{}^{\prime\prime}}\circ i_{0}^{Y^{{}^{\prime}}}=k^{{}^{\prime}} and F′′(h×1I)=FF^{{}^{\prime\prime}}\circ(h^{{}^{\prime}}\times 1_{I})=F^{{}^{\prime}}. Conversely, assume that hh^{{}^{\prime}} is a proximal cofibration. Similarly, for a prox-hom FF^{{}^{\prime}} from (Y×I,δ3)(Y^{{}^{\prime}}\times I,\delta_{3}^{{}^{\prime}}) to (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) that makes hh^{{}^{\prime}} a cofibration, there exists a prox-hom

F′′:=F(f×1I):(X×I,δ2)(X′′,δ′′)F^{{}^{\prime\prime}}:=F^{{}^{\prime}}\circ(f^{{}^{\prime}}\times 1_{I}):(X^{{}^{\prime}}\times I,\delta_{2}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})

that makes hh a proximal cofibration.

ii) Let h:(X,δ)(X,δ)h:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) and h:(X,δ)(Y,δ1)h^{{}^{\prime}}:(X^{{}^{\prime}},\delta^{{}^{\prime}})\rightarrow(Y,\delta_{1}) be two proximal cofibrations. Then for any pc-map k:(X,δ)(X′′,δ′′)k:(X^{{}^{\prime}},\delta^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) and prox-hom FF from (X×I,δ2)(X\times I,\delta_{2}) to (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) with

k(h×10)=Fi0X,k\circ(h\times 1_{0})=F\circ i_{0}^{X},

there is a prox-hom F:(X×I,δ2)(X′′,δ′′)F^{{}^{\prime}}:(X^{{}^{\prime}}\times I,\delta_{2}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) such that Fi0X=kF^{{}^{\prime}}\circ i_{0}^{X^{{}^{\prime}}}=k and F(h×1I)=FF^{{}^{\prime}}\circ(h\times 1_{I})=F, similarly, for any pc-map k:(Y,δ1)(X′′,δ′′)k^{{}^{\prime}}:(Y,\delta_{1})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) and prox-hom G:(X×I,δ2)(X′′,δ′′)G^{{}^{\prime}}:(X^{{}^{\prime}}\times I,\delta_{2}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) with

k(h×10)=Gi0X,k^{{}^{\prime}}\circ(h^{{}^{\prime}}\times 1_{0})=G^{{}^{\prime}}\circ i_{0}^{X^{{}^{\prime}}},

there is a prox-hom F′′:(Y×I,δ3)(X′′,δ′′)F^{{}^{\prime\prime}}:(Y\times I,\delta_{3}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) such that F′′i0Y=kF^{{}^{\prime\prime}}\circ i_{0}^{Y}=k^{{}^{\prime}} and F′′(h×1I)=GF^{{}^{\prime\prime}}\circ(h^{{}^{\prime}}\times 1_{I})=G^{{}^{\prime}}. Combining these results with the fact F=GF^{{}^{\prime}}=G^{{}^{\prime}}, we have the following: For a pc-map k′′:=kk^{{}^{\prime\prime}}:=k^{{}^{\prime}} and prox-hom G′′:=FG^{{}^{\prime\prime}}:=F with

k′′((hh)×10)=G′′i0X,k^{{}^{\prime\prime}}\circ((h^{{}^{\prime}}\circ h)\times 1_{0})=G^{{}^{\prime\prime}}\circ i_{0}^{X},

there is a prox-hom F′′′:=F′′F^{{}^{\prime\prime\prime}}:=F^{{}^{\prime\prime}} such that

F′′′i0Y=k′′andF′′′((hh)×1I)=G′′.F^{{}^{\prime\prime\prime}}\circ i_{0}^{Y}=k^{{}^{\prime\prime}}\ \ \text{and}\ \ F^{{}^{\prime\prime\prime}}\circ((h^{{}^{\prime}}\circ h)\times 1_{I})=G^{{}^{\prime\prime}}.

This proves that hhh^{{}^{\prime}}\circ h is a proximal cofibration.

iii) Let hj:(Xj,δj)(Xj,δj)h_{j}:(X_{j},\delta_{j})\rightarrow(X_{j}^{{}^{\prime}},\delta_{j}^{{}^{\prime}}) be a family of cofibrations for all jJj\in J. Then we shall show that jhj:(jXj,δ)(jXj,δ)\sqcup_{j}h_{j}:(\sqcup_{j}X_{j},\delta)\rightarrow(\sqcup_{j}X_{j}^{{}^{\prime}},\delta^{{}^{\prime}}) is a cofibration. Since for all jJj\in J, hj:(Xj,δj)(Xj,δj)h_{j}:(X_{j},\delta_{j})\rightarrow(X_{j}^{{}^{\prime}},\delta_{j}^{{}^{\prime}}) is cofibration, we have that for any pc-map kjk_{j} from (Xj,δ)(X_{j}^{{}^{\prime}},\delta^{{}^{\prime}}) to (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) and prox-hom Fj:(Xj×I,δ2)(X′′,δ′′)F_{j}:(X_{j}\times I,\delta_{2})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) with

kj(hj×10)=Fji0Xj,k_{j}\circ(h_{j}\times 1_{0})=F_{j}\circ i_{0}^{X_{j}},

there is a prox-hom Fj:(Xj×I,δ2)(X′′,δ′′)F^{{}^{\prime}}_{j}:(X_{j}^{{}^{\prime}}\times I,\delta_{2}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) such that Fji0Xj=kjF^{{}^{\prime}}_{j}\circ i_{0}^{X_{j}^{{}^{\prime}}}=k_{j} and Fj(hj×1I)=FjF^{{}^{\prime}}_{j}\circ(h_{j}\times 1_{I})=F_{j}. Now assume that for a pc-map kj:(jXj,δ)(X′′,δ′′)\sqcup k_{j}:(\sqcup_{j}X_{j}^{{}^{\prime}},\delta^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) and prox-hom jFj:(jXj×I,δ4)(X′′,δ′′)\sqcup_{j}F_{j}:(\sqcup_{j}X_{j}\times I,\delta_{4})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) with

jkj(jhj×10)=jFji0jXj.\sqcup_{j}k_{j}\circ(\sqcup_{j}h_{j}\times 1_{0})=\sqcup_{j}F_{j}\circ i_{0}^{\sqcup_{j}X_{j}^{{}^{\prime}}}.

Then there exists a map jFj:(jXj×I,δ5)(X′′,δ′′)\sqcup_{j}F^{{}^{\prime}}_{j}:(\sqcup_{j}X_{j}^{{}^{\prime}}\times I,\delta_{5})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) such that Fj=jFjijF^{{}^{\prime}}_{j}=\sqcup_{j}F^{{}^{\prime}}_{j}\circ i_{j} for a map ij:(Xj×I,δ2)(jXj×I,δ5)i_{j}:(X_{j}^{{}^{\prime}}\times I,\delta_{2}^{{}^{\prime}})\rightarrow(\sqcup_{j}X_{j}^{{}^{\prime}}\times I,\delta_{5}). If we define iji_{j}^{{}^{\prime}} as iji0Xji_{j}\circ i_{0}^{X_{j}^{{}^{\prime}}}, then we find that i0jXj=jiji_{0}^{\sqcup_{j}X_{j}^{{}^{\prime}}}=\sqcup_{j}i_{j}^{{}^{\prime}}. It follows that

jFji0jXj=jkj\displaystyle\sqcup_{j}F^{{}^{\prime}}_{j}\circ i_{0}^{\sqcup_{j}X_{j}^{{}^{\prime}}}=\sqcup_{j}k_{j}

and

jFj(jhj×1I)=jFj.\displaystyle\sqcup_{j}F^{{}^{\prime}}_{j}\circ(\sqcup_{j}h_{j}\times 1_{I})=\sqcup_{j}F_{j}.

Finally, we have that jhj\sqcup_{j}h_{j} is a cofibration.

iv) Let h:(X,δ)(X,δ)h:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) be a proximal cofibration, i.e., there is a prox-hom F:(X×I,δ2)(X′′,δ′′)F^{{}^{\prime}}:(X^{{}^{\prime}}\times I,\delta_{2}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) such that

Fi0X=kandF(h×1I)=FF^{{}^{\prime}}\circ i_{0}^{X^{{}^{\prime}}}=k\ \ \text{and}\ \ F^{{}^{\prime}}\circ(h\times 1_{I})=F

for any pc-map k:(X,δ)(X′′,δ′′)k:(X^{{}^{\prime}},\delta^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) and prox-hom FF from (X×I,δ2)(X\times I,\delta_{2}) to (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) with k(h×10)=Fi0Xk\circ(h\times 1_{0})=F\circ i_{0}^{X}. Since we have a pushout diagram, it follows that lh=hll^{{}^{\prime}}\circ h=h^{{}^{\prime}}\circ l holds. Now assume that k:(Y,δ1)(X′′,δ′′)k^{{}^{\prime}}:(Y^{{}^{\prime}},\delta_{1}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) is a pc-map, and F′′F^{{}^{\prime\prime}} from (Y×I,δ3)(Y^{{}^{\prime}}\times I,\delta_{3}) to (X′′,δ′′)(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) is a prox-hom with kl=kk^{{}^{\prime}}\circ l^{{}^{\prime}}=k, F′′(l×1I)F^{{}^{\prime\prime}}\circ(l\times 1_{I}), and

k(h×10)=F′′i0Y.k^{{}^{\prime}}\circ(h^{{}^{\prime}}\times 1_{0})=F^{{}^{\prime\prime}}\circ i_{0}^{Y^{{}^{\prime}}}.

Then there exists a prox-hom

F′′′:(Y×I,δ3)(X′′,δ′′)F^{{}^{\prime\prime\prime}}:(Y^{{}^{\prime}}\times I,\delta_{3})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})

such that F′′′(l×1I)=FF^{{}^{\prime\prime\prime}}\circ(l^{{}^{\prime}}\times 1_{I})=F^{{}^{\prime}}. Moreover, we have that

F′′′(l×1I)=F\displaystyle F^{{}^{\prime\prime\prime}}\circ(l^{{}^{\prime}}\times 1_{I})=F^{{}^{\prime}} \displaystyle\Rightarrow F′′′(l×1I)i0X=Fi0X\displaystyle F^{{}^{\prime\prime\prime}}\circ(l^{{}^{\prime}}\times 1_{I})\circ i_{0}^{X^{{}^{\prime}}}=F^{{}^{\prime}}\circ i_{0}^{X^{{}^{\prime}}}
\displaystyle\Rightarrow F′′′i0Yl=k\displaystyle F^{{}^{\prime\prime\prime}}\circ i_{0}^{Y^{{}^{\prime}}}\circ l^{{}^{\prime}}=k
\displaystyle\Rightarrow F′′′i0Yl=kl\displaystyle F^{{}^{\prime\prime\prime}}\circ i_{0}^{Y^{{}^{\prime}}}\circ l^{{}^{\prime}}=k^{{}^{\prime}}\circ l^{{}^{\prime}}
\displaystyle\Rightarrow F′′′i0Y=k,\displaystyle F^{{}^{\prime\prime\prime}}\circ i_{0}^{Y^{{}^{\prime}}}=k^{{}^{\prime}},

and

F(h×1I)=F\displaystyle F^{{}^{\prime}}\circ(h\times 1_{I})=F^{{}^{\prime}} \displaystyle\Rightarrow F′′(l×1I)(h×1I)=F′′(l×1I)\displaystyle F^{{}^{\prime\prime}}\circ(l^{{}^{\prime}}\times 1_{I})\circ(h\times 1_{I})=F^{{}^{\prime\prime}}\circ(l\times 1_{I})
\displaystyle\Rightarrow F′′′((lh)×1I)=F′′(l×1I)\displaystyle F^{{}^{\prime\prime\prime}}\circ((l^{{}^{\prime}}\circ h)\times 1_{I})=F^{{}^{\prime\prime}}\circ(l\times 1_{I})
\displaystyle\Rightarrow F′′′((hl)×1I)=F′′(l×1I)\displaystyle F^{{}^{\prime\prime\prime}}\circ((h^{{}^{\prime}}\circ l)\times 1_{I})=F^{{}^{\prime\prime}}\circ(l\times 1_{I})
\displaystyle\Rightarrow F′′′(h×1I)(l×1I)=F′′(l×1I)\displaystyle F^{{}^{\prime\prime\prime}}\circ(h^{{}^{\prime}}\times 1_{I})\circ(l\times 1_{I})=F^{{}^{\prime\prime}}\circ(l\times 1_{I})
\displaystyle\Rightarrow F′′′(h×1I)=F′′.\displaystyle F^{{}^{\prime\prime\prime}}\circ(h^{{}^{\prime}}\times 1_{I})=F^{{}^{\prime\prime}}.

As a consequence, hh^{{}^{\prime}} is a proximal cofibration. ∎

Theorem 3.26.

h:(X,δ)(X,δ)h:(X,\delta)\rightarrow(X^{{}^{\prime}},\delta^{{}^{\prime}}) is a proximal cofibration if and only if (X×0)(X×I)(X^{{}^{\prime}}\times 0)\cup(X\times I) is a proximal retract of X×IX^{{}^{\prime}}\times I.

Proof.

Let X′′=(X×0)(X×I)X^{{}^{\prime\prime}}=(X^{{}^{\prime}}\times 0)\cup(X\times I). If ff is a proximal cofibration, then for any pc-map k:(X,δ)(X′′,δ′′)k:(X^{{}^{\prime}},\delta^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) and prox-hom F:(X×I,δ2)(X′′,δ′′)F:(X\times I,\delta_{2})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) with k(h×10)=Fi0Xk\circ(h\times 1_{0})=F\circ i_{0}^{X}, there is a prox-hom F:(X×I,δ2)(X′′,δ′′)F^{{}^{\prime}}:(X^{{}^{\prime}}\times I,\delta_{2}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) such that Fi0X=kF^{{}^{\prime}}\circ i_{0}^{X^{{}^{\prime}}}=k and F(h×1I)=FF^{{}^{\prime}}\circ(h\times 1_{I})=F. Hence, FF^{{}^{\prime}} is a proximal retraction of X×IX^{{}^{\prime}}\times I. Conversely, let h:(X×I,δ2)(X′′,δ′′)h:(X^{{}^{\prime}}\times I,\delta_{2}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}}) be a proximal retraction. Assume that k:(X,δ)(Y,δ)k:(X^{{}^{\prime}},\delta^{{}^{\prime}})\rightarrow(Y,\delta) is a pc-map and F:(X×I,δ2)(Y,δ)F:(X\times I,\delta_{2})\rightarrow(Y,\delta) is a prox-hom with

k(h×10)=Fi0X.k\circ(h\times 1_{0})=F\circ i_{0}^{X}.

Define a map F′′:(X′′,δ′′)(Y,δ)F^{{}^{\prime\prime}}:(X^{{}^{\prime\prime}},\delta^{{}^{\prime\prime}})\rightarrow(Y,\delta) by F′′(x,t)=F(x,t)F^{{}^{\prime\prime}}(x^{{}^{\prime}},t)=F(x^{{}^{\prime}},t) and F′′(x,0)=k(x)F^{{}^{\prime\prime}}(x^{{}^{\prime}},0)=k(x). By Lemma 2.3, F′′F^{{}^{\prime\prime}} is a pc-map. Therefore, the map F=F′′kF^{{}^{\prime}}=F^{{}^{\prime\prime}}\circ k is a proximal fibration satisfying that Fi0X=kF^{{}^{\prime}}\circ i_{0}^{X^{{}^{\prime}}}=k and F(h×1I)=FF^{{}^{\prime}}\circ(h\times 1_{I})=F. This shows that hh is a proximal cofibration. ∎

4. Descriptive Proximity Definitions

This section is dedicated to describing the concepts given in Section 3 on descriptive proximity spaces. Recall that a (spatial) proximity is also a descriptive proximity, and note that, in the examples of this section, descriptions of feature vectors consider the colors of boxes or some parts of balls (see Example 4.2, Example 4.7, and Example 4.12).

Definition 4.1.

Let (X,δΦ1)(X,\delta^{1}_{\Phi}) and (Y,δΦ2)(Y,\delta^{2}_{\Phi}) be two descriptive proximity spaces. The descriptive proximal mapping space YXY^{X} is defined as the set

{α:XY|αis a dpc-map}\{\alpha:X\rightarrow Y\ |\ \alpha\ \text{is a dpc-map}\}

having the following descriptive proximity relation δΦ\delta_{\Phi} on itself: Let EE, FXF\subset X and {αi}iI\{\alpha_{i}\}_{i\in I} and {βj}jJ\{\beta_{j}\}_{j\in J} be any subsets of dpc-maps in YXY^{X}. We say that {αi}iIδΦ{βj}jJ\{\alpha_{i}\}_{i\in I}\delta_{\Phi}\{\beta_{j}\}_{j\in J} if the fact EδΦ1FE\delta^{1}_{\Phi}F implies that αi(E)δΦ2βj(F)\alpha_{i}(E)\delta^{2}_{\Phi}\beta_{j}(F).

Example 4.2.

Consider the set X={a,b,c,d,e,f,g,h}X=\{a,b,c,d,e,f,g,h\} in Figure 3.1 with the descriptive proximity δΦ\delta_{\Phi}, where Φ\Phi is a set of probe functions that admits colors of given boxes. Define three descriptive proximal paths γ1\gamma_{1}, γ2\gamma_{2}, and γ3XI\gamma_{3}\in X^{I} by

γ1:abcd,\displaystyle\gamma_{1}:a\mapsto b\mapsto c\mapsto d,
γ2:cbah,\displaystyle\gamma_{2}:c\mapsto b\mapsto a\mapsto h,
γ3:ahgf.\displaystyle\gamma_{3}:a\mapsto h\mapsto g\mapsto f.

For all tIt\in I, γ1(t)δΦγ2(t)\gamma_{1}(t)\delta_{\Phi}\gamma_{2}(t). Indeed,

γ1(t)={red,t[0,1/4]and[3/4,1]green,t[1/4,2/4]black,t[2/4,3/4]=γ2(t),\displaystyle\gamma_{1}(t)=\begin{cases}\text{red},&t\in[0,1/4]\ \text{and}\ [3/4,1]\\ \text{green},&t\in[1/4,2/4]\\ \text{black},&t\in[2/4,3/4]\end{cases}\ \ =\gamma_{2}(t),

namely that, γ1\gamma_{1} is descriptively near γ2\gamma_{2}. However, for t[1/4,2/4]t\in[1/4,2/4], we have that α1(t)=\alpha_{1}(t)= green and α3(t)=\alpha_{3}(t)= black, that is, α1\alpha_{1} and α3\alpha_{3} are not descriptively near in XX.

Definition 4.3.

We say that a map H:(X,δΦ1)(ZY,δΦ)H:(X,\delta^{1}_{\Phi})\rightarrow(Z^{Y},\delta_{\Phi}^{{}^{\prime}}) is descriptive proximally continuous if the fact EδΦ1FE\delta^{1}_{\Phi}F implies that H(E)δΦH(F)H(E)\delta_{\Phi}^{{}^{\prime}}H(F) for any subsets EE, FXF\subset X.

Definition 4.4.

For any descriptive proximity spaces (X,δΦ1)(X,\delta^{1}_{\Phi}) and (Y,δΦ2)(Y,\delta^{2}_{\Phi}), the descriptive proximal evaluation map

eX,Y:(YX×X,δΦ)(Y,δΦ2)e_{X,Y}:(Y^{X}\times X,\delta_{\Phi})\rightarrow(Y,\delta^{2}_{\Phi})

is defined by e(α,x)=α(x)e(\alpha,x)=\alpha(x).

Proposition 4.5.

The descriptive proximal evaluation map eX,Ye_{X,Y} is a dpc-map.

Proof.

We shall show that for any EE, FXF\subset X and {αi}iI\{\alpha_{i}\}_{i\in I}, {βj}jJYX\{\beta_{j}\}_{j\in J}\subset Y^{X}, ({αi}iI×E)δΦ({βj}jJ×F)(\{\alpha_{i}\}_{i\in I}\times E)\delta_{\Phi}(\{\beta_{j}\}_{j\in J}\times F) implies eX,Y({αi}iI×E)δΦ2eX,Y({βj}jJ×F)e_{X,Y}(\{\alpha_{i}\}_{i\in I}\times E)\delta^{2}_{\Phi}e_{X,Y}(\{\beta_{j}\}_{j\in J}\times F).

({αi}iI×E)δΦ({βj}jJ×F)\displaystyle(\{\alpha_{i}\}_{i\in I}\times E)\delta_{\Phi}(\{\beta_{j}\}_{j\in J}\times F)\ \ \displaystyle\Rightarrow {αi}iIδΦ{βj}jJandEδΦ1F\displaystyle\{\alpha_{i}\}_{i\in I}\delta_{\Phi}^{{}^{\prime}}\{\beta_{j}\}_{j\in J}\ \ \text{and}\ \ E\delta^{1}_{\Phi}F
\displaystyle\Rightarrow αi(E)δΦ2βj(F),iI,jJ\displaystyle\alpha_{i}(E)\delta^{2}_{\Phi}\beta_{j}(F),\ \ \forall i\in I,\ \forall j\in J
\displaystyle\Rightarrow eX,Y({αi}iI×E)δΦ2eX,Y({βj}jJ×F),\displaystyle e_{X,Y}(\{\alpha_{i}\}_{i\in I}\times E)\delta^{2}_{\Phi}e_{X,Y}(\{\beta_{j}\}_{j\in J}\times F),

where YXY^{X} has a descriptive proximity δΦ\delta_{\Phi}^{{}^{\prime}}. ∎

Definition 4.6.

A surjective and dpc-map p:(X,δΦ)(X,δΦ)p:(X,\delta_{\Phi})\rightarrow(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}}) between any descriptive proximity spaces (X,δΦ)(X,\delta_{\Phi}) and (X,δΦ)(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}}) is a descriptive proximal covering map if the following hold:

  • Let {x}X\{x^{{}^{\prime}}\}\subseteq X^{{}^{\prime}} be any subset with {x}δΦY\{x^{{}^{\prime}}\}\ll_{\delta_{\Phi}^{{}^{\prime}}}Y^{{}^{\prime}}. Then there is an index set II satisfying that

    p1(Y)=iIYi\displaystyle p^{-1}(Y^{{}^{\prime}})=\displaystyle\bigcup_{i\in I}Y_{i}

    with ViδΦYiV_{i}\ll_{\delta_{\Phi}}Y_{i}, where Vip1({x})V_{i}\in p^{-1}(\{x^{{}^{\prime}}\}) for each iIi\in I.

  • YiYjY_{i}\neq Y_{j} when iji\neq j for ii, jIj\in I.

  • p|Yi:YiYp|_{Y_{i}}:Y_{i}\rightarrow Y^{{}^{\prime}} is a descriptive proximal isomorphism for every iIi\in I.

In Definition 4.6, (X,δΦ)(X,\delta_{\Phi}) is called a descriptive proximal covering space of (X,δΦ)(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}}). For iIi\in I, YiY_{i} is said to be a descriptive proximal sheet. For any xXx^{{}^{\prime}}\in X^{{}^{\prime}}, p1({x})p^{-1}(\{x^{{}^{\prime}}\}) is called a descriptive proximal fiber of xx^{{}^{\prime}}. The map p|Yi:YiYp|_{Y_{i}}:Y_{i}\rightarrow Y^{{}^{\prime}} is a descriptive proximal isomorphism if the map p:(X,δΦ)(X,δΦ)p:(X,\delta_{\Phi})\rightarrow(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}}) is a descriptive proximal isomorphism. However, the converse is not generally true. Given any descriptive proximity space (X,δΦ)(X,\delta_{\Phi}), it is obvious that the identity map on XX is always a descriptive proximal covering map.

Example 4.7.

Consider the surjective and dpc-map p:(X,δΦ)(X,δΦ)p:(X,\delta_{\Phi})\rightarrow(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}}), defined by p(ai)=p(bi)=p(ci)=dip(a_{i})=p(b_{i})=p(c_{i})=d_{i} for any i=1,2,3,4i=1,2,3,4, in Figure 3.2, where Φ\Phi is a set of probe functions that admits colors of given shapes. Let {d1}X\{d_{1}\}\subset X^{{}^{\prime}} and Y={d1,d3,d4}Y^{{}^{\prime}}=\{d_{1},d_{3},d_{4}\} a δΦ\delta_{\Phi}^{{}^{\prime}}-neighborhood of {d1}\{d_{1}\}. For V1={a1}V_{1}=\{a_{1}\}, V2={b1}V_{2}=\{b_{1}\}, and V3={c1}V_{3}=\{c_{1}\}, we have that p1(Y)=i=13Yip^{-1}(Y^{{}^{\prime}})=\displaystyle\bigcup_{i=1}^{3}Y_{i}, where Y1={a1,a3,a4}Y_{1}=\{a_{1},a_{3},a_{4}\}, Y2={b1,b3,b4}Y_{2}=\{b_{1},b_{3},b_{4}\}, and Y3={c1,c3,c4}Y_{3}=\{c_{1},c_{3},c_{4}\}. This gives us that for all i{1,2,3}i\in\{1,2,3\}, YiY_{i} is a δΦ\delta_{\Phi}-neighborhood of ViV_{i}. We also observe that YiYjY_{i}\neq Y_{j} if iji\neq j for ii, j{1,2,3}j\in\{1,2,3\}. In addition, p|Yi:YiYp|_{Y_{i}}:Y_{i}\rightarrow Y^{{}^{\prime}} is a descriptive proximal isomorphism for each ii. If one considers d3d_{3} and d4d_{4}, the same process can be repeated. Let {d2}δΦ{d2}=Y\{d_{2}\}\ll_{\delta_{\Phi}^{{}^{\prime}}}\{d_{2}\}=Y^{{}^{\prime}} in XX^{{}^{\prime}}. Then p1(Y)=Y1Y2Y3p^{-1}(Y^{{}^{\prime}})=Y_{1}\cup Y_{2}\cup Y_{3}, where Y1={a2}Y_{1}=\{a_{2}\}, Y2={b2}Y_{2}=\{b_{2}\}, and Y3={c2}Y_{3}=\{c_{2}\}. We observe that V1={a2}δΦY1V_{1}=\{a_{2}\}\ll_{\delta_{\Phi}}Y_{1}, V2={b2}δΦY2V_{2}=\{b_{2}\}\ll_{\delta_{\Phi}}Y_{2}, and V3={c2}δΦY3V_{3}=\{c_{2}\}\ll_{\delta_{\Phi}}Y_{3}. Note that Y1Y2Y3Y_{1}\neq Y_{2}\neq Y_{3}. Furthermore, p|Yi:YiYp|_{Y_{i}}:Y_{i}\rightarrow Y^{{}^{\prime}} is a descriptive proximal isomorphism for each i=1,2,3i=1,2,3. This proves that pp is a descriptive proximal covering map.

Definition 4.8.

A dpc-map p:(X,δΦ)(X,δΦ)p:(X,\delta_{\Phi})\rightarrow(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}}) is said to have the descriptive proximal homotopy lifting property (DPHLP) with respect to a descriptive proximity space (X′′,δΦ′′)(X^{{}^{\prime\prime}},\delta_{\Phi}^{{}^{\prime\prime}}) if, for an inclusion map i0:(X′′,δΦ′′)(X′′×I,δΦ1)i_{0}:(X^{{}^{\prime\prime}},\delta_{\Phi}^{{}^{\prime\prime}})\rightarrow(X^{{}^{\prime\prime}}\times I,\delta^{1}_{\Phi}) defined by i0(x′′)=(x′′,0)i_{0}(x^{{}^{\prime\prime}})=(x^{{}^{\prime\prime}},0), for every dpc-map h:(X′′,δΦ′′)(X,δΦ)h:(X^{{}^{\prime\prime}},\delta_{\Phi}^{{}^{\prime\prime}})\rightarrow(X,\delta_{\Phi}), and dprox-hom G:(X′′×I,δΦ1)(X,δΦ)G:(X^{{}^{\prime\prime}}\times I,\delta^{1}_{\Phi})\rightarrow(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}}) with ph=Gi0p\circ h=G\circ i_{0}, then there exists a dprox-hom G:(X′′×I,δΦ1)(X,δΦ)G^{{}^{\prime}}:(X^{{}^{\prime\prime}}\times I,\delta^{1}_{\Phi})\rightarrow(X,\delta_{\Phi}) for which G(x′′,0)=h(x′′)G^{{}^{\prime}}(x^{{}^{\prime\prime}},0)=h(x^{{}^{\prime\prime}}) and pG(x′′,t)=G(x′′,t)p\circ G^{{}^{\prime}}(x^{{}^{\prime\prime}},t)=G(x^{{}^{\prime\prime}},t).

X′′\textstyle{X^{{}^{\prime\prime}}\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}h\scriptstyle{h}i0\scriptstyle{i_{0}}X\textstyle{X\ignorespaces\ignorespaces\ignorespaces\ignorespaces}p\scriptstyle{p}X′′×I\textstyle{X^{\prime\prime}\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}G\scriptstyle{G}G\scriptstyle{G^{{}^{\prime}}}X.\textstyle{X^{{}^{\prime}}.}
Definition 4.9.

A map p:(X,δΦ)(X,δΦ)p:(X,\delta_{\Phi})\rightarrow(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}}), which is a dpc-map, is said to be a descriptive proximal fibration if it has the DPHLP for any descriptive proximity space (X′′,δΦ′′)(X^{{}^{\prime\prime}},\delta_{\Phi}^{{}^{\prime\prime}}).

Definition 4.10.

Given two descriptive proximity spaces (X,δΦ)(X,\delta_{\Phi}) and (X,δΦ)(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}}), a dpc-map h:XXh:X\rightarrow X^{{}^{\prime}} is said to have a dprox-hom extension property (DPHEP) with respect to a descriptive proximity space (X′′,δΦ′′)(X^{{}^{\prime\prime}},\delta_{\Phi}^{{}^{\prime\prime}}) if there exists a dprox-hom

F:(X×I,δΦ1)(X′′,δΦ′′)F^{{}^{\prime}}:(X^{{}^{\prime}}\times I,\delta_{\Phi}^{1^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta_{\Phi}^{{}^{\prime\prime}})

satisfying the conditions Fi0X=kF^{{}^{\prime}}\circ i_{0}^{X^{{}^{\prime}}}=k and F(h×1I)=FF^{{}^{\prime}}\circ(h\times 1_{I})=F for any dpc-map k:(X,δΦ)(X′′,δΦ′′)k:(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta_{\Phi}^{{}^{\prime\prime}}), and dprox-hom F:(X×I,δΦ1)(X′′,δΦ′′)F:(X\times I,\delta_{\Phi}^{1})\rightarrow(X^{{}^{\prime\prime}},\delta_{\Phi}^{{}^{\prime\prime}}) with the equality k(h×10)=Fi0Xk\circ(h\times 1_{0})=F\circ i_{0}^{X}, where the maps i0X:(X,δΦ)(X×I,δΦ1)i_{0}^{X}:(X,\delta_{\Phi})\rightarrow(X\times I,\delta^{1}_{\Phi}) and i0X:(X,δΦ)(X×I,δΦ1)i_{0}^{X^{{}^{\prime}}}:(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}})\rightarrow(X^{{}^{\prime}}\times I,\delta_{\Phi}^{1^{\prime}}) are inclusions.

X×0\textstyle{X\times 0\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}i0X\scriptstyle{i_{0}^{X}}h×10\scriptstyle{h\times 1_{0}}X×I\textstyle{X\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}h×1I\scriptstyle{h\times 1_{I}}F\scriptstyle{F}X′′\textstyle{X^{{}^{\prime\prime}}}X×0\textstyle{X^{{}^{\prime}}\times 0\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}i0X\scriptstyle{i_{0}^{X^{{}^{\prime}}}}k\scriptstyle{k}X×I.\textstyle{X^{{}^{\prime}}\times I.\ignorespaces\ignorespaces\ignorespaces\ignorespaces}F\scriptstyle{F^{{}^{\prime}}}
Definition 4.11.

A dpc-map f:(X,δΦ)(X,δΦ)f:(X,\delta_{\Phi})\rightarrow(X^{{}^{\prime}},\delta_{\Phi}^{{}^{\prime}}) is said to be a descriptive proximal cofibration if it has the DPHEP with respect to any descriptive proximity space (X′′,δΦ′′)(X^{{}^{\prime\prime}},\delta_{\Phi}^{{}^{\prime\prime}}).

Example 4.12.

Let (X′′,δΦ′′)(X^{{}^{\prime\prime}},\delta_{\Phi}^{{}^{\prime\prime}}) be a descriptive proximity space as in Figure 3.3, where Φ\Phi is a set of probe functions which admits colors of given rounds. Assume that γ1\gamma_{1} and γ2\gamma_{2} are descriptive proximal paths on X′′X^{{}^{\prime\prime}} such that γ1\gamma_{1} is a descriptive proximal path from bb to aa and γ2\gamma_{2} is a descriptive proximal path from bb to cc. Let h:({0},δΦ)(I,δΦ)h:(\{0\},\delta_{\Phi})\rightarrow(I,\delta_{\Phi}^{{}^{\prime}}) be an inclusion map. For a dpc-map k:(I,δΦ)(X′′,δΦ′′)k:(I,\delta_{\Phi}^{{}^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta_{\Phi}^{{}^{\prime\prime}}) defined as k=γ2k=\gamma_{2}, and a dprox-hom F:({0}×I,δΦ1)(X′′,δΦ′′)F:(\{0\}\times I,\delta_{\Phi}^{1})\rightarrow(X^{{}^{\prime\prime}},\delta_{\Phi}^{{}^{\prime\prime}}) defined by F(0,t)=α(t)F(0,t)=\alpha(t) for all tIt\in I with the property k(h×10)=F×i0{0}k\circ(h\times 1_{0})=F\times i_{0}^{\{0\}}, there exists a dprox-hom

F:(I×I,δΦ1)(X′′,δΦ′′)F^{{}^{\prime}}:(I\times I,\delta_{\Phi}^{1^{\prime}})\rightarrow(X^{{}^{\prime\prime}},\delta_{\Phi}^{{}^{\prime\prime}})

defined by F(0,t1)=F(0,t1)F^{{}^{\prime}}(0,t_{1})=F(0,t_{1}) and F(t2,0)=k(t2)F^{{}^{\prime}}(t_{2},0)=k(t_{2}) for all (t1,t2)I×I(t_{1},t_{2})\in I\times I which satisfy

F(h×1I)=F,\displaystyle F^{{}^{\prime}}\circ(h\times 1_{I})=F,
Fi0I=k.\displaystyle F^{{}^{\prime}}\circ i_{0}^{I}=k.

In another saying, the diagram

{0}×0\textstyle{\{0\}\times 0\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}i0{0}\scriptstyle{i_{0}^{\{0\}}}h×10\scriptstyle{h\times 1_{0}}{0}×I\textstyle{\{0\}\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}h×1I\scriptstyle{h\times 1_{I}}F\scriptstyle{F}X′′\textstyle{X^{{}^{\prime\prime}}}I×0\textstyle{I\times 0\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces\ignorespaces}i0I\scriptstyle{i_{0}^{I}}k\scriptstyle{k}I×I\textstyle{I\times I\ignorespaces\ignorespaces\ignorespaces\ignorespaces}F\scriptstyle{F^{{}^{\prime}}}

holds.

5. Conclusion

A subfield of topology called homotopy theory investigates spaces up to continuous deformation. Although homotopy theory began as a topic in algebraic topology, it is currently studied as an independent discipline. For instance, algebraic and differential nonlinear equations emerging in many engineering and scientific applications can be solved using homotopy approaches. As an example, these equations include a set of nonlinear algebraic equations that model an electrical circuit. In certain studies, the aging process of the human body is presented using the algebraic topology notion of homotopy. In addition to these examples, one can easily observe homotopy theory once more when considering the algorithmic problem of robot motion planning. In this sense, this research is planned to accelerate homotopy theory studies within proximity spaces that touch many important application areas. Moreover, this examination encourages not only homotopy theory but also homology and cohomology theory to take place within proximity spaces. The powerful concepts of algebraic topology always enrich the proximity spaces and thus it becomes possible to see the topology even at the highest level fields of science such as artificial intelligence and medicine.

Acknowledgment.  The second author is grateful to the Azerbaijan State Agrarian University for all their hospitality and generosity during his stay. This work has been supported by the Scientific and Technological Research Council of Turkey TÜBİTAK-1002-A with project number 122F454.

References

  • [1] V.A. Efremovic, The geometry of proximity I, Matematicheskii Sbornik(New Series), 31(73), 189-200, (1952).
  • [2] C. Kuratowski, Topologie. I, Panstwowe Wydawnictwo Naukowe, Warsaw, xiii+494 pp., (1958).
  • [3] S. Leader, On products of proximity spaces, Mathemathische Annalen, 154, 185-194, (1964).
  • [4] M.W. Lodato, On topologically induced generalized proximity relations, Proceedings of the American Mathematical Society, 15, 417-422, (1964).
  • [5] S.A. Naimpally, and B.D. Warrack, Proximity Spaces, Cambridge Tract in Mathematics No. 59, Cambridge University Press, Cambridge, UK, x+128 pp., Paperback (2008), MR0278261 (1970).
  • [6] S.A. Naimpally, and J.F. Peters, Topology With Applications. Topological Spaces via Near and Far, World Scientific, Singapore (2013).
  • [7] S.G. Mrowka, and W.J. Pervin, On uniform connectedness, Proceedings of the American Mathematical Society, 15(3), 446-449, (1964).
  • [8] F. Pei-Ren, Proximity on function spaces, Tsukuba Journal of Mathematics, 9(2), 289-297, (1985).
  • [9] J.F. Peters, Near sets. General theory about nearness of objects, Applied Mathematical Sciences, 1(53), 2609-2629, (2007).
  • [10] J.F. Peters, Near sets. Special theory about nearness of objects, Fundamenta Informaticae, 75(1-4), 407-433, (2007).
  • [11] J.F. Peters, and S.A. Naimpally, Applications of near sets, American Mathematical Society Notices, 59(4), 536-542, (2012).
  • [12] J.F. Peters, Near sets: An introduction, Mathematics in Computer Science, 7(1), 3-9, (2013).
  • [13] J.F. Peters, Topology of Digital Images: Visual Pattern Discovery in Proximity Spaces (Vol. 63), Springer Science & Business Media (2014).
  • [14] J.F. Peters, and T. Vergili, Good Coverings of Proximal Alexandrov Spaces. Homotopic Cycles in Jordan Curve Theorem Extension., arXiv preprint arXiv:2108.10113 (2021).
  • [15] J.F. Peters, and T. Vergili, Descriptive Proximal Homotopy. Properties and Relations., arXiv preprint arXiv:2104.05601v1 (2021).
  • [16] J.F. Peters, and T. Vergili, Good Coverings of Proximal Alexandrov Spaces. Path Cycles in the Extension of the Mitsuishi-Yamaguchi good covering and Jordan Curve Theorems., Applied General Topology, 24(1), 25-45, (2023).
  • [17] J.F. Peters, and T. Vergili, Proximity Space Categories. Results For Proximal Lyusternik-Schnirel’man, Csaszar And Bornology Categories., submitted to Afrika Matematika (2022).
  • [18] H. Poincare, Analysis Situs, Paris, France: Gauthier-Villars (1895).
  • [19] B. Riemann, Grundlagen für eine allgemeine Theorie der Functionen einer veränderlichen complexen Grösse, Huth, (1851).
  • [20] F. Riesz, Stetigkeitsbegriff und abstrakte mengenlehre, Atti del IV Congresso Internazionale dei Matematici II, 18-24pp, (1908).
  • [21] Y.M. Smirnov, On proximity spaces, Matematicheskii Sbornik(New Series), 31(73), 543-574, (1952). English Translation: American Mathematical Society Translations: Series 2, 38, 5-35, (1964).
  • [22] A.D. Wallace, Separation Space, Annals of Mathematics, 687-697, (1941).