1.
Introduction
Graphs are useful tools for data analysis, complicated system modeling and information communication. They give us the ability to visualize and interpret data, which enables us to gain greater insights and make more rational decisions. A graph is a practical tool for understanding information about the connections between items. In operations research, system analysis and economics, graph models are widely used. Yet, in many real-life circumstances, a graph-theoretic issue may have a part that is uncertain and cannot be represented by a graph. Fuzzy models are preferable to handle problems with uncertainty. In fuzzy graph theory as well, connectivity evolved into a key idea. Connectivity is among the most essential concerns in graph theory and its applications. A difficulty that highlights the significance of connectedness in fuzzy graphs is the fact that total flow disconnection happens less frequently in physical issues than flow reduction between pairs of vertices. Understanding complex systems, optimizing resource allocation, building effective algorithms and obtaining insights into a variety of fields, ranging from social networks to biological systems, all depend on the study of connectivity in graphs.
Zadeh [1] developed the idea of a fuzzy subset of a set as an extension of the crisp set to indicate uncertainty. He considered the essential idea of a membership value. Since the crisp set only contains the truth values, 0 (which means "false") and 1 (which means "true"), it cannot be used to solve ambiguous real-life problems. A fuzzy set (FS) allowed an object to have the value of membership value within [0,1]. The FS only considers a member's value of membership in a set. Atanassov [2] proposed the idea of intuitionistic fuzzy sets (IFSs) as a generalization of fuzzy sets. IFSs also consider the nonmembership value of a member such that the sum of membership and nonmembership values of a member is less than or equal to 1. Xu and Yager [3] called ϕ=(σϕ,μϕ), an intuitionistic fuzzy number (IFN). Yager [4,5] regarded Pythagorean fuzzy sets (PFSs) as a novel extension of IFSs, defined by the membership value and the nonmembership value fulfilling the condition that the sum of their squares is less than 1. PFSs are more capable than IFSs to model the uncertainties in real life decision-making problems. Yager and Abbasov [6] developed a relation between Pythagorean membership values and complex numbers.
Kaufmann [7] was the first to propose the idea of fuzzy graphs (FGs). Rosenfeld [8] explored a number of theoretical ideas, such as paths, cycles and connectedness in the FGs. Mathew and Sunitha [9] investigated node and arc connectivity in FGs. Chakraborty and Mahapatra [10] introduced the concept of intuitionistic fuzzy graphs (IFGs). Connectivity status of intuitionistic fuzzy graphs was discussed by Bera et al. [11] with application. Naz et al. [12] proposed the concept of Pythagorean fuzzy graphs (PFGs) along with their applications in decision-making. Akram and Naz [13] studied the energy of PFGs with applications. For other extensions and notations, the readers are referred to [14,15,16,17]. Akram et al. [18,19] illustrated connectivity concepts in m-polar fuzzy network models. Ahmad and Nawaz [20,21] studied connectivity in directed rough fuzzy graphs (DRFGs) and introduced the Wiener index of a DRFG. Ahmad and Batool [22] proposed the idea of domination in DRFGs with an application. For other applications, the readers are referred to [23,24,25,26].
The drawback of FGs is that they provide no information concerning the impact of vertices on edges. For example, suppose vertices indicate different hostels, and edges represent the roads linking these hostels. We can create a graph to show the volume of traffic moving between the hostels. The hostel with the most guests will have the most ramps. If S1 and S2 are two different hostels, and S1S2 is a road connecting them, then (S2,S1S2) could reveal the ramp system from the road S1S2 to the hostel S2. The introduction of the concept of fuzzy incidence graphs was necessary to address this gap in these graphs. In interconnection networks with influenced flows, fuzzy incidence graphs are crucial. Thus, it is important to examine their connection features. Dinesh [27,28] proposed the concept of fuzzy incidence graphs (FIGs). Mordeson [29] developed the concept of incidence cut pairs in FIGs. Malik et al. [30] investigated complementary FIGs. Mathew and Mordeson [31] illustrated a variety of connectivity concepts in FIGs. They also covered other structural characteristics of FIGs. Fang et al. [32] discussed the connectivity index and Wiener index of FIGs. They also developed three different types of nodes in FIGs. Subsequently, Nazeer et al. [33] proposed the concepts of order, size, dominance and strong pair domination in FIGs. Strong and weak fuzzy incidence dominance as well as other forms of domination were also covered by them. Nazeer et al. [34] was the first to propose the concepts of cyclic connectivity, fuzzy incidence cycle, cyclic connectivity index and average cyclic connectivity index. For a more detailed and comprehensive study on FIGs, we may suggest [35] to the reader. Nazeer et al. [36] proposed the concept of intuitionistic fuzzy incidence graphs (IFIGs), which they describe as a generalization of FIGs with unique characteristics. In IFIGs, they discussed several types of product, such as the Cartesian product, composition, tensor product and normal product.
FIGs and IFIGs have potential applications in a wide range of sectors, particularly in those electrical, electronic and social networks where not only the edges and vertices are of interest, but additionally how they are related to one another is crucial. However, several issues in real life cannot be described using FIGs and IFIGs. We need a more general graph to handle these certain situations since FIGs and IFIGs may not handle them effectively. The Pythagorean fuzzy incidence graphs (PFIGs) would be a prominent research direction since uncertainties are well expressed using the PFSs. In interconnection networks with influenced flows, PFIGs are essential. Therefore, it is important to study their connectivity properties. Motivated by the factors above, the goal of our study is to propose a generalization of connectivity of FIGs that operate effectively in a Pythagorean fuzzy environment. The novel contributions of our study might be summed up as follows:
● The concept of incidence graph in Pythagorean fuzzy environment is introduced.
● In this work, we examine how the removal of a vertex, edge and pair from the PFIGs affects the strength of a vertex-edge pair's connectivity.
● We propose the idea of Pythagorean fuzzy incidence cycles and trees. Strong incidence pairs are used to describe Pythagorean fuzzy incidence trees and cut pairs.
● We establish the existence of the strongest incidence path between every vertex and edge. Also, we introduce some types of Pythagorean fuzzy incidence pairs, namely, α-strong incidence pairs, β-strong incidence pairs and δ-weak incidence pairs.
● The issue of illicit trafficking in wildlife is addressed by the proposed tools.
The other contents of this paper are structured as follows: In Section 2 we define Pythagorean fuzzy incidence graphs (PFIGs), incidence path and strength of incidence path. Pythagorean fuzzy incidence cycle and Pythagorean fuzzy incidence trees and a few relevant propositions are presented in Section 3. In Section 4, we discuss Pythagorean fuzzy incidence cut vertices, Pythagorean fuzzy incidence bridges and Pythagorean fuzzy incidence cut pairs in PFIG and prove several propositions on Pythagorean fuzzy incidence cut pairs. In Section 5, we define the strong Pythagorean fuzzy incidence pairs and demonstrate the types of strong Pythagorean fuzzy incidence pairs. This section also includes the concept of a strong incidence path and some results on strong Pythagorean incidence pairs. Section 6 provides an application of PFIG in illegal wildlife trade. In Section 7, a comparison between our research work and existing models is given. Lastly, in Section 8, we give some final remarks. The list of abbreviations is given in Table 1.
2.
Pythagorean fuzzy incidence graphs
In this section, we first present a few essential concepts that are relevant to this research article. We then define the Pythagorean fuzzy incidence graphs, subgraphs of Pythagorean fuzzy incidence graphs, complete Pythagorean fuzzy incidence subgraphs and the strength of connectedness between any two vertices of PFIGs. The notion of incidence graphs proposed by Dinesh [28] is defined below:
Definition 2.1. [28] An incidence graph on a non-empty set V is a triplet G=(V,E,I), where E⊆V×V,I⊆V×E.
Definition 2.2. [28] In an incidence graph G=(V,E,I), if (ˇq,ˇtˇw)∈I, then (ˇq,ˇtˇw) is called a pair or incidence pair (IP).
Definition 2.3. [29] Let G=(V,E,I) be an incidence graph. A sequence
is called a walk. It is closed if ˇq0=ˇqn. It is a trail if the edges are distinct and an incidence trail if the pairs are distinct. It is said to be a path if the vertices are distinct. A path is called a cycle if it is closed.
All IPs are distinct by the definition of a cycle. We consider the below sequences to be walks:
The latter is closed if ˇqˇq0=ˇqnˇqn+1. They are called incidence paths (IPt if the vertices are distinct). According to the definition of an incidence path, if ˇqˇt is on the incidence path, then IPs of the type (ˇq,ˇqˇt) and (ˇt,ˇqˇt) are also on the incidence path but not an IP of the form (ˇq,ˇtˇw) with ˇq≠ˇt≠ˇw.
Definition 2.4. [28] An incidence graph is said to be a connected incidence graph if every pair of vertices is connected by an incidence path. A tree is an incidence connected graph with no cycles. It is a forest if it is not connected.
Definition 2.5. [4,5] A Pythagorean fuzzy set (PFS) on a universe ˇX is an object of the form
where σˇA:ˇX→[0,1] and μˇA:ˇX→[0,1] represent the membership and non-membership functions of ˇA, respectively, such that for all ˇx∈ˇX,
πˇA(ˇx)=√1−(σˇA(ˇx))2−(μˇA(ˇx))2 is called the degree of indeterminacy of element ˇx∈ˇX.
We now define a Pythagorean fuzzy incidence graph.
Definition 2.6. A PFIG on a non-empty set V is an ordered triplet ˇG=(ˇJ,ˇK,ˇL), where ˇJ=<V,σˇJ,μˇJ> is a PFS on V, ˇK=<E,σˇK,μˇK> is a PFS on E⊆V×V such that
for all ˇa,ˇk∈V, and ˇL=<I,σˇL,μˇL> is a PFS on I⊆V×E such that
for all (ˇa,ˇkˇl)∈I.
Although pairs of the type (ˇa,ˇkˇl), where ˇa≠ˇk≠ˇl are allowed under the definition of PFIG, only pairs of the form (ˇa,ˇaˇk) will be taken into consideration.
Example 2.1. Consider an incidence graph G=(V,E,I), where V={ˇq,ˇt,ˇa,ˇk,ˇl}, E={ˇaˇk,ˇaˇl,ˇkˇq,ˇlˇq,ˇaˇq,ˇkˇt,ˇqˇt}⊆V×V, and I={(ˇa,ˇaˇk),(ˇk,ˇaˇk),(ˇa,ˇaˇl),(ˇl,ˇaˇl),(ˇl,ˇlˇq),(ˇq,ˇlˇq),(ˇk,ˇkˇq),(ˇq,ˇkˇq),(ˇa,ˇaˇq),(ˇt,ˇkˇt),(ˇt,ˇqˇt)}⊆V×E. Let ˇJ, ˇK and ˇL be the PFSs defined on V, E and I, respectively:
ˇJ={(ˇq,0.5,0.6),(ˇt,0.4,0.9),(ˇa,0.7,0.5),(ˇk,0.6,0.7),(ˇl,0.8,0.5)}, ˇK={(ˇaˇk,0.4,0.6),(ˇaˇl,0.6,0.4),(ˇkˇq,0.5,0.5),(ˇlˇq,0.5,0.6),(ˇaˇq,0.4,0.6),(ˇkˇt,0.3,0.8),(ˇqˇt,0.4,0.9)},
ˇL={((ˇa,ˇaˇk),0.4,0.5),((ˇk,ˇaˇk),0.4,0.7),((ˇa,ˇaˇl),0.4,0.5),((ˇl,ˇaˇl),0.6,0.5),((ˇl,ˇlˇq),0.4,0.6),((ˇq,ˇlˇq),0.4,0.4),((ˇk,ˇkˇq),0.5,0.7),((ˇq,ˇkˇq),0.4,0.5),((ˇa,ˇaˇq),0.4,0.6),((ˇt,ˇkˇt),0.2,0.7),((ˇt,ˇqˇt),0.3,0.7)}.
By routine calculations, it is easy to see from Figure 1 that ˇG=(ˇJ,ˇK,ˇL) is a PFIG.
Definition 2.7. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. The supports of ˇJ, ˇK and ˇL, respectively, are defined below:
supp(ˇJ)={ˇa∈V|σˇJ(ˇa)≠0orμˇJ(ˇa)≠0},
supp(ˇK)={ˇaˇk∈E|σˇK(ˇaˇk)≠0orμˇK(ˇaˇk)≠0},
supp(ˇL)={(ˇa,ˇaˇk)∈V×E|σˇL(ˇa,ˇaˇk)≠0orμˇL(ˇa,ˇaˇk)≠0}.
Definition 2.8. The incidence strength of PFIG ˇG=(ˇJ,ˇK,ˇL) is defined by ISˇG=(σISˇG,μISˇG), where σISˇG=min{σˇL(ˇa,ˇaˇk)|(ˇa,ˇaˇk)∈supp(ˇL)} and μISˇG=max{μˇL(ˇa,ˇaˇk)|(ˇa,ˇaˇk)∈supp(ˇL)}.
Example 2.2. Consider a PFIG ˇG=(ˇJ,ˇK,ˇL) as shown in Figure 2.1. Then,
σISˇG=min{σˇL(ˇa,ˇaˇk),σˇL(ˇk,ˇaˇk),σˇL(ˇa,ˇaˇl),σˇL(ˇl,ˇaˇl),σˇL(ˇl,ˇlˇq),σˇL(ˇq,ˇlˇq),σˇL(ˇk,ˇkˇq),σˇL(ˇq,ˇkˇq),σˇL(ˇa,ˇaˇq),σˇL(ˇt,ˇkˇt),σˇL(ˇt,ˇqˇt)} =min{0.4,0.4,0.4,0.6,0.4,0.4,0.5,0.4,0.4,0.2,0.3}=0.2,
μISˇG=max{μˇL(ˇa,ˇaˇk),μˇL(ˇk,ˇaˇk),μˇL(ˇa,ˇaˇl),μˇL(ˇl,ˇaˇl),μˇL(ˇl,ˇlˇq),μˇL(ˇq,ˇlˇq),μˇL(ˇk,ˇkˇq),μˇL(ˇq,ˇkˇq),μˇL(ˇa,ˇaˇq),μˇL(ˇt,ˇkˇt),μˇL(ˇt,ˇqˇt)} =max{0.5,0.7,0.5,0.5,0.6,0.4,0.7,0.5,0.6,0.7,0.7}=0.7.
Thus, the incidence strength of ˇG is ISˇG=(σISˇG,μISˇG)=(0.2,0.7).
Definition 2.9. A PFIG ˇH=(ˇM,ˇQ,ˇS) is called a partial Pythagorean fuzzy incidence subgraph of PFIG ˇG=(ˇJ,ˇK,ˇL) if
Example 2.3. Consider a PFIG ˇG=(ˇJ,ˇK,ˇL) as shown in Figure 1. Let ˇM={(ˇq,0.4,0.7),(ˇt,0.2,0.9),(ˇa,0.5,0.7),(ˇk,0.4,0.8),(ˇl,0.6,0.7)}, ˇQ={(ˇaˇk,0.2,0.7),(ˇaˇl,0.4,0.5),(ˇlˇq,0.3,0.6),(ˇaˇq,0.3,0.6),(ˇkˇq,0.4,0.7),(ˇqˇt,0.2,0.9)}, and ˇS={((ˇa,ˇaˇk),0.2,0.6),(ˇk,ˇaˇk),0.2,0.7),((ˇa,ˇaˇl),0.3,0.6),((ˇl,ˇaˇl),0.4,0.6),((ˇl,ˇlˇq),0.2,0.7),((ˇq,ˇlˇq),0.3,0.5),((ˇa,ˇaˇq),0.3,0.7),((ˇk,ˇkˇq),0.3,0.8),((ˇt,ˇqˇt),0.2,0.8)}.
By direct calculations, it is easy to see from Figure 2 that ˇH=(ˇM,ˇQ,ˇS) is a partial Pythagorean fuzzy incidence subgraph of ˇG=(ˇJ,ˇK,ˇL).
Definition 2.10. A PFIG ˇH=(ˇM,ˇQ,ˇS) is called a Pythagorean fuzzy incidence subgraph (PFIS) of PFIG ˇG=(ˇJ,ˇK,ˇL) if
A PFIS ˇH=(ˇM,ˇQ,ˇS) is said to bea spanning Pythagorean fuzzy incidence subgraph (SPFIS) of PFIG ˇG=(ˇJ,ˇK,ˇL) if supp(ˇM)=supp(ˇJ).
Example 2.4. Consider a PFIG ˇG=(ˇJ,ˇK,ˇL) as shown in Figure 1. A PFIS and SPFIS of ˇG=(ˇJ,ˇK,ˇL) are shown in Figures 3 and 4, respectively.
Definition 2.11. A PFIG ˇG=(ˇJ,ˇK,ˇL) is called a complete PFIG if
Example 2.5. Let ˇG=(ˇJ,ˇK,ˇL) with ˇJ={(x,0.5,0.3),(y,0.7,0.4),(z,0.6,0.2)} be a PFIG as shown in Figure 5. By routine calculations, it is easy to see that ˇG=(ˇJ,ˇK,ˇL) is a complete PFIG.
Definition 2.12. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. An IPt is a walk of distinct vertices, edges and pairs.
Definition 2.13. The incidence strength of an IPt ˇP of PFIG ˇG=(ˇJ,ˇK,ˇL), denoted by ISˇP, is defined by
where σISˇP=min{σˇL(ˇa,ˇaˇk):(ˇa,ˇaˇk)∈ˇP} and μISˇP=max{μˇL(ˇa,ˇaˇk):(ˇa,ˇaˇk)∈ˇP} are σ-incidence strength and μ-incidence strength of ˇP, respectively.
Example 2.6. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG as shown in Figure 6. Consider a ˇq−ˇl IPt between vertices ˇq and ˇl,
The σ-incidence strength of an IPt ˇP1 is given by
and the μ-incidence strength of an IPt ˇP1 is given by
Thus, the incidence strength of an IPt ˇP1 is ISˇP1=(σISˇP1,μISˇP1)=(0.2,0.7). Consider an ˇa−ˇtˇw IPt between a vertex ˇa and an edge ˇtˇw,
The σ-incidence strength of an IPt ˇP2 is given by
and the μ-incidence strength of an IPt ˇP2 is given by
Thus, the incidence strength of an IPt ˇP2 is ISˇP2=(σISˇP2,μISˇP2)=(0.1,0.8). Consider an ˇaˇk−ˇwˇl IPt between edges ˇaˇk and ˇwˇl,
The σ-incidence strength of an IPt ˇP3 is given by
and the μ-incidence strength of an IPt ˇP3 is given by
Thus, the incidence strength of an IPt ˇP3 is ISˇP3=(σISˇP3,μISˇP3)=(0.1,0.8).
Definition 2.14. In a PFIG ˇG=(ˇJ,ˇK,ˇL), the incidence strength of connectedness between vertices ˇa and ˇk, denoted by ICONNˇG(ˇa,ˇk), is defined by
where ICONNσ(ˇG)(ˇa,ˇk)=max{σISˇPi} and ICONNμ(ˇG)(ˇa,ˇk)=min{μISˇPi} are σ-incidence strength and μ-incidence strength of connectedness between ˇa and ˇk, respectively. Here ˇPi represents all possible IPts between ˇa and ˇk.
We now define incidence strength of connectedness of an IPt between a vertex and an edge in PFIG.
Definition 2.15. A vertex ˇa and an edge ˇkˇl in a PFIG ˇG=(ˇJ,ˇK,ˇL) are called connected if an IPt exists between them.
Definition 2.16. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. The incidence strength of connectedness between a vertex ˇa and an edge ˇkˇl, denoted by ICONNˇG(ˇa,ˇkˇl), is defined by
where ICONNσ(ˇG)(ˇa,ˇkˇl)=max{σISˇPi} and ICONNμ(ˇG)(ˇa,ˇkˇl)=min{μISˇPi} are σ-incidence strength and μ-incidence strength of connectedness between ˇa and ˇkˇl, respectively. Here ˇPi represents all possible IPts between ˇa and ˇkˇl.
Example 2.7. Consider a PFIG ˇG=(ˇJ,ˇK,ˇL) as shown in Figure 6. All possible ˇq−ˇkˇl IPts are
ˇP1:ˇq,(ˇq,ˇqˇa),ˇqˇa,(ˇa,ˇqˇa),ˇa,(ˇa,ˇaˇk),ˇaˇk,(ˇk,ˇaˇk),ˇk,(ˇk,ˇkˇt),ˇkˇt,ˇt,(ˇt,ˇtˇw),ˇtˇw,(ˇw,ˇtˇw),ˇw,(ˇw,ˇwˇl),ˇwˇl,(ˇl,ˇwˇl),ˇl,(ˇl,ˇlˇk),ˇlˇk=ˇkˇl;
ˇP2:ˇq,(ˇq,ˇqˇt),ˇqˇt,(ˇt,ˇqˇt),ˇt,(ˇt,ˇtˇw),ˇtˇw,(ˇw,ˇtˇw),ˇw,(ˇw,ˇwˇl),ˇwˇl,(ˇl,ˇwˇl),ˇl,(ˇl,ˇlˇk),ˇlˇk=ˇkˇl; ˇP3:ˇq,(ˇq,ˇqˇt),ˇqˇt,(ˇt,ˇqˇt),ˇt,(ˇt,ˇtˇk),ˇtˇk,(ˇk,ˇtˇk),ˇk,(ˇk,ˇkˇl),ˇkˇl;
ˇP4:ˇq,(ˇq,ˇqˇa),ˇqˇa,(ˇa,ˇqˇa),ˇa,(ˇa,ˇaˇk),ˇaˇk,(ˇk,ˇaˇk),ˇk,(ˇk,ˇkˇl),ˇkˇl.
The incidence strengths of these IPts are given by
The σ-incidence strength and μ-incidence strength of connectedness are given by
Thus, ICONNˇG(ˇq,ˇkˇl)=(ICONNσ(ˇG)(ˇq,ˇkˇl),ICONNμ(ˇG)(ˇq,ˇkˇl))=(0.2,0.7).
All possible ˇq−ˇqˇa IPts are
ˇP1:ˇq,(ˇq,ˇqˇt),ˇqˇt,(ˇt,ˇqˇt),ˇt,(ˇt,ˇtˇw),ˇtˇw,(ˇw,ˇtˇw),ˇw,(ˇw,ˇwˇl),ˇwˇl,(ˇl,ˇwˇl),ˇl,(ˇl,ˇlˇk),ˇlˇk,(ˇk,ˇlˇk),ˇk,(ˇk,ˇkˇa),ˇkˇa,(ˇa,ˇkˇa),ˇa,(ˇa,ˇaˇq),ˇaˇq=ˇqˇa;
ˇP2:ˇq,(ˇq,ˇqˇt),ˇqˇt,(ˇt,ˇqˇt),ˇt,(ˇt,ˇtˇk),ˇtˇk,(ˇk,ˇtˇk),ˇk,(ˇk,ˇkˇa),ˇkˇa,(ˇa,ˇkˇa),ˇa,(ˇa,ˇaˇq),ˇaˇq=ˇqˇa;
ˇP3:ˇq,(ˇq,ˇqˇa),ˇqˇa.
The incidence strengths of these IPts are given by
The σ-incidence strength and μ-incidence strength of connectedness are given by
Thus, ICONNˇG(ˇq,ˇqˇa)=(ICONNσ(ˇG)(ˇq,ˇqˇa),ICONNμ(ˇG)(ˇq,ˇqˇa))=(0.3,0.4).
Proposition 2.1. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG and ˇH=(ˇM,ˇQ,ˇS) be a PFIS of ˇG. Then, for every (ˇa,ˇaˇk)∈supp(ˇS),
Definition 2.17. An ˇa−ˇkˇl IPt ˇP in a PFIG ˇG=(ˇJ,ˇK,ˇL) is called a strongest ˇa−ˇkˇl IPt if its incidence strength equals ICONNˇG(ˇa,ˇkˇl), i.e., ICONNσ(ˇG)(ˇa,ˇkˇl)=σISˇP and ICONNμ(ˇG)(ˇa,ˇkˇl) =μISˇP.
Remark 2.1. An IPt ˇP1 is said to have more incidence strength than an IPt ˇP2 if σISˇP1>σISˇP2, μISˇP1<μISˇP2. Note that the strongest IPt does not have to be unique.
Example 2.8. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG as shown in Figure 7. All possible ˇn−ˇmˇn IPts are
ˇP1:ˇn,(ˇn,ˇmˇn),ˇmˇn;
ˇP2:ˇn,(ˇn,ˇnˇo),ˇnˇo,(ˇo,ˇnˇo),ˇo,(ˇo,ˇoˇm),ˇoˇm,(ˇm,ˇoˇm),ˇm,(ˇm,ˇmˇn),ˇmˇn;
ˇP3:ˇn,(ˇn,ˇnˇp),ˇnˇp,(ˇp,ˇnˇp),ˇp,(ˇp,ˇpˇm),ˇpˇm,(ˇm,ˇpˇm),ˇm,(ˇm,ˇmˇn),ˇmˇn.
The incidence strengths of these IPts are given by
Thus, ICONNˇG(ˇn,ˇmˇn)=(ICONNσ(ˇG)(ˇn,ˇmˇn),ICONNμ(ˇG)(ˇn,ˇmˇn))=(0.4,0.6).
The IPt ˇP2:ˇn,(ˇn,ˇnˇo),ˇnˇo,(ˇo,ˇnˇo),ˇo,(ˇo,ˇoˇm),ˇoˇm,(ˇm,ˇoˇm),ˇm,(ˇm,ˇmˇn),ˇmˇn is a strongest n−nm IPt since ICONNσ(ˇG)(ˇn,ˇmˇn)=σISˇP2 and ICONNμ(ˇG)(ˇn,ˇmˇn)=μISˇP2. Similarly, ˇn,(ˇn,ˇnˇo),ˇnˇo is the strongest ˇn−ˇnˇo IPt. Note that both ˇp,(ˇp,ˇpˇm),ˇmˇp,(ˇm,ˇpˇm),ˇm,(ˇm,ˇmˇn),ˇmˇn,(ˇn,ˇmˇn),ˇn,(ˇn,ˇnˇp),ˇnˇp and ˇp,(ˇp,ˇpˇm),ˇmˇp,(ˇm,ˇpˇm),ˇm,(ˇm,ˇmˇo),ˇmˇo,(ˇo,ˇmˇo),ˇo,(ˇo,ˇoˇn),ˇoˇn,(ˇn,ˇoˇn),ˇn,(ˇn,ˇnˇp),ˇnˇp are strongest ˇp−ˇnˇp IPts.
3.
Pythagorean fuzzy incidence cycles and trees
In this section, we define the notions of Pythagorean fuzzy incidence cycles and Pythagorean fuzzy incidence trees in Pythagorean fuzzy incidence graphs and then prove several propositions. The definition of cycle of a PFIG ˇG=(ˇJ,ˇK,ˇL) is given below:
Definition 3.1. A PFIG ˇG=(ˇJ,ˇK,ˇL) is a cycle if (supp(ˇJ),supp(ˇK),supp(ˇL)) is a cycle.
Definition 3.2. A PFIG ˇG=(ˇJ,ˇK,ˇL) is a Pythagorean fuzzy cycle if (supp(ˇJ),supp(ˇK),supp(ˇL)) is a cycle and no unique ˇqˇt∈supp(ˇK) exists such that σˇK(ˇqˇt)=min{σˇK(ˇaˇk)|ˇaˇk∈supp(ˇK)} and μˇK(ˇqˇt)=max{μˇK(ˇaˇk)|ˇaˇk∈supp(ˇK)}.
Definition 3.3. A PFIG ˇG=(ˇJ,ˇK,ˇL) is a Pythagorean fuzzy incidence cycle (PFIC) if it is a Pythagorean fuzzy cycle and no unique (ˇq,ˇqˇt)∈supp(ˇL) exists such that
Definition 3.4. A pair (ˇq,ˇqˇt) is called a weakest IP of cycle C if σˇL(ˇq,ˇqˇt)=min{σˇL(ˇa,ˇaˇk)|(ˇa,ˇaˇk)∈C} and μˇL(ˇq,ˇqˇt)=max{μˇL(ˇa,ˇaˇk)|(ˇa,ˇaˇk)∈C}.
Example 3.1. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG as shown in Figure 8. Consider the walk
It is a cycle.
Since
and also σˇK(ˇpˇm)=0.3,μˇK(ˇpˇm)=0.7. Thus, ˇG is a Pythagorean fuzzy cycle. ˇG is a PFIC since
and also σˇL(ˇo,ˇoˇp)=σˇL(ˇm,ˇmˇp)=0.2,μˇL(ˇo,ˇoˇp)=μˇL(ˇm,ˇmˇp)=0.7.
Definition 3.5. A connected PFIG ˇG=(ˇJ,ˇK,ˇL) is a tree if (supp(ˇJ),supp(ˇK),supp(ˇL)) is a tree.
Definition 3.6. A connected PFIG ˇG=(ˇJ,ˇK,ˇL) is a Pythagorean fuzzy tree if it has a Pythagorean fuzzy incidence spanning subgraph ˇT=(ˇM,ˇQ,ˇS) which is a tree such that for all ˇaˇk not in ˇT,σˇK(ˇaˇk)<CONNσ(ˇT)(ˇa,ˇk) and μˇK(ˇaˇk)>CONNμ(ˇT)(ˇa,ˇk).
Definition 3.7. A connected PFIG ˇG=(ˇJ,ˇK,ˇL) is a Pythagorean fuzzy incidence tree (PFIT) if it has a Pythagorean fuzzy incidence spanning subgraph ˇT=(ˇM,ˇQ,ˇS) which is a tree such that for every (ˇa,ˇaˇk) not in ˇT,
Example 3.2. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG as shown in Figure 9. It is easy to see that ˇG has a SPFIS ˇT=(ˇM,ˇQ,ˇS) as shown in Figure 10. This is a tree, and CONNσ(ˇT)(ˇq,ˇa)=0.3,CONNμ(ˇT)(ˇq,ˇa)=0.7.
Since ˇqˇa∈supp(ˇK)∖supp(ˇQ) satisfies
Thus, ˇG is a Pythagorean fuzzy tree. Also, ICONNσ(ˇT)(ˇa,ˇaˇq)=0.2,ICONNμ(ˇT)(ˇa,ˇaˇq)=0.7, and ICONNσ(ˇT)(ˇq,ˇaˇq)=0.2,ICONNμ(ˇT)(ˇq,ˇaˇq)=0.7. Since (ˇa,ˇaˇq)∈supp(ˇL)∖supp(ˇS),
and (ˇq,ˇaˇq)∈supp(ˇL)∖supp(ˇS),
Thus, ˇG=(ˇJ,ˇK,ˇL) is a PFIT.
Definition 3.8. The PFIG ˇG=(ˇJ,ˇK,ˇL) is a forest if (supp(ˇJ),supp(ˇK),supp(ˇL)) is a forest.
Definition 3.9. A PFIG ˇG=(ˇJ,ˇK,ˇL) is a Pythagorean fuzzy forest if it has a Pythagorean fuzzy incidence spanning subgraph ˇF=(ˇM,ˇQ,ˇS) which is a forest such that for all ˇaˇk not in ˇF,σˇK(ˇaˇk)<CONNσ(ˇF)(ˇa,ˇk) and μˇK(ˇaˇk)>CONNμ(ˇF)(ˇa,ˇk).
Definition 3.10. A PFIG ˇG=(ˇJ,ˇK,ˇL) is a Pythagorean fuzzy incidence forest (PFIF) if it has a Pythagorean fuzzy incidence spanning subgraph ˇF=(ˇM,ˇQ,ˇS) which is a forest such that for every (ˇa,ˇaˇk) not in ˇF,
Proposition 3.1. A PFIG ˇG=(ˇJ,ˇK,ˇL) is a PFIF if and only if in any cycle of ˇG, there is an IP (ˇa,ˇaˇk) such that σˇL(ˇa,ˇaˇk)<ICONNσ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk) and μˇL(ˇa,ˇaˇk)>ICONNμ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk).
Proof. The result is trivially true if there are no cycles. Let (ˇa,ˇaˇk)∈ˇG and let (ˇa,ˇaˇk) belong to a Pythagorean fuzzy cycle such that
Let (ˇa,ˇaˇk) be a weakest IP of cycle C, i.e.,
Then, PFIS obtained after deleting IP (ˇa,ˇaˇk) is a PFIF. Remove IPs in a similar manner if there are any further cycles. The removed IP will always have a lesser incidence strength than those eliminated previously at each step. The PFIS that remains after deletion of IPs is a PFIF. As a result, there is an IPt ˇP between ˇa and ˇaˇk such that σISˇP>σˇL(ˇa,ˇaˇk), μISˇP<μˇL(ˇa,ˇaˇk) and does not include (ˇa,ˇaˇk). If earlier deleted IPs still exist in ˇP, we can use an IPt with more incidence strength to bypass them.
Conversely, if ˇG is a PFIF, and C is any cycle, then by definition, there exist (ˇa,ˇaˇk) IPs of C not in ˇF such that
where ˇF is as in the PFIF definition. □
Proposition 3.2. If there is at most one IPt between any vertex ˇa and edge ˇaˇk of PFIG ˇG=(ˇJ,ˇK,ˇL) with the most incidence strength, then ˇG is a PFIF.
Proof. Let ˇG be not a PFIF. Then, by Proposition 3.1, ∃ a cycle C in ˇG such that for all (ˇa,ˇaˇk)∈C, σˇL(ˇa,ˇaˇk)≥ICONNσ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk) and μˇL(ˇa,ˇaˇk)≤ICONNμ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk). Therefore ˇa,(ˇa,ˇaˇk),ˇaˇk is the strongest IPt between ˇa and ˇaˇk. Suppose (ˇa,ˇaˇk) is a weakest IP of cycle C, and then σˇL(ˇa,ˇaˇk)=min{σˇL(ˇq,ˇqˇt)|(ˇq,ˇqˇt)∈C} and μˇL(ˇa,ˇaˇk)=max{μˇL(ˇq,ˇqˇt)|(ˇq,ˇqˇt)∈C}. Hence, the remaining portion of C is the IPt between ˇa and ˇaˇk with the most incidence strength, which is a contradiction. Hence, ˇG is a PFIF. □
Proposition 3.3. Let ˇG=(ˇJ,ˇK,ˇL) be a cycle. Then, ˇG is a PFIC if and only if ˇG is not a PFIT.
Proof. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIC. Then, ∃ at least two IPs (ˇa,ˇaˇk)∈supp(ˇL) with
Let ˇT=(ˇM,ˇQ,ˇS) be a spanning PFIT in ˇG=(ˇJ,ˇK,ˇL). Then{, } there exists (ˇq,ˇqˇt) such that supp(ˇL)∖supp(ˇS)={(ˇq,ˇqˇt)}. Hence, IPt between ˇq and ˇqˇt in ˇT=(ˇM,ˇQ,ˇS) such that σˇL(ˇq,ˇqˇt)<ICONNσ(ˇT)(ˇq,ˇqˇt) and μˇL(ˇq,ˇqˇt)>ICONNμ(ˇT)(ˇq,ˇqˇt) does not exist. Thus, ˇT=(ˇM,ˇQ,ˇS) is not a PFIT.
Conversely, assume that ˇG=(ˇJ,ˇK,ˇL) is not a PFIT. Since ˇG is a cycle, ∀(ˇq,ˇqˇt)∈supp(ˇL), we have a SPFIS ˇT=(ˇM,ˇQ,ˇS) which is a tree. σˇS(ˇq,ˇqˇt)=0=μˇS(ˇq,ˇqˇt),
and σˇS(ˇa,ˇaˇk)=σˇL(ˇa,ˇaˇk),μˇS(ˇa,ˇaˇk)=μˇL(ˇa,ˇaˇk),∀(ˇa,ˇaˇk)∈supp(ˇL)∖{(ˇq,ˇqˇt)}. Hence, IP (ˇa,ˇaˇk) for which σˇL(ˇa,ˇaˇk)=min{σˇL(ˇq,ˇqˇt)|(ˇq,ˇqˇt)∈supp(ˇL)} and μˇL(ˇa,ˇaˇk)=max{μˇL(ˇq,ˇqˇt)|(ˇq,ˇqˇt)∈supp(ˇL)} holds is not unique. Thus, ˇG=(ˇJ,ˇK,ˇL) is a PFIC. □
Proposition 3.4. If ˇG=(ˇJ,ˇK,ˇL) is a PFIT, and ˇG∗=(supp(ˇJ),supp(ˇK),supp(ˇL)) is not a tree, then there exists at least one IP (ˇa,ˇaˇk) such that σˇL(ˇa,ˇaˇk)<ICONNGσ(ˇG)(ˇa,ˇaˇk) and μˇL(ˇa,ˇaˇk)>ICONNGμ(ˇG)(ˇa,ˇaˇk).
Proof. Since ˇG is a PFIT, there exists a SPFIS ˇT=(ˇM,ˇQ,ˇS) that is a tree, and for every (ˇa,ˇaˇk)∉ˇT,
Also,
Thus, for every (ˇa,ˇaˇk)∉ˇT,
Thus, one IP (ˇa,ˇaˇk)∉ˇT exists. □
Proposition 3.5. If ˇG=(ˇJ,ˇK,ˇL) is a PFIT, then ˇG is not a complete PFIG.
Proof. Let ˇG=(ˇJ,ˇK,ˇL) be a complete PFIG, and then for all (ˇa,ˇaˇk)
Since ˇG=(ˇJ,ˇK,ˇL) is a PFIT, for every (ˇa,ˇaˇk) not in ˇT=(ˇM,ˇQ,ˇS)
where ˇT=(ˇM,ˇQ,ˇS) is a SPFIS of ˇG=(ˇJ,ˇK,ˇL), which is a tree. Thus,
which is not possible. Hence, ˇG=(ˇJ,ˇK,ˇL) is not a complete PFIG. □
4.
Pythagorean fuzzy incidence cut vertices, bridges and cut pairs
In this section, we define Pythagorean fuzzy incidence cut vertices, Pythagorean fuzzy incidence bridges and Pythagorean fuzzy incidence cut pairs. We also establish some results about Pythagorean fuzzy incidence cut pairs. The notion of Pythagorean fuzzy incidence cut vertices is defined below:
Definition 4.1. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. Let ˇl∈V and ˜E be the set difference of E and the set of edges with ˇl as an end vertex. Then, ˇl is called a Pythagorean fuzzy incidence cut vertex (PFICV) of ˇG, if ICONNGσ(ˇG∖{ˇl})(ˇa,ˇaˇk)<ICONNGσ(ˇG)(ˇa,ˇaˇk) and ICONNGμ(ˇG∖{ˇl})(ˇa,ˇaˇk)>ICONNGμ(ˇG)(ˇa,ˇaˇk) for some pair (ˇa,ˇaˇk)∈VטE such that ˇa≠ˇl≠ˇk.
Definition 4.2. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. Let ˇqˇt∈E and ˜E=E∖{ˇqˇt}. Then, ˇqˇt is called a Pythagorean fuzzy incidence bridge of ˇG, if ICONNGσ(ˇG∖{ˇqˇt})(ˇa,ˇaˇk)<ICONNGσ(ˇG)(ˇa,ˇaˇk) and ICONNGμ(ˇG∖{ˇqˇt})(ˇa,ˇaˇk)>ICONNGμ(ˇG)(ˇa,ˇaˇk) for some pair (ˇa,ˇaˇk)∈VטE such that ˇaˇk≠ˇqˇt.
Definition 4.3. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. A pair (ˇq,ˇqˇt)∈supp(ˇL) is called a Pythagorean fuzzy incidence cut pair (PFICP) of ˇG if ICONNGσ(ˇG∖{(ˇq,ˇqˇt)})(ˇa,ˇaˇk)<ICONNGσ(ˇG)(ˇa,ˇaˇk) and ICONNGμ(ˇG∖{(ˇq,ˇqˇt)})(ˇa,ˇaˇk)>ICONNGμ(ˇG)(ˇa,ˇaˇk) for some pair (ˇq,ˇqˇt) in ˇG.
Example 4.1. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG as shown in Figure 11.
By routine calculation, it is easy to see that ˇt3 is PFICV, ˇt2ˇt3,ˇt3ˇt4,ˇt4ˇt1 are Pythagorean incidence bridges, and all incidence pairs except (ˇt2,ˇt1ˇt2) are PFICPs.
Proposition 4.1. If ˇG=(ˇJ,ˇK,ˇL) is a PFIF, then the pairs of ˇF=(ˇM,ˇQ,ˇS) (as in Definition 3.10) are exactly the PFICPs of ˇG.
Proposition 4.2. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. Then, the following statements are equivalent:
(1) ICONNGσ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk)<σˇL(ˇa,ˇaˇk),ICONNGμ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk)>μˇL(ˇa,ˇaˇk).
(2) (ˇa,ˇaˇk) is a PFICP.
(3) (ˇa,ˇaˇk) is not the weakest IP of any cycle.
Proof. The following three implications are proven by contrapositive:
(1)⇒(2) If (ˇa,ˇaˇk) is not a PFICP, then
(2)⇒(3) Assume that the weakest IP in a cycle is (ˇa,ˇaˇk). Then, the rest of the cycle can be used as an IPt between ˇa and ˇaˇk to convert any IPt involving (ˇa,ˇaˇk) into a path not involving (ˇa,ˇaˇk) but at least as strong. Thus, (ˇa,ˇaˇk) is not a PFICP.
(3)⇒(1) Let
Then, there exists an IPt ˇP between ˇa and ˇaˇk that does not include (ˇa,ˇaˇk) such that σISˇP≥σˇL(ˇa,ˇaˇk) and μISˇP≤μˇL(ˇa,ˇaˇk). Then, the cycle formed by this IPt ˇP and (ˇa,ˇaˇk) has (ˇa,ˇaˇk) as its weakest IP. □
Proposition 4.3. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG and (ˇa,ˇaˇk)∈supp(ˇL). If (ˇa,ˇaˇk) is a PFICP, then σˇL(ˇa,ˇaˇk)=ICONNGσ(ˇG)(ˇa,ˇaˇk) and μˇL(ˇa,ˇaˇk)=ICONNGμ(ˇG)(ˇa,ˇaˇk).
Proof. Suppose σˇL(ˇa,ˇaˇk)<ICONNGσ(ˇG)(ˇa,ˇaˇk) and μˇL(ˇa,ˇaˇk)>ICONNGμ(ˇG)(ˇa,ˇaˇk). Then, a strongest ˇa−ˇaˇk IPt exists between such that for all IPs (ˇq,ˇqˇt) in the IPt, σˇL(ˇq,ˇqˇt)>σˇL(ˇa,ˇaˇk) and μˇL(ˇq,ˇqˇt)<μˇL(ˇa,ˇaˇk). Then, the cycle formed by this IPt ˇP and (ˇa,ˇaˇk) has (ˇa,ˇaˇk) as its weakest IP. Hence, (ˇa,ˇaˇk) is not a PFICP. □
Proposition 4.4. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. If ˇl is a common vertex of at least two PFICPs, then ˇl is a PFICV.
Proof. Let (ˇl,ˇaˇl) and (ˇl,ˇkˇl) be two PFICPs. Then, ˇq,ˇt∈V exists such that every strongest ˇq−ˇt IPt contains (ˇl,ˇaˇl). If ˇl≠ˇq and ˇl≠ˇt, then ˇl is a PFICV. Let ˇl=ˇq or ˇl=ˇt, and then every strongest ˇq−ˇl IPt contains (ˇl,ˇaˇl), or every strongest ˇl−ˇt IPt contains (ˇl,ˇkˇl). Assume that ˇl is not a PFICV. Then, at least one strongest IPt ˇP exists between every two vertices not containing ˇl. Then, the cycle C is formed by this IPt ˇP with (ˇl,ˇaˇl) and (ˇl,ˇkˇl). We will discuss two cases:
(1) Suppose that ˇa,(ˇa,ˇaˇl),ˇaˇl,(ˇl,ˇaˇl),ˇl,(ˇl,ˇlˇk),ˇlˇk,(ˇk,ˇlˇk),ˇk is not a strongest IPt. Then, (ˇl,ˇaˇl) or (ˇl,ˇkˇl) or both will be weakest IPs of the cycle C, which contradicts that (ˇl,ˇaˇl) and (ˇl,ˇkˇl) are PFICPs.
(2) Suppose that ˇP1:ˇa,(ˇa,ˇaˇl),ˇaˇl,(ˇl,ˇaˇl),ˇl,(ˇl,ˇlˇk),ˇlˇk,(ˇk,ˇlˇk),ˇk is a strongest IPt joining ˇl and ˇk. Then, σISˇP1=min{σˇL(ˇl,ˇaˇl),σˇL(ˇl,ˇkˇl)}, and μISˇP1=max{μˇL(ˇl,ˇaˇl),μˇL(ˇl,ˇkˇl)}), the strength of IPt ˇP. Thus, for all (ˇq,ˇqˇt)∈ˇP
and
Thus, both (ˇl,ˇaˇl) and (ˇl,ˇkˇl) are the weakest IPs of the cycle C, which is a contradiction.
□
Proposition 4.5. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIT. Then, the internal vertices of a SPFIS ˇT=(ˇM,ˇQ,ˇS), which is a tree, are the PFICVs of ˇG.
Proof. Let ˇl∈ˇG=(ˇJ,ˇK,ˇL) such that ˇl is not an end vertex of ˇT. Then, ˇl is the common vertex of at least two IPs in ˇT, which are PFICPs of ˇG. Thus, by Proposition 4.4, ˇl is a PFICV. Let ˇl be an end vertex of ˇT. Then, ˇl is not a PFICV, or else there would exist ˇa≠ˇl and ˇk≠ˇl such that every strongest ˇa−ˇk IPt contains ˇl, and one such IPt lies in ˇT, which is impossible since ˇl is an end vertex of ˇT. □
Corollary 4.1. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIT. Then, a PFICV ˇl of ˇG is the common vertex of at least two PFICPs.
5.
Connectivity in Pythagorean fuzzy incidence graphs
In this section, we first define the strong incidence pairs and weak incidence pairs in Pythagorean fuzzy incidence graphs. Then, we obtain the characterization of Pythagorean fuzzy incidence cut pairs using a-strong incidence pairs. Also, we examine the connectivity of Pythagorean fuzzy incidence graphs by various theorems.
Definition 5.1. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. An IP (ˇq,ˇqˇt) is said to be a strong incidence pair (SIP) if σˇL(ˇq,ˇqˇt)≥ICONNσ(ˇG∖{(ˇq,ˇqˇt)})(ˇq,ˇqˇt) and μˇL(ˇq,ˇqˇt)≤ICONNμ(ˇG∖{(ˇq,ˇqˇt)})(ˇq,ˇqˇt).
A SIP is called an α-SIP if σˇL(ˇq,ˇqˇt)>ICONNσ(ˇG∖{(ˇq,ˇqˇt)})(ˇq,ˇqˇt) and μˇL(ˇq,ˇqˇt)<ICONNμ(ˇG∖{(ˇq,ˇqˇt)})(ˇq,ˇqˇt).
A SIP is called a β-SIP if σˇL(ˇq,ˇqˇt)=ICONNσ(ˇG∖{(ˇq,ˇqˇt)})(ˇq,ˇqˇt) and μˇL(ˇq,ˇqˇt)=ICONNμ(ˇG∖{(ˇq,ˇqˇt)})(ˇq,ˇqˇt).
Note that a SIP need not to be an α-SIP or a β-SIP.
Definition 5.2. An IP (ˇq,ˇqˇt) is said to be a δ-weak IP if σˇL(ˇq,ˇqˇt)≤ICONNσ(ˇG∖{(ˇq,ˇqˇt)})(ˇq,ˇqˇt) and μˇL(ˇq,ˇqˇt)≥ICONNμ(ˇG∖{(ˇq,ˇqˇt)})(ˇq,ˇqˇt).
Definition 5.3. An IP (ˇq,ˇqˇt) is said to be a δ∗-IP if it is a δ-weak IP with σˇL(ˇq,ˇqˇt)>min{σˇL(ˇa,ˇaˇk)|(ˇa,ˇaˇk)∈supp(ˇL)} and μˇL(ˇq,ˇqˇt)<max{μˇL(ˇa,ˇaˇk)|(ˇa,ˇaˇk)∈supp(ˇL)}.
Example 5.1. In Figure 7. Consider an IP (ˇm,ˇmˇn) of ˇG with σˇL(ˇm,ˇmˇn)=0.4 and μˇL(ˇm,ˇmˇn)=0.5. All possible ˇm−ˇmˇn IPts in ˇG are
ˇP1:ˇm,(ˇm,ˇmˇn),ˇmˇn;
ˇP2:ˇm,(ˇm,ˇmˇp),ˇmˇp,(ˇp,ˇmˇp),ˇp,(ˇp,ˇpˇn),ˇpˇn,(ˇn,ˇpˇn),ˇn,(ˇn,ˇnˇm),ˇnˇm=ˇmˇn;
ˇP3:ˇm,(ˇm,ˇmˇo),ˇmˇo,(ˇo,ˇmˇo),ˇo,(ˇo,ˇoˇn),ˇoˇn,(ˇn,ˇoˇn),ˇn,(ˇn,ˇnˇm),ˇnˇm=ˇmˇn.
The incidence strengths of these IPts are given by
The σ-incidence strength and μ-incidence strength of connectedness are given by
After the deletion of IP (ˇm,ˇmˇn), we have ICONNσ(ˇG∖{(ˇm,ˇmˇn)})(ˇm,ˇmˇn)=0.3,ICONNμ(ˇG∖{(ˇm,ˇmˇn)})(ˇm,ˇmˇn)=0.6. σˇL(ˇm,ˇmˇn)>ICONNσ(ˇG∖{(ˇm,ˇmˇn)})(ˇm,ˇmˇn) and μˇL(ˇm,ˇmˇn)<ICONNμ(ˇG∖{(ˇm,ˇmˇn)})(ˇm,ˇmˇn). Thus, (ˇm,ˇmˇn) is an α-SIP. Similarly, (ˇp,ˇpˇm),(ˇm,ˇmˇo),(ˇn,ˇnˇo),(ˇo,ˇoˇn),(ˇm,ˇmˇp),(ˇo,ˇoˇm),(ˇn,ˇnˇp) are SIPs. In particular, (ˇn,ˇnˇo),(ˇo,ˇoˇn),(ˇm,ˇmˇp),(ˇo,ˇoˇm),(ˇn,ˇnˇp) are α-strong IPs. The IPs (ˇp,ˇpˇn),(ˇn,ˇnˇm) are δ-weak IPs. In particular, (ˇn,ˇnˇm) is δ∗-IP. There is no β-SIP in ˇG.
Definition 5.4. In a PFIG ˇG=(ˇJ,ˇK,ˇL), an ˇa−ˇaˇk IPt ˇP is called a strong IPt if all (ˇq,ˇqˇt) in ˇP are strong. In particular, it is called an α-strong IPt if all the IPs in an IPt are α-strong, and it is called a β-strong IPt if all the pairs in an IPt are β-strong. A strong IPt that is closed is called a strong incidence cycle.
Remark 5.1. In a PFIG, the strongest IPt need not be a strong IPt, and a strong IPt need not be the strongest IPt.
Example 5.2. Consider a PFIG as shown in Figure 7. The ˇn−ˇmˇn IPt ˇP:ˇn,(ˇn,ˇnˇo),ˇnˇo,(ˇo,ˇnˇo),ˇo,(ˇo,ˇoˇm),ˇoˇm,(ˇm,ˇoˇm),ˇm,(ˇm,ˇmˇn),ˇmˇn is a strong IPt since all the IPs in ˇP are strong. Similarly, an ˇp−ˇnˇp IPt ˜P:ˇp,(ˇp,ˇpˇm),ˇmˇp,(ˇm,ˇpˇm),ˇm,(ˇm,ˇmˇo),ˇmˇo,(ˇo,ˇmˇo),ˇo,(ˇo,ˇoˇn),ˇoˇn,(ˇn,ˇoˇn),ˇn,(ˇn,ˇnˇp),ˇnˇp is a strong IPt. Note that both ˇP and ˜P IPts are strong as well as strongest IPts. Now consider another strongest ˇp−ˇpˇn IPt ˇp,(ˇp,ˇpˇm),ˇmˇp,(ˇm,ˇpˇm),ˇm,(ˇm,ˇmˇn),ˇmˇn,(ˇn,ˇmˇn),ˇn,(ˇn,ˇnˇp),ˇnˇp. This is not a strong IPt since IP (ˇn,ˇmˇn) is δ-weak IP. The ˇo−ˇmˇo IPt ˇo,(ˇo,ˇoˇn),ˇoˇn,(ˇn,ˇoˇn),ˇn,(ˇn,ˇnˇm),ˇnˇm,(ˇm,ˇnˇm),ˇm,(ˇm,ˇmˇo),ˇmˇo is neither a strongest IPt nor a strong IPt.
Proposition 5.1. A PFIC is a strong incidence cycle.
Proof. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIC. Then, (supp(ˇJ),supp(ˇK),supp(ˇL)) is a cycle such that ∃ no unique ˇqˇt∈supp(ˇK) such that σˇK(ˇqˇt)=min{σˇK(ˇaˇk)|ˇaˇk∈supp(ˇK)} and μˇK(ˇqˇt)=max{μˇK(ˇaˇk)|ˇaˇk∈supp(ˇK)} and ∃ no unique (ˇq,ˇqˇt)∈supp(ˇL) such that
We have to show that each IP in ˇG is a SIP. Let (ˇa,ˇaˇk)∈supp(ˇL) be not a SIP. Then, it is a δ-weak IP. Thus, σˇL(ˇa,ˇaˇk)≤ICONNσ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk), and μˇL(ˇa,ˇaˇk)≥ICONNμ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk). Since (supp(ˇJ),supp(ˇK),supp(ˇL)) is a cycle, every IP (ˇq,ˇqˇt)∈ˇG∖{(ˇa,ˇaˇk)} satisfies
which contradict that ˇG is a PFIC. Thus, (ˇa,ˇaˇk) is a SIP. □
Proposition 5.2. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. An IP (ˇa,ˇaˇk) in ˇG such that σˇL(ˇa,ˇaˇk)=max{σˇL(ˇq,ˇqˇt)|(ˇq,ˇqˇt)∈supp(ˇL)} and μˇL(ˇa,ˇaˇk)=min{μˇL(ˇq,ˇqˇt)|(ˇq,ˇqˇt)∈supp(ˇL)} is a SIP.
Proof. Let ˇP be an ˇa−ˇaˇk IPt in ˇG. Then, σISˇP≤σˇL(ˇa,ˇaˇk) and μISˇP≥μˇL(ˇa,ˇaˇk). If (ˇa,ˇaˇk) is a unique IP such that σˇL(ˇa,ˇaˇk)=max{σˇL(ˇq,ˇqˇt)|(ˇq,ˇqˇt)∈supp(ˇL)} and μˇL(ˇa,ˇaˇk)=min{μˇL(ˇq,ˇqˇt)|(ˇq,ˇqˇt)∈supp(ˇL)}, then for every ˇq−ˇqˇt IPt ˜P in ˇG,
where (ˇq,ˇqˇt) is a pair other than (ˇa,ˇaˇk), and hence,
Thus, (ˇa,ˇaˇk) is a SIP. If (ˇa,ˇaˇk) is not unique, then for every ˇq−ˇqˇt IPt ˜P in ˇG∖{(ˇa,ˇaˇk)}, σIS˜P=σˇL(ˇa,ˇaˇk), and μIS˜P=μˇL(ˇa,ˇaˇk). If ∃ an ˇa−ˇaˇk IPt ˇP in ˇG∖{(ˇa,ˇaˇk)}, then σISˇP=σˇL(ˇa,ˇaˇk) and μISˇP=μˇL(ˇa,ˇaˇk). Hence (ˇa,ˇaˇk) is a β-SIP. Otherwise, it is an α-SIP. □
The converse of this proposition is not necessarily true.
Example 5.3. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG as shown in Figure 12. Note that σˇL(ˇm,ˇmˇn)=0.5=max{σˇL(ˇq,ˇqˇt)|(ˇq,ˇqˇt)∈supp(ˇL)}, and μˇL(ˇm,ˇmˇn)=0.5=min{μˇL(ˇq,ˇqˇt)|(ˇq,ˇqˇt)∈supp(ˇL)}, and (ˇm,ˇmˇn) is a SIP. (ˇa,ˇmˇa) is also an α-SIP in ˇG but σˇL(ˇa,ˇmˇa)≠0.5,μˇL(ˇa,ˇmˇa)≠0.4.
Proposition 5.3. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. An IP (ˇa,ˇaˇk) in ˇG such that σˇL(ˇa,ˇaˇk)=min{σˇJ(ˇa),σˇK(ˇaˇk)} and μˇL(ˇa,ˇaˇk)=max{μˇJ(ˇa),μˇK(ˇaˇk)} is a SIP.
Proof. Consider the PFIG ˇG∖{(ˇa,ˇaˇk)}. If ˇG∖{(ˇa,ˇaˇk)} is disconnected, then (ˇa,ˇaˇk) is a PFICP, and so
Hence, (ˇa,ˇaˇk) is a SIP by definition. If ˇG∖{(ˇa,ˇaˇk)} is connected, then ∃ pairs (ˇa,ˇaˇl) for some ˇl≠ˇk such that (ˇa,ˇaˇl),(ˇk,ˇaˇk)∈ˇP, where ˇP is an IPt between ˇa and ˇaˇk in ˇG∖{(ˇa,ˇaˇk)}. Hence,
That is, σˇL(ˇa,ˇaˇk)≥ICONNσ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk) and μˇL(ˇa,ˇaˇk)≤ICONNμ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk). Hence (ˇa,ˇaˇk) is a SIP. □
The converse of the above proposition is not necessarily true. In Figure 7, σˇL(ˇn,ˇnˇa)=min{σˇJ(ˇn),σˇK(ˇnˇa)} and μˇL(ˇn,ˇnˇa)=max{μˇJ(ˇn),μˇK(ˇnˇa)}, and (ˇn,ˇnˇa) is a SIP. The IP (ˇa,ˇnˇa) is also a SIP but σˇL(ˇa,ˇnˇa)≠min{σˇJ(ˇa),σˇK(ˇnˇa)} and μˇL(ˇa,ˇnˇa)≠max{μˇJ(ˇa),μˇK(ˇnˇa)}.
Proposition 5.4. In a PFIG, every PFICP is a SIP.
Proof. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. Let (ˇa,ˇaˇk)∈supp(ˇL) be a PFICP. Then, by definition,
If possible, assume that (ˇa,ˇaˇk) is not a SIP. Then, σˇL(ˇa,ˇaˇk)<ICONNGσ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk), and μˇL(ˇa,ˇaˇk)>ICONNGμ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk). Let ˇP be a strongest ˇa−ˇaˇk IPt in ˇG∖{(ˇa,ˇaˇk)}, and then ˇP together with (ˇa,ˇaˇk) forms a PFIC whose weakest IP is (ˇa,ˇaˇk). By Proposition 4.2, it is impossible since (ˇa,ˇaˇk) is a PFICP. □
Proposition 5.5. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. An IP (ˇa,ˇaˇk) in ˇG is a SIP if and only if σˇL(ˇa,ˇaˇk)=ICONNσ(ˇG)(ˇa,ˇaˇk),μˇL(ˇa,ˇaˇk)=ICONNμ(ˇG)(ˇa,ˇaˇk).
Proof. Assume that (ˇa,ˇaˇk)∈supp(ˇL) is a SIP. Since ˇP:ˇa,(ˇa,ˇaˇk),ˇaˇk is a IPt between ˇa and ˇaˇk, ICONNGσ(ˇG)(ˇa,ˇaˇk)≥σˇL(ˇa,ˇaˇk),ICONNGμ(ˇG)(ˇa,ˇaˇk)≤μˇL(ˇa,ˇaˇk). If ˇP is a unique IPt between ˇa and ˇaˇk, the result is trivial. Now, let ˇQ be another IPt between ˇa and ˇaˇk in ˇG. Then, ˇQ is an IPt in ˇG∖{(ˇa,ˇaˇk)} such that σISˇQ≤ICONNGσ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk),μISˇQ≥ICONNGμ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk). Since (ˇa,ˇaˇk) is a SIP, ICONNGσ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk)≤σˇL(ˇa,ˇaˇk), and ICONNGμ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk)≥μˇL(ˇa,ˇaˇk). Thus, σˇL(ˇa,ˇaˇk)≥σISˇQ and μˇL(ˇa,ˇaˇk)≤μISˇQ. σˇL(ˇa,ˇaˇk)=σISˇP and μˇL(ˇa,ˇaˇk)=μISˇP. Thus, ICONNσ(ˇG)(ˇa,ˇaˇk)=σˇL(ˇa,ˇaˇk) and ICONNμ(ˇG)(ˇa,ˇaˇk)=μˇL(ˇa,ˇaˇk).
Conversely, If σˇL(ˇa,ˇaˇk)=ICONNσ(ˇG)(ˇa,ˇaˇk) and μˇL(ˇa,ˇaˇk)=ICONNμ(ˇG)(ˇa,ˇaˇk), then
Hence, (ˇa,ˇaˇk) is a SIP. □
Proposition 5.6. Let ˇG=(ˇJ,ˇK,ˇL) be a connected PFIG. Then, there exists a SIP between any vertex ˇa and edge ˇkˇl in ˇG.
Proposition 5.7. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG. An IP (ˇa,ˇaˇk) is a PFICP if and only if it is an α-SIP.
Proof. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG and let (ˇa,ˇaˇk)∈supp(ˇL) be a PFICP of ˇG. Then, by Proposition 4.2, ICONNGσ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk)<σˇL(ˇa,ˇaˇk), and ICONNGμ(ˇG∖{(ˇa,ˇaˇk)})(ˇa,ˇaˇk)>μˇL(ˇa,ˇaˇk). Thus, (ˇa,ˇaˇk) is an α-SIP.
Conversely, assume that (ˇa,ˇaˇk) is an α-SIP. Then, by definition, ˇP:ˇa,(ˇa,ˇaˇk),ˇaˇk is the unique strongest ˇa−ˇaˇk IPt, and so
Hence, (ˇa,ˇaˇk) is a PFICP. □
Proposition 5.8. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIT. An IP (ˇq,ˇqˇt) in ˇG is a SIP if and only if (ˇq,ˇqˇt)∈ˇT, where ˇT=(ˇM,ˇQ,ˇS) is a tree in the definition of PFIT.
Proof. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIT and let ˇT=(ˇM,ˇQ,ˇS) be a tree such that for every (ˇa,ˇaˇk)∉ˇT, σˇL(ˇa,ˇaˇk)<ICONNσ(ˇT)(ˇa,ˇaˇk), and μˇL(ˇa,ˇaˇk)>ICONNμ(ˇT)(ˇa,ˇaˇk). Let (ˇq,ˇqˇt) be a strong IP. If (ˇq,ˇqˇt)∉ˇT, then σˇL(ˇq,ˇqˇt)<ICONNσ(ˇT)(ˇq,ˇqˇt), and μˇL(ˇq,ˇqˇt)>ICONNμ(ˇT)(ˇq,ˇqˇt). Since (ˇq,ˇqˇt) is a SIP,
Since (ˇq,ˇqˇt)∉ˇT, ˇT is PFIS of ˇG∖(ˇq,ˇqˇt), and hence
which is a contradiction.
Conversely, assume that (ˇq,ˇqˇt)∈supp(ˇL). By Proposition 4.1, (ˇq,ˇqˇt) is a PFICP. By Proposition 5.7, (ˇq,ˇqˇt) is a SIP. □
Proposition 5.9. A connected PFIG is a PFIT if and only if it has no β-SIP.
Proof. Let ˇG=(ˇJ,ˇK,ˇL) be a connected PFIG. If ˇG is PFIT, then ∃ a unique ˇT=(ˇM,ˇQ,ˇS) so that every (ˇa,ˇaˇk)∈ˇT is PFICP and hence is an α-SIP. By definition of a PFIT, all (ˇq,ˇqˇt) such that (ˇq,ˇqˇt)∈ˇG but (ˇq,ˇqˇt)∉ˇT are δ-weak IPs. Thus, ˇG has no β-SIP.
Conversely, assume that ˇG has β-SIP. If ˇG has no cycles, then ˇG is a PFIT. Now, let C be a cycle in ˇG. Then, C has only an α-SIP and δ-weak IP. All pairs of C cannot be α-SIPs. Hence, in every cycle of ˇG ∃ a unique δ-weak IP. By Proposition 3.1, ˇG is a PFIT. □
Proposition 5.10. Let ˇG=(ˇJ,ˇK,ˇL) be a connected PFIG. ˇG is a PFIT if and only if there exists a unique strong IPt between any vertex and edge. In particular, this IPt will be an α-strong IPt.
Proof. Let ˇG be a PFIT. Then, by Proposition 5.6, there exists a strong IPt ˇP between any vertex ˇa∈supp(ˇJ) and edge ˇkˇl∈supp(ˇK). By Proposition 5.8, this IPt ˇP lies entirely in the associated maximum spanning tree ˇT=(ˇM,ˇQ,ˇS), and such IPt is unique since ˇT is a tree. Since ˇT has no β-SIPs, this IPt will be an α-strong IPt.
Conversely, suppose ∃ a unique strong IPt between any vertex and edge of ˇG. Let ˇG be not a PFIT, and then there is a cycle C in ˇG such that for every (ˇa,ˇaˇk)∈C,
That is, every IP (ˇa,ˇaˇk)∈ˇL is a SIP, which is a contradiction to our assumption that the strong IPt is unique. □
Proposition 5.11. Let ˇG=(ˇJ,ˇK,ˇL) be a PFIG and let ˇP be a strong IPt between ˇq and ˇqˇt. Then, ˇP is a strongest ˇq−ˇqˇt IPt in the following cases:
(1) ˇP contains only α-SIPs. {(1)} ˇP is a unique strong ˇq−ˇqˇt IPt.
(2) Incidence strengths of all ˇq−ˇqˇt IPts in ˇG are equal.
Proposition 5.12. A complete PFIG has no δ-weak IPs.
Proof. Suppose that (ˇq,ˇqˇt) is a δ-weak IP of a complete PFIG ˇG=(ˇJ,ˇK,ˇL). Then,
Thus, there exists a stronger IPt ˇP than path ˜P:ˇq,(ˇq,ˇqˇt),ˇqˇt between ˇq and ˇqˇt in ˇG. Then,
That is, for every (ˇa,ˇaˇk)∈ˇP, σˇL(ˇa,ˇaˇk)>σˇL(ˇq,ˇqˇt), and μˇL(ˇa,ˇaˇk)<μˇL(ˇq,ˇqˇt). Let ˇw be the first vertex in ˇP after ˇq. Then σˇL(ˇq,ˇqˇw)>σˇL(ˇq,ˇqˇt), and μˇL(ˇq,ˇqˇw)<μˇL(ˇq,ˇqˇt). This is not possible as σˇL(ˇq,ˇqˇw)=min{σˇJ(u),σˇK(uw)}=σˇL(ˇq,ˇqˇt), and μˇL(ˇq,ˇqˇw)=max{μˇJ(u),μˇK(uw)}=μˇL(ˇq,ˇqˇt). Thus, ˇG has no δ-weak IP. □
6.
Application: Recognition of countries participating in illegal wildlife trade
Graph theory has expanded greatly as a result of a wide variety of applications in optimization issues, combinatorial issues, linguistics, chemistry, physics and other fields. In this section, we describe a real-world application of PFIGs.
Due to the numerous benefits of wildlife on human life, wildlife trading is getting more popular with the increase of population. People in many countries are accustomed to a lifestyle that fuels the demand for wildlife. Wildlife crime is seen as a low-risk, high-reward dirty industry, currently estimated to be the fourth most profitable global crime, after the trafficking of drugs, humans and firearms. From the Americas to Asia to Africa, wildlife trade is unfortunately still common in many continents. It is a big business, bringing in estimated billions of dollars of illegal revenue. The golden triangle of Laos, Thailand and Myanmar is a global hub for illegal wildlife trade and trafficking. China is the largest importer of illegal wildlife and animal products, driving demands for animals from around the world. Wildlife trade alone is a major threat to some species, but its impact is frequently made worse by habitat loss and other pressures. Criminals mostly devise many creative ways to transport illegal wildlife focusing on the path of processing and sale. It is difficult for law enforcement to find concealed ways. So, police should be aware of the latest trends in the illegal wildlife trade market. We can use the PFIG to highlight the safest path chosen by dangerous international networks for illegal wildlife trade between two countries and can also tell the removal of which country reduces the safety of that path. Consider a few countries of the world, where illegal wildlife trade is a major threat to wildlife, in the following set: ˇW={SouthAfrica,Mozambique,Kenya,Uganda,China,Myanmar,Thailand,Laos,Vietnam}. PFS ˇJ defined on set ˇW is presented in Table 2.
In Table 2, σˇJ indicates law enforcement efforts of the country for illegal wildlife trade, μˇJ indicates the involvement of the country in organized illegal wildlife trade, and the neutral approach of the country to illegal wildlife trade can be considered as a degree of indeterminacy. We define PFS ˇK on ˇZ⊆ˇWסW in Table 3. An element of PFS ˇK represents illegal wildlife trade between those two countries.
In Table 3, σˇK indicates the rate of illegal wildlife trade between countries, and μˇK indicates the rate of the world's negative thinking or disliking for that illegal wildlife trade. Membership and non-membership values of each pair of countries are according to σˇK(ˇaˇk)≤min{σˇJ(ˇa),σˇJ(ˇk)},μˇK(ˇaˇk)≤max{μˇJ(ˇa),μˇJ(ˇk)} and (σˇK(ˇaˇk))2+(μˇK(ˇaˇk))2≤1 for all ˇa,ˇk∈ˇW. Let us use the following alphabets for country names:
S = South Africa, MZ = Mozambique, K = Kenya, U = Uganda, C = China, M = Myanmar, TH = Thailand, LS = Laos, VT = Vietnam. Now, we define a PFS ˇL on ˇY⊆ˇWסZ in Table 4.
Trafficking routes for illegal wildlife frequently do not follow direct lines between source and destination countries; they can be circuitous and involve multiple transit stages. Let σˇL(VT,(VT,TH)) and μˇL(VT,(VT,TH)) represent the degree of safety and degree of risk for illegal wildlife trade, respectively, to use Vietnam as a source country, travel on (VT,TH) and arrive at destination country Thailand. Similarly, the membership and non-membership values of the other pairs of PFIGs are shown in Table 4. Membership and non-membership values of each pair of countries are according to σˇL(ˇa,ˇaˇk)≤min{σˇJ(ˇa),σˇK(ˇa,ˇaˇk)},μˇL(ˇa,ˇaˇk)≤max{μˇJ(ˇa),μˇK(ˇa,ˇaˇk)} and (σˇL(ˇa,ˇaˇk))2+(μˇL(ˇa,ˇaˇk))2≤1 for all (ˇa,ˇaˇk)∈ˇWסZ. PFIG ˇG=(ˇJ,ˇK,ˇL) is shown in Figure 13.
To make clear how the values in Table 5 are obtained, we show a calculation example for the bold entry (K,(K,C)) in Table 5. The possible incidence paths are as follows:
The incidence strengths of these IPts are given by
The σ-incidence strength and μ-incidence strength of connectedness are given by
Thus, ICONNˇG(K,(K,C))=(ICONNσ(ˇG)(K,(K,C)),ICONNμ(ˇG)(K,(K,C)))=(0.5,0.3).
If we remove the pair (K,(K,C)) from the graph, then
σˇL(K,(K,C))>ICONNσ(ˇG∖{(K,(K,C))})(K,(K,C)),andμˇL(K,(K,C))<ICONNμ(ˇG∖{(K,(K,C))})(K,(K,C)). Thus, (K,(K,C)) is an α-SIP and so by Proposition 5.7 is a PFICP of ˇG.
σˇL(ˇa,ˇaˇk), and μˇL(ˇa,ˇaˇk) represent the degree of safety and degree of risk of IPts between ˇa and ˇk. Then, ICONNˇG(ˇa,ˇaˇk) represents the IPt with the highest safety and smallest risk among all such IPts. Hence such an IPt is the safest path to travel.
The incidence strength of connectedness between a vertex ˇa and an edge ˇaˇk of ˇG=(ˇJ,ˇK,ˇL) are calculated in Table 5. Suppose (ˇq,ˇqˇt) is a PFICP. Then, (ˇa,ˇaˇk) exists such that ICONNGσ(ˇG∖{(ˇq,ˇqˇt)})(ˇa,ˇaˇk)<ICONNGσ(ˇG)(ˇa,ˇaˇk) and ICONNGμ(ˇG∖{(ˇq,ˇqˇt)})(ˇa,ˇaˇk)>ICONNGμ(ˇG)(ˇa,ˇaˇk). Thus, the removal of pair (ˇq,ˇqˇt) would make the path less safe. There are two safest IPts between from VT to (VT, M) with ICONNˇG(VT,(VT,M))=(0.2,0.8) such as:
These paths are shown in Figure 14. Let us remove PFICP (C,(C,K)) from ˇG. Then, ICONNG(ˇG−{(C,(C,K))})(VT,(VT,M))=(0.1,0.9). Thus, removal of (C,(C,K)) reduces the safety of the path. Similarly, removal of pairs (C,(C,LS)),(U,(U,S)),(K,(K,C)),(M,(M,S)),(K,(K,U)),(U,(U,K)),(M,(M,TH)), (TH,(M,TH)),(V,(V,C)),(C,(C,V)),(U,(U,MZ)),(M,(M,MZ)),(S,(S,K)) and (M,(M,VT)) reduce the safety of the path.
7.
Comparative analysis
In this section, we compare our suggested model to existing models to determine its validity and superiority. In comparison to FS and IFS, PFS is a more powerful model for handling uncertainty. FSs address the ambiguity of belongingness, whereas IFSs provide information regarding the hesitancy of a statement. A PFS effectively handles uncertain data by extending the range for the assignment of membership and non-membership values. In Figure 13, a PFIG indicates a network of illegal wildlife trade in nine different countries, South Africa, Mozambique, Kenya, Uganda, China, Myanmar, Thailand, Laos and Vietnam. The edge between any two countries represents the illegal wildlife trade between those two countries, and the membership and non-membership values indicate the rate of illegal wildlife trade between countries and the rate of the world's negative thinking or disliking of that illegal wildlife trade, respectively. Moreover, let σˇL(C1,(C1C2)) and μˇL(C1,(C1C2)), represent the degree of safety and degree of risk for illegal wildlife trade, respectively, to use C1 as a source country, travel on (C1C2) and arrive at destination country C2. The ICONNˇG(C1,(C1C2)) represents the IPt with the highest safety and smallest risk among all such IPts. In the case of graphs, incidence pairs are not present, and thus graphs do not provide any information about the safety of the path. In the case of incidence graphs, the membership value of each incidence pair is 1, and thus ICONNˇG(C1,(C1C2))=1 for all incidence pairs of the network. Hence, we cannot determine the safest path to travel. In the case of fuzzy incidence graphs, the nonmembership value of each pair is missing, and thus the ICONNˇG(C1,(C1C2)) does not provide any information about how risky the path is to travel on. In the case of IFIGs, σˇJ(C1)+μˇJ(C1)≤1, for each C1. Thus, in the case of (SouthAfrica,0.5,0.7), 0.5+0.7≰1, and hence IFIGs do not have the ability to handle such information. Similarly, all the other concepts in fuzzy incidence graph structure as fuzzy incidence cycle, fuzzy incidence tree, fuzzy incidence cut vertices, fuzzy incidence bridges, fuzzy incidence cut pairs, and strong incidence pairs, use only the membership value of the problem. Nevertheless, in real-life scenarios where the non-satisfactory factor is also present, the fuzzy incidence graph model fails to illustrate non-satisfactory factors along with satisfactory factors. In comparison to fuzzy incidence graphs, Pythagorean fuzzy incidence graphs offer a more comprehensive description of relationships between objects. Problems where an element does not belong to a particular subset or where the degree of exclusion varies, can be modeled using nonmembership values. This adaptability enables more precise modeling of challenging real-world scenarios.
8.
Conclusions
One of the key factors that affect a network is connectivity. Particularly in real life, connectivity is essential for problems like internet routing and transport network flow. We introduce the idea of PFIGs in this article. We discussed the strength IPt between a vertex and an edge in PFIGs. We also proposed the concepts of PFICs and PFITs and provided some important related results. We illustrated the notions of PFICVs and PFICPs in PFIG and proved essential propositions concerning PFICPs. We discussed α-strong, β-strong and δ-weak IPs. We proved that any vertex and edge have a strong IPt between them. Moreover, we used strong pairs to characterize various Pythagorean incidence structures. We also obtained the characterization of PFICs using α-SIPs and determined the relation between PFITs and α-SIPs. Finally, we provided an application of PFIGs in the illegal wildlife trade network.
This research work can be further extended to include the analysis of (1) connectivity indices of PFIGs, (2) cyclic connectivity index of PFIGs, (3) average cyclic connectivity index of PFIGs and (4) Wiener index of PFIGs.
Acknowledgments
The authors extend their appreciation to the Deanship of Scientific Research at King Khalid University for funding this work through the Large Group Research Project under grant number (R.G.P.2/181/44).
Conflict of interest
The authors declare no conflict of interest.