• 検索結果がありません。

Maximal Objects and Minimal Objects in the Sets with Operations

N/A
N/A
Protected

Academic year: 2021

シェア "Maximal Objects and Minimal Objects in the Sets with Operations"

Copied!
7
0
0

読み込み中.... (全文を見る)

全文

(1)

Maximal Objects and Minimal Objects in the Sets with Operations

Fumie Nakaoka

1)

and Nobuyuki Oda

1)

(Received November 25, 2014)

Abstract

Maximal objects and minimal objects in families of subsets are studied by imposing axioms on the families to generalize some common properties of maximal open sets and maximal closed sets, dually those of minimal open sets and minimal closed sets, in topological spaces. Sets with operationsκon families of subsets are studied as examples, and some properties of maximalκ-open sets and minimalκ-closed sets are obtained.

Introduction

In a series of papers [5, 6, 7] we studied some outstanding properties of minimal open sets and maximal open sets and their duals, namely maximal closed sets and minimal closed sets.

Recently, various types of maximal open sets and minimal closed sets and their duals are studied by many mathematicians, for example, [15, 14, 1]. Therefore, it seems reasonable to study axiomatic approach to these objects which appear in many fields. The purpose of this paper is to formulate some of the results obtained in [5, 6, 7] imposing an axiom on a family S ⊂ P(X), where P(X) is the power set of a set X.

In Section 1 we consider any family S ⊂ P(X) and define maximal objects and minimal objects in S. Here, any set A ∈ S is called an object of S. To generalize some of the results in [5, 6, 7], we consider the following axioms for S which state that S is closed under finite unions and finite intersections, respectively:

Axiom S(FU) If U, V ∈ S, thenU ∪V ∈ S. Axiom S(FI) If U, V ∈ S, thenU ∩V ∈ S.

These axioms are, for example, those of inf (sup) semilattice in Definition O-1.8 of Gierz et al. [2], or join (meet) semilattice in Section 3.1 of Wood [16]. We prove key Lemmas 1.4 and 1.14 with Axioms S(FU) and S(FI), respectively.

Let F be any family of subsets of a set X such that F contains at least one non-empty set. An operation κ on F is a function κ : F → P(X) such that U Uκ for each U ∈ F, where Uκ =κ(U) (see Section 2). If κ:F → P(X) is an operation, we call a triple (X,F, κ) a space. If (X,F, κ) is a space, we callX a (F, κ)-space or a κ-space [12]. Letκ :F → P(X) be an operation. A subset A of X is called a κ-open set of X if for each x A there exists a set U ∈ F such that x U Uκ A; a subset C of X is called a κ-closed set of X if its complement X−C is a κ-open set in X (Definition 2.2).

1) Department of Applied Mathematics, Faculty of Science, Fukuoka University, 8-19-1 Nanakuma, Jonan- ku, Fukuoka, 814-0180, Japan

email address: [email protected]; [email protected]

(2)

In Section 2 we consider the families ofκ-open sets and κ-closed sets in a κ-space. Then it is possible to define maximal κ-open sets and minimal κ-closed sets. Since any union of κ-open sets is a κ-open set (see Section 2), the family of κ-open sets and the family of κ- closed sets satisfy Axioms S(FU) andS(FI), respectively. Therefore some results for maximal κ-open sets are obtained by setting S = Fκ for the family S in Section 1.1, where Fκ is the family of allκ-open sets; and the dual results for minimal κ-closed sets are obtained by setting S ={F | X−F ∈ Fκ} for the family S in Section 1.2. We see that intersections of (finite) κ-open sets are not necessarily κ-open sets, and the results corresponding to Lemmas 1.4 and 1.14 are not proved for maximal κ-closed sets and minimal κ-open sets in general; therefore, these properties of maximal κ-closed sets and minimal κ-open sets are not obtained by the general results in Section 1.

If κ : F → P(X) is regular (Definition 2.5), then A1 ∩A2 ∈ Fκ for any A1, A2 ∈ Fκ (Proposition 2.6). Therefore, if ∪F = X and κ : F → P(X) is a regular operation, then Fκ

is a topology on X, and hence the results in [5, 6, 7] hold forFκ.

1 Maximal objects and minimal objects

In this section we fix a family S ⊂ P(X) to consider maximal objects and minimal objects in S imposing Axiom S(FU) and Axiom S(FI) on S in Sections 1.1 and 1.2, respectively. If a subset A of X satisfies A ∈ S, then A is called an object of S. If A ̸= X, then A is called a proper object; if =, then A is called a non-emptyobject.

1.1 Maximal objects in S .

Definition 1.1. A proper non-empty object U ∈ S is called a maximal object in S if any object in S which contains U isX or U.

We note that in Definition 1.1 we need not assume thatX ∈ S. Example 1.2. Let X ={a, b, c}.

(1) If S ={∅,{a, b}, X}, then {a, b} is a maximal object in S. (2) If S ={∅,{a, b}}, then {a, b} is a maximal object in S. (3) If S ={{a, b}, X}, then {a, b} is a maximal object in S. (4) If S ={∅, X}, then there is no maximal object in S.

Example 1.3. Let [0,2] ={x | 0≤x≤ 2}, U ={x | 0< x < 2}, Un ={x | 1n < x <2} (for any positive integer n) be intervals on the real line. Let S ={Un |n 1} ∪ {U}. (1) If S is regarded asS ⊂ P(X) for X = [0,2], thenU is a maximal object in S.

(2) If S is regarded as S ⊂ P(X) for X =U, then U is nota maximal object in S and there exists no maximal object in S.

To study some properties of maximal objects in S, we consider the following axiom for S which states that S is closed under finite unions:

Axiom S(FU) If U, V ∈ S, thenU ∪V ∈ S.

The following result is obtained by Definition 1.1.

Lemma 1.4. If S satisfies Axiom S(FU), then the following results hold.

(1) If U is a maximal object in S and W ∈ S, then U ∪W =X or W ⊂U. (2) If U and V are maximal objects in S, then U ∪V =X or U =V.

(3)

Example 1.5. Let U ={a, b},V ={b, c}, W ={a, c}, X ={a, b, c} and Y ={a, b, c, d}. (1) If we regardS ={U, V, W} ⊂ P(X), then U,V, W are maximal objects in S andS does not satisfy Axiom S(FU).

(2) If we regard S ={U, V, W, X} ⊂ P(X), then U, V, W are maximal objects in S and S satisfies Axiom S(FU).

(3) If we regard S ={U, V, W, X} ⊂ P(Y), then X is a maximal object in S and S satisfies Axiom S(FU).

(4) If we regard S = {U, V} ⊂ P(Y), then U, V are maximal objects in S and S does not satisfy Axiom S(FU). We see U∪V ={a, b, c} ̸=Y and U ̸=V.

Theorem 1.6. Assume that S satisfies Axiom S(FU). Let U and Uλ be maximal objects in S for any element λ of Λ and U ̸=Uλ for any element λ of Λ. Then X − ∩λΛUλ U and hence λ∈ΛUλ ̸=∅.

Proof. SinceX =λ∈Λ(U∪Uλ) = U∪(λ∈ΛUλ) by Lemma 1.4(2), we haveX− ∩λ∈ΛUλ ⊂U. If λΛUλ =, then we have X =U, which contradicts our assumption that U is a maximal object in S and hence λΛUλ ̸=.

Corollary 1.7. Assume that S satisfiesAxiom S(FU). Let Uλ and Uν be maximal objects in S for any element λ of Λ andν of N. If there exists an element ν of N such that Uλ ̸=Uν for any element λ of Λ, then λΛUλ ̸⊂ ∩νNUν.

Proof. Let ν N be an element such that Uλ ̸= Uν for any element λ of Λ. If λΛUλ

νNUν, thenλΛUλ ⊂Uν. It follows thatX = (λΛUλ)∪Uν ⊂Uν by Theorem 1.6, which contradicts our assumption that Uν is a maximal object in S.

Corollary 1.8. Assume that S satisfies Axiom S(FU). Let Uλ be a maximal object in S for any element λ of Λ and Uλ ̸=Uµ for any elements λ and µ of Λ with λ̸=µ. If N is a proper non-empty subset of Λ, then

(1) (λΛ\NUλ)(νNUν) = X, where Λ\N is the difference of index sets.

(2) λ∈ΛUλν∈NUν.

Proof. (1) is obtained by Theorem 1.6.

(2) Ifν Λ\N, thenUν(λΛUλ) =Uν and Uν(νNUν) =X by Theorem 1.6. It follows then that if λΛUλ =νNUν, then X =Uν, which contradicts our assumption that Uν is a maximal object.

Corollary 1.9. Assume that S satisfiesAxiom S(FU). Let Uα, Uβ, Uγ be maximal objects in S such that Uα ̸=Uβ. If Uα∩Uβ ⊂Uγ, then Uα =Uγ or Uβ =Uγ.

Proof. Set Λ ={α, β} and N ={γ} in Corollary 1.7, then the result follows.

Corollary 1.10. Assume that S satisfiesAxiomS(FU). IfUα, Uβ, Uγ are maximal objects in S which are different from each other, then Uα∩Uβ ̸⊂Uα∩Uγ.

Proof. Set Λ ={α, β} and N ={α, γ} in Corollary 1.7, then the result follows.

We denote by |Λ| the cardinality of the index set Λ.

Theorem 1.11 (Decomposition theorem for maximal objects in S). Assume that S satisfies AxiomS(FU). Assume that |Λ| ≥2 and letUλ be a maximal object in S for any elementλ of Λ and Uλ ̸=Uµ for any elements λ and µ of Λ with λ ̸=µ. Then for any element µ of Λ,

Uµ = (λΛUλ)(X− ∩λΛ\{µ}Uλ).

(4)

Proof. Since λΛUλ = (λΛ\{µ}Uλ)∩Uµ, we have

(λΛUλ)(X− ∩λ∈Λ\{µ}Uλ) = Uµ(X− ∩λ∈Λ\{µ}Uλ) =Uµ

by Theorem 1.6.

Theorem 1.12. Assume that S satisfiesAxiom S(FU). Assume that |Λ| ≥2 and let Uλ be a maximal object in S for any element λ of Λ and Uλ ̸=Uµ for any elements λ and µof Λ with λ̸=µ. If∩λΛUλ =∅, then {Uλ | λ∈Λ} is the set of all maximal objects in S.

Proof. If there exists another maximal object Uν in S which is not equal to Uλ for any el- ement λ of Λ, then = λΛUλ = λ∪{ν})\{ν}Uλ. However, by Theorem 1.6, we have

λ∪{ν})\{ν}Uλ ̸=, which contradicts our assumption.

1.2 Minimal objects in S .

Definition 1.13. A proper non-empty object F ∈ S is called a minimal object in S if any object in S which is contained in F is orF.

We need not assume that∅ ∈ S in Definition 1.13. We see that F ∈ S is a minimal object inS if and only ifX−F is a maximal object inX− S ={X−F | F ∈ S}. Hence the proofs of the statements in this section are obtained by dual arguments of Section 1.1 and they are omitted.

We consider the following axiom onS: Axiom S(FI) If U, V ∈ S, thenU ∩V ∈ S.

Lemma 1.14. If S satisfiesAxiom S(FI), then the following results hold.

(1) If F is a minimal object in S and N ∈ S, then F ∩N = or F ⊂N. (2) If F and S are minimal objects in S, then F ∩S = or F =S.

Example 1.15. Let U ={a, b},V ={b, c} and X ={a, b, c}.

(1) If we regard S = {U, V} ⊂ P(X), then U, V are minimal objects in S and S does not satisfy Axiom S(FI).

(2) If we regard S ={{b}, U, V} ⊂ P(X), then {b} is a minimal object in S and S satisfies Axiom S(FI).

(3) If we regard S ={∅,{b}, U, V} ⊂ P(X), then {b}is a minimal object in S andS satisfies Axiom S(FI) and∅ ∈ S.

Theorem 1.16. Assume that S satisfies Axiom S(FI). Let F and Fλ be minimal objects in S for any element λ of Λ and F ̸= Fλ for any element λ of Λ. Then F X − ∪λ∈ΛFλ and hence λΛFλ ̸=X.

Corollary 1.17. Assume that S satisfiesAxiom S(FI). Let Fλ and Fν be minimal objects in S for any element λ of Λ and ν of N. If there exists an element ν of N such that Fλ ̸=Fν for any element λ of Λ, then νNFν ̸⊂ ∪λΛFλ.

Corollary 1.18. Assume that S satisfies Axiom S(FI). Let Fλ be a minimal object in S for any element λ of Λ and Fλ ̸=Fµ for any elements λ and µ of Λ with λ̸=µ. If N is a proper non-empty subset of Λ, then

(1) (λΛ\NFλ)(νNFν) =∅. (2) ν∈NFνλ∈ΛFλ.

(5)

Corollary 1.19. Assume that S satisfiesAxiom S(FI). Let Fα, Fβ, Fγ be minimal objects in S such that Fα ̸=Fβ. If Fα∪Fβ ⊃Fγ, then Fα =Fγ or Fβ =Fγ.

Corollary 1.20. Assume that S satisfies Axiom S(FI). If Fα, Fβ, Fγ are minimal objects in S which are different from each other, then Fα∪Fβ ̸⊃Fα∪Fγ.

Theorem 1.21 (Recognition principle for minimal objects in S). Assume that S satisfies Axiom S(FI). Assume that |Λ| ≥2 and letFλ be a minimal object in S for any element λ of Λ and Fλ ̸=Fµ for any elements λ and µ of Λ with λ ̸=µ. Then for any element µof Λ,

Fµ= (λ∈ΛFλ)(X− ∪λ∈Λ\{µ}Fλ).

Theorem 1.22. Assume that S satisfies Axiom S(FI). Assume that |Λ| ≥ 2 and let Fλ be a minimal object in S for any element λ of Λ and Fλ ̸=Fµ for any elements λ and µ of Λ with λ̸=µ. If λΛFλ =X, then {Fλ|λ∈Λ} is the set of all minimal objects in S of X.

2 Maximal κ-open sets and minimal κ-closed sets

Let P(X) be the power set of a set X and F ⊂ P(X) such that F contains at least one non-empty set. However, we do notassume that∪F :=U∈FU =X. An operation κonF is a function

κ:F → P(X)

such that U Uκ for each U ∈ F, where Uκ = κ(U). We call a triple (X,F, κ) a space for any operation κ : F → P(X); if (X,F, κ) is a space, we call X a (F, κ)-space or a κ-space [12].

Remark 2.1. Let τ be any family of sets. Kasahara [3] defined an operation α as a function α :τ → P(∪τ) such that G ⊂Gα for any G ∈τ, where ∪τ is the union of the sets in τ. We remove the restriction ∪τ inP(∪τ) to define our operation in [12] and in this paper.

Definition 2.2. ([12]) Let κ : F → P(X) be an operation. A subset A of X is called a κ-open set of X if for each x∈ A there exists a setU ∈ F such that x ∈U ⊂Uκ ⊂A. The family of allκ-open sets is denoted by Fκ. A subset C of X is called aκ-closed set of X if its complement X−C is a κ-open set in X.

IfAλ ∈ Fκ for any λ of Λ, then λ∈ΛAλ ∈ Fκ by Proposition 2.8 of [12]. If we set S =Fκ

in Definition 1.1 for any space (X,F, κ), then a maximal object inS is amaximal κ-open set; moreover, since S =Fκ satisfies Axiom S(FU), we have the results for maximal κ-open sets by theorems and corollaries for maximal objects in Section 1.1. If S = {F | X −F ∈ Fκ} in Definition 1.13 for a space (X,F, κ), then a minimal object inS is a minimal κ-closed set.

The family S = {F | X−F ∈ Fκ} satisfies Axiom S(FI), since Fκ satisfies Axiom S(FU), and the results for minimal κ-closed sets are obtained by the results in Section 1.2.

Example 2.3. Let U be a subset of a set X such thatUX and F ={U}. (1) If κ(U) =X, then Fκ ={∅}.

(2) If κ(U) =U, then Fκ ={∅, U}.

We consider the following axiom for the set F, which states that F is closed under finite intersections and that F is closed under arbitrary unions, respectively:

(6)

Axiom F(FI) If U, V ∈ F, then U ∩V ∈ F.

Axiom F(AU) If Aλ ∈ F for any λ∈Λ, then λΛAλ ∈ F.

Then the following results are immediate consequences of the definition of theκ-open set.

Proposition 2.4. Let κ:F → P(X) be an operation. If F satisfies Axiom F(AU), then the following results hold.

(1) Fκ ⊂ F ∪ {∅}.

(2) If Uκ =U for any U ∈ F, then Fκ =F ∪ {∅}.

Definition 2.5. An operationκ:F → P(X) isregularif for anya∈X and any setsU, V ∈ F with a ∈U ∩V, there exists a set W ∈ F such that a ∈W Wκ ⊂Uκ∩Vκ. An operation κ:F → P(X) is monotone if U ⊂V and U, V ∈ F impliesUκ ⊂Vκ (cf. p.98 of [3]).

We have the following results by the arguments similar to the proofs of Proposition 2.9(1) of [13] and p.98 of [3].

Proposition 2.6. (1) If κ :F → P(X) is regular, then A1 ∩A2 ∈ Fκ for any A1, A2 ∈ Fκ. (2) If an operation κ : F → P(X) is monotone and F satisfies Axiom F(FI), then κ is regular.

IfF satisfies AxiomF(AU) and∅ ∈ F, then we haveFκ =F by Proposition 2.4(2) for the κ defined there. If ∪F =X and κ :F → P(X) is a regular operation, then Fκ is a topology on X by Proposition 2.6(1), and hence the results in [5, 6, 7] hold for minimal and maximal κ-open sets and maximal and minimal κ-closed sets.

Remark 2.7. Whenγ :τ → P(X) is the operation in Kasahara [3] for some topological space (X, τ), we studied maximalγ-open sets and its dual, minimal γ-closed sets in [4, 8, 9, 11] to generalize the results in [5, 6, 7]. Some of the results in these articles are generalized in this paper making use of the operation κ which is more general than those studied in previous articles. More precisely:

If we set S = τγ for an operation γ : τ → P(X) for some topological space (X, τ), then Lemma 1.4 implies Lemma 2.2 of [8]; Theorems 1.6, 1.11 and 1.12 imply Theorems 2.4, 2.7 and 2.8 of [11], respectively; Corollaries 1.7 and 1.8 imply Corollaries 2.5 and 2.6 of [11];

Corollaries 1.9 and 1.10 imply Theorems 2.3 and 2.4 of [8].

If γ : τ → P(X) is an operation for some topological space (X, τ) and the family S is defined by S = {F | X−F τγ}, then Lemma 1.14 implies Lemma 2.2 of [9]; Theorems 1.16, 1.21 and 1.22 imply Theorems 2.3, 2.8 and 2.9 of [9]; Corollaries 1.17 and 1.18 imply Corollaries 2.5, 2.4 and Theorem 2.6 of [9].

In the attempts to generalize these results further, we studied some aspects of maximal objects and minimal objects in topological spaces in [4, 10].

References

[1] M. Caldas, S. Jafari and S. P. Moshokoa, On some new maximal and minimal sets via θ-open sets, Commun. Korean Math. Soc.25 (2010), 623–628.

(7)

[2] G. Gierz, K. H. Hofmann, K. Keimel, J. D. Lawson, M. Mislove, D. S. Scott, Contin- uous lattices and domains, Encyclopedia Math. Appl. 93, Cambridge University Press, Cambridge, 2003.

[3] S. Kasahara, Operation-compact spaces, Math. Japon. 24 (1979), no.1, 97–105.

[4] F. Nakaoka and N. Oda, Minimal objects in locally finite spaces, The 5th Meetings on Topological Spaces Theory and its Applications, August 19–20, 2000, Yatsushiro College of Technology, 21–26.

[5] F. Nakaoka and N. Oda, Some Applications of minimal open sets, Int. J. Math. Math. Sci.

27 (2001), Issue 8, 471–476.

[6] F. Nakaoka and N. Oda, Some properties of maximal open sets, Int. J. Math. Math. Sci.

2003 (2003), Issue 21, 1331–1340.

[7] F. Nakaoka and N. Oda, Minimal closed sets and maximal closed sets, Int. J. Math. Math.

Sci. 2006 (2006) Art. ID 18647, 8 pp.

[8] F. Nakaoka and N. Oda, Maximal γ-open sets, The 10th Meetings on Topological Spaces Theory and its Applications, August 20–21, 2005, Fukuoka University Seminar House, 21–25.

[9] F. Nakaoka and N. Oda, Minimal γ-closed sets, The 11th Meetings on Topological Spaces Theory and its Applications, August 19–20, 2006, Fukuoka University Seminar House, 11–15.

[10] F. Nakaoka and N. Oda, Minimal objects and maximal objects, The 12th Meetings on Topological Spaces Theory and its Applications, August 18–19, 2007, Fukuoka University Seminar House, 11–14.

[11] F. Nakaoka and N. Oda,Maximal γ-open sets and minimal γ-closed sets, The 13th Meet- ings on Topological Spaces Theory and its Applications, August 23–24, 2008, Fukuoka University Seminar House, 35–39.

[12] F. Nakaoka and N. Oda, Interiors and closures in a set with an operation, Commun.

Korean Math. Soc. 29 (2014), 555–568.

[13] H. Ogata, Operations on topological spaces and associated topology, Math. Japon. 36 (1991), no.1, 175–184.

[14] S. Rajakumar, A. Vadivel and K. Vairamanickam, Minimal τ-g-open sets and maximal τ-g-closed sets in topological spaces, J. Adv. Stud. Topol. 3 (2012), 48–54.

[15] B. Roy and R. Sen,On maximalµ-open and minimalµ-closed sets via generalized topology, Acta Math. Hungar.136 (2012), 233–239.

[16] R. J. Wood, Ordered sets via adjunctions, Categorical foundations, 5–47, Encyclopedia Math. Appl. 97, Cambridge Univ. Press, Cambridge, 2004.

参照

関連したドキュメント

In Section 7, we classify the sets reduced decompositions which can correspond to paths between two flags in a semimodular lattice, and in Section 8, we use our results to derive

Next we show that the traces of maximal clones defined by bounded partial orders, equivalence, affine and h–regular relations are not subsets of the trace of a maximal clone defined

Denote by N q the number of non-singular points lying in PG(2, q) of an absolutely irreducible plane curve of degree

We show that every maximal clone determined by a prime affine or h-universal relation on A is contained in a family of partial clones on A of continuum cardinality.. MSC 2000:

The purpose of this paper is to prove some fundamental properties of maximal open sets and establish a part of the foundation of the theory of maximal open sets in topological

The case when the space has atoms can easily be reduced to the nonatomic case by “putting” suitable mea- surable sets into the atoms, keeping the values of f inside the atoms

In the next section, we introduce dependency graphs and the Arratia-Goldstein-Gordon version of the Chen-Stein method, and perform preliminary computations.. After these

It is easy to prove that B X (D) is a semigroup with respect to the operation of multiplication of binary relations, which is called a complete semigroup of