The notation of fuzzy set field is introduced. A fuzzy metric is redefined on fuzzy set field and on arbitrary fuzzy set in a field. The metric redefined is between fuzzy points and constitutes both fuzziness and crisp property of vector. In addition, a fuzzy magnitude of a fuzzy point in a field is defined.
1. Introduction
Different researchers introduced the concept of fuzzy field and notion of fuzzy metric on fuzzy sets. How to define a fuzzy metric on a fuzzy set is still active research topic in fuzzy set theory which is very applicable in fuzzy optimization and pattern recognition. The notion of fuzzy sets has been applied in recent years for studying sequence spaces by Tripathy and Baruah [1], Tripathy and Sarma [2], Tripathy and Borgohain [3], and others.
Wenxiang and Tu [4] introduced the concept of fuzzy field in field and fuzzy linear spaces over fuzzy field. Furthermore, different authors are attempting to define fuzzy normed linear spaces, fuzzy inner product space, fuzzy Hilbert space, fuzzy Banach spaces, and so forth (cf. [5–8]).
Many authors introduced different notion of fuzzy metric on a fuzzy set from different points of view. Kaleva and Seikkala [9] introduced the notion of a fuzzy metric space where metric was defined between fuzzy sets. The idea behind this notion was to fuzzify the classical metric by replacing real values of a metric by fuzzy values (fuzzy numbers). For the further research work and the properties of this type of fuzzy metric space see for instance Fang [10], Quan Xia and Guo [11], and others.
Wong [12] defined fuzzy point and discussed its topological properties and there after Deng [13] defined Pseudo-metric spaces where metric was defined between fuzzy points rather than between fuzzy sets. Hsu [14] introduced fuzzy metric space with metric defined between fuzzy points and examined the completion of fuzzy metric space. For different notions of fuzzy metric space and for further research work see for instance Shi [15], Shi and Zheng [16], Shi [17] and others.
This paper is an attempt to define a fuzzy set field in a field which is assumed to be the generalization of a fuzzy field introduced by [4]. We restate fuzzy set in more general form by allowing a particular fuzzy set to consist a family of membership functions. A fuzzy metric on fuzzy set and on fuzzy set field is reintroduced in such way that the classical metric is considered as a special type of fuzzy metric. In the sequel, a notion of magnitude of a fuzzy point in a field is introduced for the first time (up to our knowledge) and some of its properties are investigated.
2. Brief Summary of Fuzzy Set, Fuzzy Point, and Fuzzy FieldDefinition 1 (see [18]).
Let X be a nonempty set and μ:X→[0,1] be a mapping. Then Aμ={(α,μ(α)):α∈X} is said to be a fuzzy set in X with membership function μ.
Definition 2.
Let Aμ={(α,μ(α)):α∈K} be a fuzzy in K. Then
Aμ is called normal if there is at least one point x∈R(K=R) with μ(x)=1 (see [19]).
A fuzzy set Aμ is convex if for any x,y∈R(K=R) and any λ∈[0,1], we have μ(λx+(1-λ)y)≥minμ(x),μ(y) (see [19]).
A fuzzy number is a fuzzy set on the real line that satisfies the conditions of normality and convexity (see [19]).
Aμ is said to be fuzzy point if {α∈K:μ(α)≠0} = singleton set. It is usually denoted by xλ, (see [12, 18]), where
(1)xλ(y)={λify=x0ify≠x.
Support of Aμ is the crisp set, sprt(Aμ)={α∈K:μ(α)≠0} (see [12, 18]).
The fuzzy point xλ is said to be contained in a fuzzy set, Aμ, or to belong to Aμ, denoted by xλ∈Aμ, if and only if λ≤μ(x) (see [18]).
Definition 3 (see [20]).
Let X be a Cartesian product of universes X=X1×⋯×Xn, Aμ1,…,Aμn be n fuzzy sets in X1,…,Xn, respectively, and f be a mapping from X to a universe Y, y=f(x1,…,xn). Then the extension principle allows us to define a fuzzy set Bγ in Y given by Bγ={(y,γ(y)):y=f(x1,…,xn)}, where f-1 is inverse of f and
(2)γ(y)={supinf(x1,…,xn)∈f-1(y){μ1(x1),…,μn(xn)}iff-1(y)≠∅0otherwise.
Lemma 4 (see [21]).
Provided that M is either a positive or a negative fuzzy number and that N and P are together either positive or negative fuzzy numbers, then M⊙(N⊕P)=(M⊙N)⊕(M⊙P).
Definition 5 (see [4]).
Let F be a field and let K be a fuzzy set in F with membership function μ. Suppose the following conditions hold:
μ(x+y)≥min{τ(x),μ(y)},
μK(-x)≥τ(x),
μ(xy)≥min{τ(x),μ(y)},
μ(x-1)≥τ(x)
then K is said to be a fuzzy field in F and is denoted by (Kμ,F).
Lemma 6 (see [4, 8]).
If K is a fuzzy field in F with membership function μ, then
μ(0)≥τ(x), x∈F,
μ(1)≥τ(x), x≠0,
τ(x)=μ(-x), x∈F,
μ(x-1)=τ(x), x≠0.
3. Fuzzy Set Field and Fuzzy Metric on a Fuzzy Set in Fuzzy Set Field3.1. Fuzzy Set Field
In this section the concept of fuzzy set field is introduced which is assumed to generalize a fuzzy field defined by [4]. Since we are dealing with a subset of original field, the axioms imposed on it are similar to that of the ordinary field. However, the operations (fuzzy sum and fuzzy product) are performed according to the extension principle.
Notation 1.
The following notations might be used
Every mapping from a a non-empty set K into the unit interval, [0;1] is denoted by small Greek letters;
I=[0;1]; IK={τ:τ:K[0;1]};K=∅;
xτ=(x,τ(x)), we mean a fuzzy point x with membership value τ(x)≠0;
Aγ={xτ:x∈K,τ(x)≠0,τ∈γ⊂IK};
xτ=γη, we mean x=y (in usual sense) and τ(x)=η(y).
Definition 7.
Let K be a field and let μ:K→[0,1] be mapping. A point x is said to be a fuzzy point of K if τ(x)≠0 and τ(x) is said to be membership value of a fuzzy point x.
Definition 8.
Let γ⊂IK.Aγ={xν:x∈Kforsome,ν∈γ,ν(x)≠0} is said to be a fuzzy set in K.
In Definition 8, if γ=singletonset, then Aγ is said to be ordinary fuzzy set on K.
Notation 2.
Let Aγ be a fuzzy set in a field F. A fuzzy point x belongs to a fuzzy Aγ (or x∈Aγ); we mean that there exists τ∈γ such that xτ∈Aγ. Therefore, we will use x∈Aγ and xτ∈Aγ alternatively.
Definition 9.
Two fuzzy points xμ and yν are said to be equal if and only if x=y (in usual sense) and μ(x)=ν(y) and we denote it by xμ=yν.
Using Definition 3 (that is extension principle), we define operations fuzzy product, fuzzy sum, and fuzzy difference between fuzzy points according to the following definition.
Definition 10.
Let xτ and yη be fuzzy points in a field K. Then the fuzzy sum, fuzzy product, and fuzzy difference of fuzzy points are defined by
xτ⊕yη:=(x+y,τ(x)⊕η(y)), where τ(x)⊕η(y)=supinfa=x+y{τ(x),η(y)}.
xτ⊙yη:=(xy,τ(x)⊙η(y)), where τ(x)⊙η(y)=supinfa=xy{τ(x),η(y)}.
xτ⊖yη:=(x-y,τ(x)⊖η(y)), where τ(x)⊖η(y)=supinfa=xy{τ(x),η(y)}.
Remark 11.
Let K be a fuzzy set in a field K and let xμ;yν∈Ky. Then
τ(x±y)≥min{μ(x),ν(y)};
τ(xy)≥min{μ(x),ν(y)};
τ(x-y)=τ(y-x) for all x, y∈K.
Definition 12.
Let Aγ be a fuzzy set in K. Fuzzy points w and l in Aγ with respective membership values ρ(w) and ψ(l) are said to be fuzzy additive and fuzzy multiplicative points, respectively, in Aγ if they satisfy the following conditions:
for each x∈Aγ, xτ⊕wρ=xτ=wρ⊕xτ;
for each x∈Aγ, xτ⊙lψ=xτ=lψ⊙xτ.
Definition 13.
Let Aμ be an ordinary fuzzy set in a field F with membership function μ. Let γμ={ν:ν=μorcombinationsofμcombinedby,⊕orboth}. A fuzzy set, Kγμ={xν:ν∈γμ,x∈F}, is said to be a generated fuzzy set in F with a generating membership function μ denoted by (Kγμ,F). Alternatively, we call (Kγμ,F) the collection of all fuzzy points generated by μ.
Theorem 14.
Let Kγμ be a generated fuzzy set in F. Then for all x,y,z∈Kγμ
xτ⊕yη∈Kγμ, xτ⊕yη=yη⊕xτ, and (xτ⊕yη)⊕zκ=xτ⊕(yη⊕zκ);
xτ⊙yη∈Kγμ, xτ⊙yη=yη⊙xτ, and (xτ⊙yη)⊙zκ=xτ⊙(yη⊙kκ).
Proof.
It follows from Definitions 10 and 13.
Theorem 15.
Let (Kμ,F) be a fuzzy field
(3)ρ(x)={μ(0)ifx=00ifx≠0,ψ(x)={μ(1)ifx=10ifx≠1,γ′=γμ∪{ρ,ψ},
where γμ is as in Definition 13. If x, y, z∈Kγ′ with respective membership values λ(x), η(y), and κ(z), then
xτ⊕yη=yη⊕xτ;
xτ⊙yη=yη⊙xτ;
xτ⊕(yη⊕zκ)=(xτ⊕yη)⊕zκ;
xτ⊙(yη⊙zκ)=(xτ⊙yη)⊙zκ;
0ρ⊕xτ=xτ=xτ⊕0ρ;
1ψ⊙xτ=xτ=xτ⊙1ψ;
for each x∈Kγ′, there exists x′∈Kγ′ such that xτ′′⊕xτ=0ρ=xτ⊕xτ′′;
given 0≠x∈Kγ′, there exists an x′∈Kγ′ such that xτ′′⊙xτ=1μl=xτ⊙xτ′′.
Proof.
The results follow from Lemma 6 and Definition 10.
According to Theorem 15, if (Kμ,F) is a fuzzy field, γμ′={η:η∈γμorη∈{ρ,μl}}, then (Kγμ′,F) represents almost a field over binary operations (·) and (⊕). It is clear that if τ(x)⊙(μ(y)⊕μ(z))=(τ(x)⊙μ(y))⊕(τ(x)⊙μ(y)) for all x,y,z∈Kγμ′, then (Kγμ′,F) satisfies most axioms of a field under the binary operations ⊕ and ⊙.
With motivation of Theorem 15 and discussion that follows it, we have the following definition.
Definition 16.
Let K be a field and let γ⊂IK. A fuzzy set Kγ is said to be a fuzzy set field in K if the binary operators ⊕:Kγ×Kγ→Kγ and ⊙:Kγ×Kγ→Kγ satisfy the following conditions:
xτ⊕yη=yη⊕xτ, ∀x,y∈Kγ, and for all τ, η∈γ;
xτ⊕(yη⊕zκ)=(xτ⊕yη)⊕zκ, ∀x,y,z∈Kγ, and for all τ,η,κ∈γ;
xτ⊙yη=yη⊙xτ, ∀x,y∈Kγ, and for all τ,η∈γ;
xτ⊙(yη⊕zk)=(xτ⊙yη)⊙zκ, ∀x,y,z∈Kγ, and for all τ, η, κ∈γ;
there exists unique θ∈Kγ with θρ such that θρ⊕xτ=xτ=xτ⊕θρ, ∀x∈Kγ, and for all τ∈γ;
for each x∈Kγ, there exists x′∈Kγ such that xτ′⊕xτ=θρ=xτ⊕xτ′′ and for some τ, τ′∈γ;
xτ⊙(yη⊕zk)=(xτ⊙yη)⊕(xτ⊙zk), ∀x,y,z∈Kγ;
there exists l∈Kγ with lψ such that lψ⊙xτ=xτ=xτ⊙lψ;
given x≠θ and x∈Kγ, there exists an x′∈Kγ such that xτ′′⊙xτ=lμ=xτ⊙xτ′′.
We call θ, l∈Kγ the fuzzy additive identity and fuzzy unit points of Kγ, respectively.
Theorem 17.
Let Kγ be a fuzzy set field in K. Then
θ∈K is the fuzzy additive identity point of Kγ if and only if θ=0, there exists ρ∈γ such that λ(x)=supinfx=a+b{λ(a),ρ(b)}, ∀x∈K, ∀λ∈γ;
l∈K is the fuzzy multiplicative identity point of Kγ if and only if l=1 and there exists ψ∈γ such that ∀x∈Kγ∀λ∈γ, λ(x)=supinfx=ab{λ(a),μl(b)};
0ρ⊙xτ=0ρ=xτ⊙0ρ, where ρ is as in (i);
xτ⊙yη=0ρ⇒xτ=0ρ or yη=0ρ.
Proof.
Let K be a fuzzy set field in a field K, let τ,η,λ∈γ be arbitrary and let ρ,ψ∈γ be as in Definitions 16 and 21 respectively. We need to show that (i) to (iv) of the theorem.
Let w be the fuzzy additive identity point of Kγ with membership ρ(w). Then we claim that w=0. For if w≠0, then wρ⊕xλ=(x+w,η(x+w))≠(x,λ(x)), hence the claim. Now let x∈Kγμ be arbitrary. Then xλ⊕wρ = (w+x,η(w+x))=(x,λ(x))⇒η(0+x) = λ(x)=supinfx=a+b{λ(a),ρ(b)}. The converse is clear.
Let l be as stated in Definition 16 with membership value ψ(l). Let x∈Kγ with corresponding membership value τ(x). Then xτ⊙1ψ=(bx,η(bx))=(x,λ(x)). Thus l=1 and xτ⊙1ψ=(x,η(x))=(x,λ(x))⇒η(x)=λ(x)=supinfx=ab{λ(a),ψ(b)}. The converse holds trivially.
Let x,y∈Kγ be arbitrary with membership values λ(x) and ν(y), respectively. Then by (5) and (7), we have xτ⊙yη=xτ⊙(yη⊕0ρ)=(xτ⊙yη)⊕(xτ⊙0ρ)=(xη⊙0ρ)⊕(xτ⊙yη)⇒xτ⊙0ρ=0ρ by uniqueness of 0 in Kγ.
The proof of (iv) follows from (iii). Therefore, the theorem is proved.
Remark 18.
If Kγ is a fuzzy field set in a field K and ρ is as stated in Theorem 17, then for each τ∈γ, either τ(0)=ρ(0) or τ(0)=0.
Theorem 19.
Let Kγ be a fuzzy set field in a field K.
If there exists a ρ∈γ such that ρ(0)≥τ(x) for all x∈K and supinfx=a+b,b≠0{τ(a),ρ(b)}≤min{τ(x),ρ(0)} for all τ∈γ, then 0∈Kγ with membership value ρ(0) is the additive element of Kγ.
If there exists a ψ∈γ such that ψ(1)≥τ(x) for all 0≠x∈K and supinfx=ab,b≠1{τ(a),ψ(b)}≤min{τ(x),ψ(1)} for all τ∈γ, then 1∈Kγ with membership value ψ(1) is the multiplicative element of Kγ.
For ongoing discussion, we keep the notations ρ and ψ as they are stated in Theorems 17 and 19.
Remark 20.
If Kγ is a fuzzy set field in K, μ, τ, k∈γ and ρ,ψ∈γ be as stated in Theorem 19, then
ρ(0)≥ρ(x), x∈K.
ρ(0)=supx∈Kτ(x) for each τ∈γ.
xτ⊖xτ=0ρ for each x∈Kγ.
ψ(1)≥ψ(x),x∈K.
ψ(1)=sup0≠x∈Kτ(x) for each τ∈γ.
τ(x)≠0⇒η(x-1)≠0, x≠0 for some η∈γ.
τ(x)⊙(μ(y)⊕k(z))=(τ(x)⊙k(y))⊕(τ(x)⊙k(z)) for all x,y,z∈K.
Definition 21.
Let F be a field. A function μ:F→[0,1] is said to have property (Du) if it satisfies the following conditions:
μ(0)≥μ(x)∀x∈F,0≠μ(1)≥μ(x)∀x∈F, x≠0, and {x∈F:τ(x)≠0}, is a subfield.
Let Aμ be an ordinary fuzzy set in a field F, let γμ be as in Definition 13. If μ has property (Du),
(4)ρ(x)={μ(0)ifx=00ifx≠0,ψ(x)={μ(1)ifx=10ifx≠1;
then Fω={xv:v∈ω} is a fuzzy set field.
Proof.
Definition 16 ((1)–(4)) follows from Theorem 14 and Definition 16 (7) follows from Definition 21. Therefore, we need to verify (5), (6), (8), and (9) of Definition 16. From Definition 21 and definition of Fω, for all τ∈ω, we can easily verify that
ρ(0)≥τ(x), x∈F,
ρ(0)=τ(0),
τ(1)≥τ(x), x∈F,x≠0,
ψ(1)=τ(1).
Using (i) to (iv) and definition of ψ and ρ, we have for each λ∈ω,λ(x)≤ρ(0), we have λ(x)≤ψ(1), ρ(x)≤λ(x), x≠0, and ψ(x)≤λ(x), x≠1. Consequently, supinfx=a+b,b≠0{λ(a),ρ(b)}≤min{λ(x),ρ(0)}, and supinfx=ab,b≠1{λ(a),ψ(b)}≤min{λ(x),ψ(1)},∀x∈K,∀λ∈γ. Therefore, by Theorem 19, θ=0 and l=1 with respective membership values ρ(0) and ψ(1) are fuzzy zero and fuzzy unit of Kω, respectively. Hence ((5) and (8)) of Definition 16 are proved. Since K={x∈F:τ(x)≠0} is subfield, given x∈K, x≠0, there exists y∈K such that xy=1 and xτ⊙yλ=(1,η(1)). Using ((iii) and (iv)) of this theorem, η(1)=supinf1=xy{τ(x),ν(y)}=ψ(1). Thus xτ⊙yν=lψ=yν⊙xτ. Similarly, given x∈K, there is a unique y∈K such that xτ⊕yη=0ρ=yη⊕xτ, hence the theorem.
Theorem 23.
Let F and K be fields and let Fγ be fuzzy set field of F. If there exists ψ,ρ∈γ as stated in Theorem 19 and f:F→K is a homomorphism, then f(Fγ) is a fuzzy set field in K.
Proof.
Let 1 and 1′ be units (multiplicative) of F and K, respectively, 0 and 0′ be neutral elements with respect to addition (+) of F and K, respectively and M=f(F). Since f is homomorphism, we have f(1)=1′ and f(0)=0. Since Fγ is fuzzy set in F, Mω is a corresponding fuzzy set in K with membership function(s) given by ωτ(y)=sup{τ(x):y=f(x),τ∈γ} [20].
Now, we shall verify Definition 16 ((1) to (9)).
Let a,b∈Mω with membership values ωμ(a) and ων(a), respectively. Then aωμ⊕bων = (a+b,ωτ(a+b)) and there are x,y∈F such that f(x) = a,f(y)=b. But
(5)ωτ(a+b)=sup{τ(x+y):a+b=f(x+y)}=sup{τ(y+x):b+a=f(y+x)}=ωτ(a+b).
Thus aωμ⊕bων=bων⊕aωμ.
Similar to proof of (1), we can easily verify that (aωμ⊕bων)⊕cωκ=aωμ⊕(bων⊕cωκ).
Suppose a,b∈Mω. Since a,b∈M, then there are fuzzy points x,y∈F such that f(x) = a and f(y)=b. Since f is a homomorphism, ab=f(xy)∈M. So aωμ⊙bων = (ab,ω(ab))=(ab,sup{τ(xy):ab=f(xy)})∈Mω. But sup{τ(xy):ab = f(xy)}=sup{τ(yx):ba = f(yx)}=ω(ba). Thus, aωμ⊙bων=bων⊙aωμ.
Similar to the proof of (3), we show that (aωμ⊙bων)⊙cωκ=aωμ⊙(bων⊙cωκ).
Let y∈Mω. Since f(0)=0′,0′∈M and by Theorem 17, for every x∈Fγ with corresponding membership value τ(x), there exists ρ∈γ such that τ(x) = supinfx=a+b{τ(a),ρ(b)}. Since ωτ(y) = sup{τ(x):y=f(x)}, ∀y∈M, we have supinfy=y1+y2(ωτ(y1),ωρ(y2)) = supinfy=y1+y2(supy1=f(a)τ(a),supy2=f(b)ρ(b)) = supinfy=y1+y2{supx=a+b,y=f(x)(τ(a),ρ(b))} = supx,y=f(x){supinfx=a+b,y1=f(a),y2=f(b)(τ(a),ρ(b))}=supx,y=f(x)τ(x) = ωτ(y). Therefore, by Theorem 17, 0′=0 with membership value ωρ(0′) is the fuzzy additive identity point of Mω and for every y∈Mω, 0ωρ′⊕yωτ=yωτ⊕0ωρ′=yωτ.
Let a∈Mω. Since a∈M, there exist x,y∈Fγ such that f(x)=a and xτ⊕yν=0ρ, by Definition 16. But f(x+y)=f(x)+f(y)=a+b=0′, where f(y)=b, and aωμ⊕bωv=(0,ωτ(0′)). Since x⊕y = 0ρ and ρ(0)=ωρ(0) (be extension principle followed by Remak 20), it follows that ωτ(0=a+b) = sup{τ(x+y):0=f(x+y)}=ρ(0)=ωρ(0).
Thus, aωμ⊕bων=0ωρ′=bων⊕aωμ. Since a∈Mω was arbitrary, so given y∈Mω there exists z such that yων⊕zωκ=0ωρ′=zωκ⊕yων.
Similar to the proof of (3) above, we show that aωμ⊙(bων⊕cωκ)=(aωμ⊙bων)⊕(aωμ⊙cωκ).
Since Fγ is a fuzzy field set, by Theorem 17, for every x∈Fγ with corresponding membership value τ(x), there exists ψ∈γ such that τ(x)=supinfx=ab{τ(a),ψ(b)}. Since ωτ(y)=sup{τ(x):y=f(x)},∀y∈M, we have supinfy=y1y2(ωτ(y1),ωψ(y2)) = supinfy=y1y2(supy1=f(a)τ(a),supy2=f(b)ψ(b)) = supinfy=y1y2{supx=ab,y=f(x)(τ(a),ψ(b))} = supx,y=f(x){supinfx=ab,y1=f(a),y2=f(b)(τ(a),ψ(b))} = supx,y=f(x)τ(x)=ωτ(y). Therefore, by Theorem 17, l′=1′ with membership value ωψ(1′) is the fuzzy multiplicative identity point of Mω and for every y∈Mω,1ωψ′⊕yωτ=yωτ⊕1ωψ′=yωτ.
Let a∈Mω,a≠0′. Since a∈M, there exist x,y∈Fγ such that f(x)=a and xτ⊙yη=1ψ and f(xy) = f(x)f(y)=ab=1′, where f(y)=b. Thus a⊙b=(1′,ωτ(1′)). Since xt⊙yη=lψ and ψ(1)=ωψ(1′) (by extension prnciple followed by Remark 20), it follows that ωτ(1′) = sup{τ(xy):1′=f(xy)}=ψ(1)=ωψ(1′). Thus, aωμ⊙bων=1ωψ′=bων⊙aωμ.
Corollary 24.
Let F, K be fields, and let Kω be a fuzzy set field in K. If f:F→K is injective homomorphism, then f-1(Kω) is a fuzzy set field in F.
Example 25.
Let F be a field and let τ:F→[0,1] be given by τ(x)=1∀x∈F. Then Kτ={xτ:τ(x)≠0}=F×{1} is a fuzzy set field in F.
Proof.
(1)–(9) of Definition 16 hold trivially, with 0ρ=(0,1)=0τ and 1ψ=(1,1)=1τ.
Example 26.
Let F=R and let μ:R→[0,1] be given by
(6)τ(x)={11+|x|if-12≤x≤1211+|x+1|ifx≤-1211+|x-1|ifx≥12,ρ(x)={μ(0)=1ifx=00ifx≠0,ψ(x)={μ(1)=1ifx=10ifx≠1.
Let γμ be as in Definition 13 and ω={η:η∈γμ or η∈{ρ,ψ}}, then Fω={xv:v∈ω} is a fuzzy set field in R.
Proof.
First we show that μ has property (Du). Clearly, μ satisfies Definition 21(i). To verify Definition 21(ii), Lemma 4 will be applied. Since τ(x)=μ(-x) for all x∈R, we can consider only for x≥0. Moreover, Aμ=Aμ1∪Aμ2, where
(7)μ1(x)={11+xifx≥00ifx<0,μ2(x)={11+|x-1|ifx≥00ifx<0,
and τ(x)=max{μ1(x),μ2(x)} [20]. Therefore, μ1 and μ2 are both positive fuzzy numbers. Hence, by Lemma 4, μj⊙(μi⊕μj) = (μj⊙μi)⊕(μj⊙μj)i,j=1,2,…. Thus, τ(x)⊙(μ(y)⊕μ(z)) = (τ(x)⊙μ(y))⊕(τ(x)⊙μ(z))∀x,y,z∈F. Therefore, applying Theorem 22 the result follows.
3.2. Fuzzy Metric of a Fuzzy Set and Fuzzy Magnitude of a Fuzzy Point in a Fuzzy Set Field
Now we introduce definition of a fuzzy magnitude of a fuzzy point and fuzzy metric on a fuzzy set in a fuzzy set field, Kγ, such that ρ∈γ satisfies (1) in Theorem 19. For a fuzzy set field Kγ, such that ρ∈γ satisfies (1) in Theorem 19, it follows that sup{τ(x):x∈K}=ρ(0) (Remark 20). That is τ(xo)≤ρ(x) for all τ∈γ and for each x∈K. Therefore, we assume that any fuzzy point, xτ with a membership value τ(x)=ρ(0) (that is τ(x)=ρ(0)) as a nonfuzzy point of a particular fuzzy set field under discussion. Hence we have the following definition.
Definition 27.
Let Kγ be a fuzzy field set of a field K as in Theorem 19, and let τ∈γ. A fuzzy point x with membership value τ(x) is said to be nonfuzzy point of a fuzzy set field Kγ if and only if τ(x)=ρ(0).
Definition 28.
Let Kγ be a fuzzy set field of a field K, μ∈γ, and let x be a fuzzy point with membership value μ(x). Then the fuzzy magnitude of xτ, denoted by |xτ|F, is given by |xτ|F=|x|+ρ(0)-τ(x). Where ρ is as in Theorem 19.
From Definition 28, |x|F=|x| if and only x is nonfuzzy in Kγ provided that ρ∈γ is as in Theorem 19. Therefore one can redefine the magnitude of given vector in a field as a fuzzy magnitude. Moreover, |x|F=0 if and only if x=0 and nonfuzzy and |x-y|F=0 if and only if x=y and x-y is nonfuzzy. Furthermore, |xτ|F→|x| whenever τ(x)→ρ(0).
Theorem 29.
Let xτ, yν, and zk be fuzzy points in a fuzzy set field Kγ. Then
Let Kγ be a fuzzy set field of a field K such that ρ∈γ is as in Theorem 19, μ⊂γ and Aμ be a nonempty fuzzy set in field K. A fuzzy metric on A is a mapping dF:Aμ×Aμ→[0,∞) satisfying the following conditions:
dF(xτ,yη)≥ρ(0)-τ(x-y) for all x,y∈Aμ,
dF(xτ,yη)=0 if and only if x=y and both are nonfuzzy in Kγ,
dF(yη,xτ)=dF(yη,xτ)forallx,y∈Aμ,
dF(xτ,yη)≤dF(xτ,zκ)+dF(zκ,yη)forallx,y,z∈Aμ.
A pair (Aμ,dF) is said to be fuzzy metric spaces.
If d is a metric on a field K and A is any subset of K, then by considering characteristic function on A as its membership function, d defines a fuzzy metric on A. Thus, one can consider an ordinary metric as fuzzy metric.
Theorem 31.
Let Kγ be a fuzzy set field of a field K such that ρ∈γ is as in Theorem 19, μ⊂γ and Aμ be a nonempty fuzzy set in field K. A mapping dF:Aμ×Aμ→[0,∞) given by dF(xτ,yη)=d(x,y)+ρ(0)-min{τ(x),η(y)} defines a fuzzy metric on Aμ.
Proof.
Let xτ,yη,zκ∈Aμ. We need to verify (1) to (4) of Definition 30.
dF(xτ,yη)=d(x,y)+ρ(0)-min{τ(x),η(y)}≤d(x,y)+ρ(0) − min{τ(x),η(y),κ(z)}≤d(x,y) + ρ(0)-min{τ(x),κ(z)}+ρ(0) − min{κ(z),η(y)}≤dF(xτ,zκ)+dF(zκ,yη). Hence, the theorem is proved.
Corollary 32.
Let Kγ be a fuzzy set field of a field K such that ρ∈γ is as in Theorem 19, v⊂γ and Aν be a non empty fuzzy set in field K. A mapping dF:Aν×Aν→[0,∞) defined by dF(xτ,yη)=|x-y|+ρ(0)-min{τ(x),η(y)} is fuzzy metric on Aν.
Theorem 33.
Let Kγ be a fuzzy set of a field K such that ρ∈γ is as in Theorem 19, μ⊂γ and Aμ be a nonempty fuzzy set in field K. A mapping dF:Aμ×Aμ→[0,∞) defined by dF(xτ,yη)=|xτ⊖yη|F+ρ(0)-min{τ(x),η(y)} is fuzzy metric on Aμ.
Proof.
Let xτ,yη,zκ∈Aν. We need to verify (1)–(4) of Definition 30.
In Section 3, we have introduced the definition of fuzzy magnitude and a fuzzy metric of a fuzzy point and a fuzzy set, respectively, in a fuzzy set field, Kγ, in which for all x∈K and for all τ∈γ,τ(x)≤ρ(0). In this section, we try to define a fuzzy magnitude and fuzzy metric of a fuzzy point and fuzzy set in a field, respectively. The definitions introduced differ slightly from that of preceding section due to targeted highest membership value which is assumed to be 1≥ρ(0), where ρ is as in Theorem 19. However, the two definitions coincide whenever ρ(0)=1.
Definition 34.
Let Aγ be fuzzy set in a field. A fuzzy point xτ∈Aγ is said to be nonfuzzy point if and only if τ(x)=1.
Definition 35.
Let xτ be a fuzzy point in a field K. Then the fuzzy magnitude of xτ, denoted by |xτ|F, is given by
(8)|xτ|F=|x|+1-τ(x).
From Definition 34 a fuzzy magnitude of a nonfuzzy point coincides with that of usual magnitude. Therefore one can redefine the magnitude of a given vector in a field as a fuzzy magnitude.
Theorem 36.
If xτ,yη, and zκ are fuzzy points in a field K. Then
|xτ⊖yη|F=|yη⊖xτ|F,
|xτ⊕yη|F≤|xτ|F+|yη|F,
|xτ⊖yη|F≤|xτ⊖zκ|F+|zκ⊖yη|F.
Proof.
Let xτ, γη, and zk be fuzzy points in a field K. Then
Let Aμ be a nonempty fuzzy set in fuzzy field K. A fuzzy metric on Aμ is a mapping dF:Aμ×Aμ→[0,∞) satisfying the following conditions:
dF(xτ,yη)≥1-ϕ(x-y),
dF(xτ,yη)=0 if and only if x is nonfuzzy in Aμ and x=y,
dF(xτ,yη)=dF(yη,xτ) for all x,y∈Aμ,
dF(xτ,yη)≤dF(xτ,zκ)+dF(zκ,yη) for all x,y,z∈Aμ.
A function dF is said to be fuzzy metric on Aμ and a pair (Aμ,dF) is said to be fuzzy metric space.
Theorem 38.
Let K be a field, let d be metric on K and Aμbe a nonempty fuzzy set in K. A mapping dF:Aμ×Aμ→[0,∞) defined by dF(xτ,yη)=d(x,y)+1-min{τ(x),η(y)} is a fuzzy metric on Aμ.
Proof.
Let x,y,z∈Aμ with respective membership values τ(x),η(y)and κ(z). We will verify that (1) to (4) of Definition 37.
dF(xτ,yη)=d(x,y)+1-min{τ(x),η(y)}⇒dF(xτ,yη)≥1-ϕ(x-y), by Remark 11.
dF(xτ,yη)=0⇔d(x,y)+1-min{τ(x),η(y)}=0⇔d(x,y)=0,min{τ(x),η(y)}=1⇔x=y, 1=τ(x)=η(y)⇔x=y and by Definition 34 both are nonfuzzy.
In the preceding theorem, dF(xτ,yη)=d(x,y) provided that x,y∈Aμ are nonfuzzy in Aμ. Therefore, if μ is characteristic function, then fuzzy metric coincides with usual metric of Aμ. So we call a fuzzy metric given as in Theorem 38 a standard fuzzy metric on a fuzzy set.
Corollary 39.
Let Aμ be a fuzzy set in a field K. A mapping dF:Aμ×Aμ→[0,∞) given by dF(xτ,yη)=|x-y|+1-min{τ(x),η(y)} defines a fuzzy metric on Aμ.
Proof.
The result follows from Theorem 38.
Theorem 40.
Let Aν be a nonempty fuzzy set in a field K. A mapping dF:Aν×Aν→[0,∞) defined by dF(xτ,yη)=|xτ⊖yη|F+1-min{τ(x),η(y)} is fuzzy metric on Aν.
Proof.
Proof is similar to Theorem 33, so it is omitted.
5. Conclusion
In this paper, fuzzy set field in a field is defined and some of its properties are discussed. We believe that our results will be helpful to develop similar notion for fuzzy linear space. The definition of the fuzzy metric introduced in this paper resembles distance between two fuzzy points and it involves both fuzziness and crisp property of fuzzy points that constitute classical metric properties. The definition is more general and it can be applied to define a fuzzy metric on linguistic variables also. We believe that our metric might be applied in fuzzy-decision theory, pattern recognition, and image processing.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
TripathyB. C.BaruahA.Nörlund and Riesz mean of sequences of fuzzy real numbers201023565165510.1016/j.aml.2010.02.0062-s2.0-77949484965TripathyB. C.SarmaB.Some double sequence spaces of fuzzy numbers defined by Orlicz functions201131113414010.1016/S0252-9602(11)60215-42-s2.0-78650985171TripathyB. C.BorgohainS.Sequence spaces of fuzzy real numbers using fuzzy metric2014541112210.5666/KMJ.2014.54.1.11MR3190405ZBL06318432WenxiangG.TuL.Fuzzy linear spaces199249337738010.1016/0165-0114(92)90288-FZBL0791.150012-s2.0-44049125212TaghaviA.MehdizadhM.Adjoint operator in fuzzy normed linear spaces20112453458Seob RheG.HwangI.On fuzzy complete normed spaces2009222SaadatiR.VaezpourS. M.Some results on fuzzy banach spaces2005171-247548410.1007/BF02936069ZBL1077.460602-s2.0-11844286363SanthoshC. P.RamakrishnanT. V.Norm and inner product on fuzzy linear spaces over fuzzy fields201181135144MR27524592-s2.0-79954990737KalevaO.SeikkalaS.On fuzzy metric spaces198412321522910.1016/0165-0114(84)90069-1MR7400952-s2.0-0021408107FangJ.An improved completion theorem of fuzzy metric spaces2011166657410.1016/j.fss.2010.12.002MR27633792-s2.0-78751645897Quan XiaZ.GuoF. F.Fuzzy metric spaces200416371381WongC. K.Fuzzy points and local properties of fuzzy topology19744631632810.1016/0022-247X(74)90242-XMR0348694ZBL0278.540042-s2.0-0016058640DengZ.Fuzzy pseudo-metric spaces1982861749510.1016/0022-247X(82)90255-4ZBL0501.540032-s2.0-0002256558HsuN.On completion of fuzzy metric spaces199237385392ShiF.-G.(L,M)-fuzzy metric spaces2010522231250MR2681491ShiF.ZhengC.Metrization theorems in L-topological spaces2005149345547110.1016/j.fss.2004.02.003MR21118852-s2.0-10444273315ShiF.-G.Pointwise pseudo-metrics in L-fuzzy set theory2001121220921610.1016/S0165-0114(00)00013-0MR18345062-s2.0-0035898980PuP. M.LiuY. M.Fuzzy topology. I. Neighborhood structure of a fuzzy point and Moore-Smith convergence198076257159910.1016/0022-247X(80)90048-7MR587361ZBL0447.54006NasseriH.Fuzzy numbers: positive and nonnegative200833617771780MR2460852ZimmermannH. J.1996SecondBoston, Mass, USAKluwer Academic PublishersDuboisD.PradeH.1980San Diego, Calif, USAAcademic PressMR589341