Abstract
Information in many real life problems collect from multi agents, i.e., "multipolar information" exists. This multipolar information cannot be properly modeled by m- polar fuzzy graph or intutionistic fuzzy graph. An m-polar neutrosophic model is very much efficient for such real word problems which can construct more precise, flexible, and comparable system as compared to the classical, fuzzy and neutrosophic graph models. In this paper, we present the definition of m-polar neutrosophic graph model. Some new operations, such as union, join, composition and ring sum of two m-polar neutrosophic graph are defined here. We define six new products on m-polar neutrosophic graphs namely strong product, semi strong product, complete product, direct product, cartesian product and lexicographic product. Some idea of complement, isomorphism, weak and co weak isomorphism on m-polar neutrosophic graph are introduced here. We also present several associated properties and theorems of m-polar neutrosophic graph. We introduce a model of m-polar neutrosophic graph, which is applied in evaluating the teacher’s performance of a college. The performances of the teachers are computed based on the response score (feedback) of the students of the college. We also present a numerical example to illustrate our proposed model.
Introduction
Zadeh [28] introduced the concept of fuzzy set theory as the advanced version of classical set theory to handle many real world problems with uncertainties. Fuzzy set theory is a powerful mathematical tool for dealing with incomplete, inexact, indeterminate and inconsistent information in real world problem. Zhang [29] proposed the concept of bipolar fuzzy sets for handling positive and negative both situation in a given real-world problem. But there are some cases that we cannot solve with the concept of bipolar fuzzy set because they involve multi-agent, multi-attribute, multi-object, multi-index, multi-polar information.
For solving this type of problem, Juanjuan Chen et al. [13] extended the idea of the bipolar fuzzy set to m-polar fuzzy set [7, 15– 19]. Due to the complexity of facts and opacity of the human mind fuzzy set theory is not sufficient to solve many real-life problem. We cannot properly describe the concept of non-membership term in Zadeh’s fuzzy environment. To overcome this shortcoming of the fuzzy set, Atanassov [8] define the idea of intuitionistic fuzzy set(IFS), which is more efficient to handle real life problem compare to fuzzy set. Intuitionistic fuzzy sets handle incomplete information i.e., the grade of membership function and non-membership function but not the indeterminate information and inconsistent information which exists obviously in belief system.
To overcome the demerits of IFS, Smarandache [10, 25] neutrosophic set has been introduced as a remedy. A single-valued neutrosophic set has three components: truth membership degree, indeterminacy membership degree and falsity membership degree. Many papers about neutrosophic set theory and their application have been done by various researchers [1, 22]. In a neutrosophic set, the membership value is associated with truth, false and indeterminacy degrees but there is no restriction on their sum. Deli et al. [2, 14] extended the ideas of bipolar fuzzy sets and neutrosophic sets to bipolar neutrosophic sets and studied its operations and applications in decision making problems.
In 1736, as a new branch of mathematics Graph theory was born with the Eulers paper in which he solved the famous Konigsberg bridge problem. A graph G is defined by a ordered pair of two sets namely vertex set V (G) and edge set E (G) respectively, such that |V (G) | = n and |E (G) | = m. The degree of a vertex v is the number of vertices in G which are connected to v by an edge and denoted by dG(v). There are a very wide application of graph theory in computer science, image segmentation, clustering, image capturing, netwoking. Single-valued neutrosophic graph is the generalization of graphs, fuzzy graphs and intuitionistic graphs. Akram and Shahzadi [3] defined some operations on single-valued neutrosophic graph and they also showed the representation of graphs using neutrosophic soft sets in [4]. J. Ye [27] described single-valued neutrosophic minimum spanning tree and its clustering method.
Graph is an efficient method for modeling many combinatorial optimization problems especially in areas of expert system, neural network, medical diagnosis, operations research and computer science. The optimization problems in real life scenarios generally involve with multiple agents, multiple attributes, multiple objects and multiple indexes and information about the problems are generated from multiple sources, i.e., multi-polar information. We cannot represent this information properly using bipolar fuzzy graphs or m-polar fuzzy graphs.
The motivation behind the work in this paper is to introduce a new neutrosophic graph model which will be simple enough and effective to model the uncertainties in the real world problems. In this manuscript, we define the m-polar neutrosophic graph and several operations/products between two m-polar neutrosophic graph. These operations/products are illustrated by some examples. Some properties of complement, isomorphism, weak and co weak isomorphism of m-polar neutrosophic graph are presented here. We describe an application of m-polar neutrosophic graph in evaluating the professor performance of a college.
The paper is organized as follows. Section 2 briefly describes some basic ideas and definitions on fuzzy set, fuzzy graph, neutrosophic Set, single-valued neutrosophic graph(SVNG) and complete SVNG. In Section 3, we present the idea of m-polarSVNG. The corresponding some operation on m-polar SVNG are also presented in this Section 4. We present some products on m-polar neutrosophic graphs in Section 5. An application of m-polar SVNG is presented in the Section 6. Finally, we conclude in Section 7.
Preliminaries
We note that, if Γ is not a symmetric relation on V G , then G is called directed fuzzy graph.
Here T N : X → [0, 1], I N : X → [0, 1], F N : X → [0, 1] are called truth membership function, indeterminacy membership function and falsity membership function respectively.
0 ≤ T
ϒ
(ξ) + I
ϒ
(ξ) + F
ϒ
(ξ) ≤3, ∀ξ ∈ V
G
; T
Γ
: V
G
× V
G
→ [0, 1], I
Γ
: V
G
× V
G
→ [0, 1] and F
Γ
: V
G
× V
G
→ [0, 1] are the truth membership function, indeterminacy membership function and falsity membership fuction of the edge (ξ, η) respectively, s.t., T
Γ
(ξη) ≤ T
ϒ
(ξ) ∧ T
ϒ
(η), I
Γ
(ξη) ≤ I
ϒ
(ξ) ∧ I
ϒ
(η) and F
Γ
(ξη) ≤ F
ϒ
(ξ) ∨ F
ϒ
(η), ∀ ξη ∈ E
G
.
The m-polar SVNG
Throughout the paper [0, 1] r (r-th power of [0, 1]) is considered a poset with the point-wise order ≤, where r is an arbitrary ordinal number (we make an appointment that m = {n|n < m} when m > 0), ≤ is defined by u ≤ v ⇔ ϕ j (u) ≤ ϕ j (y) for each j ∈ m (u, v ∈ [0, 1] m ) and ϕ j : [0, 1] m → [0, 1] is the j-th projection mapping (j ∈ m).
ϕ
j
(Γ (xy)) ≤ ϕ
j
(ϒ (x)) ∧ ϕ
j
(ϒ (y)),
for each j = 1, 2, 3, . . . , m}.
Here T N : X → [0, 1] m , I N : X → [0, 1] m , F N : X → [0, 1] m are called truth membership function, indeterminacy membership function and falsity membership function respectively.
0 ≤ ϕ
j
(T
ϒ
(ξ)) + ϕ
j
(I
ϒ
(ξ)) + ϕ
j
(F
ϒ
(ξ)) ≤3, ∀ξ ∈ V
G
;
Some operation on m-polar SVNG
∀ (α, β) ∈ V
G
× V
H
, ϕ
j
((T
ϒ
1
× T
ϒ
2
) (α, β)) = ϕ
j
(T
ϒ
1
(α)) ∧ ϕ
j
(T
ϒ
2
(β)); ϕ
j
((I
ϒ
1
× I
ϒ
2
) (α, β)) = ϕ
j
(I
ϒ
1
(α)) ∧ ϕ
j
(I
ϒ
2
(β)); ϕ
j
((F
ϒ
1
× F
ϒ
2
) (α, β)) = ϕ
j
(F
ϒ
1
(α)) ∨ ϕ
j
(F
ϒ
2
(β)); ∀γ ∈ V
G
and ∀αβ ∈ E
H
, ϕ
j
((T
Γ
1
× T
Γ
2
) ((γ, α) (γ, β))) = ϕ
j
(T
ϒ
1
(γ)) ∧ ϕ
j
(T
Γ
2
(αβ)); ϕ
j
((I
Γ
1
× I
Γ
2
) ((γ, α) (γ, β))) = ϕ
j
(I
ϒ
1
(γ)) ∧ ϕ
j
(I
Γ
2
(αβ)); ϕ
j
((F
Γ
1
× F
Γ
2
) ((γ, α) (γ, β))) = ϕ
j
(F
ϒ
1
(γ)) ∨ ϕ
j
(F
Γ
2
(αβ)); ∀γ ∈ V
H
and ∀αβ ∈ E
G
, ϕ
j
((T
Γ
1
× T
Γ
2
) ((α, γ) (β, γ))) = ϕ
j
(T
Γ
1
(αβ)) ∧ ϕ
j
(T
ϒ
2
(γ)); ϕ
j
((I
Γ
1
× I
Γ
2
) ((α, γ) (β, γ))) = ϕ
j
(I
Γ
1
(αβ)) ∧ ϕ
j
(I
ϒ
2
(γ)); ϕ
j
((F
Γ
1
× F
Γ
2
) ((α, γ) (β, γ))) = ϕ
j
(F
Γ
1
(αβ)) ∨ ϕ
j
(F
ϒ
2
(γ));
ϕ
j
((T
Γ
1
× T
Γ
2
) ((α, β) (γ, δ))) =0; ϕ
j
((I
Γ
1
× I
Γ
2
) ((α, β) (γ, δ))) =0; ϕ
j
((F
Γ
1
× F
Γ
2
) ((α, β) (γ, δ))) =0.
Here E = {((α, γ) (β, γ)) : αβ ∈ E
G
, γ ∈ V
H
} ∪ {((α, γ) (α, δ)) : α ∈ V
G
, γδ ∈ E
H
}.
Compute G × H. Then we have the following:
(T Γ G × T Γ H ) ((α, γ) (α, δ)) = <0.3, 0.4, 0.6 >,
(I Γ G × I Γ H ) ((α, γ) (α, δ)) = <0.4, 0.5, 0.4 >,
(F Γ G × F Γ H ) ((α, γ) (α, δ)) = <0.5, 0.5, 0.7 >.
Similarly, we get (T ϒ G × T ϒ G ) (α, γ) = <0.3, 0.4, 0.6 >. See Fig. 1.

CP of two 3-polar SVNG’s.
∴ ϕ j ((T Γ G × T Γ H ) ((γ, α) (γ, β))) ≤ ϕ j ((T ϒ G × T ϒ H ) (γ, α)) ∧ ϕ j ((T ϒ G × T ϒ H ) (γ, β)).
Let γ ∈ V H and αβ ∈ E G . Then for each j = 1, 2, 3, . . . , m,
∴ ϕ j ((T Γ G × T Γ H ) ((α, γ) (β, γ))) ≤ ϕ j ((T ϒ G × T ϒ H ) (α, γ)) ∧ ϕ j ((T ϒ G × T ϒ H ) (β, γ)).
Let
ϕ j ((T Γ G × T Γ H ) ((α, β) (γ, δ))) =0 ≤ ϕ j ((T ϒ G × T ϒ H ) (α, β)) ∧ ϕ j ((T ϒ G × T ϒ H ) (γ, δ)).
Let γ ∈ V G and αβ ∈ E H . Then for each j = 1, 2, 3, . . . , m,
∴ ϕ j ((I Γ G × I Γ H ) ((γ, α) (γ, β))) ≤ ϕ j ((I ϒ G × I ϒ H ) (γ, α)) ∧ ϕ j ((I ϒ G × I ϒ H ) (γ, β)).
Let γ ∈ V H and αβ ∈ E G . Then for each j = 1, 2, 3, . . . , m,
∴ ϕ j ((I Γ G × I Γ H ) ((α, γ) (β, γ))) ≤ ϕ j ((I ϒ G × I ϒ H ) (α, γ)) ∧ ϕ j ((I ϒ G × I ϒ H ) (β, γ)).
Let
ϕ j ((I Γ G × I Γ H ) ((α, β) (γ, δ))) =0 ≤ ϕ j ((I ϒ G × I ϒ H ) (α, β)) ∧ ϕ j ((I ϒ G × I ϒ H ) (γ, δ)).
Let γ ∈ V G and αβ ∈ E H . Then for each j = 1, 2, 3, . . . , m,
∴ ϕ j ((F Γ G × F Γ H ) ((γ, α) (γ, β))) ≤ ϕ j ((F ϒ G × F ϒ H ) (γ, α)) ∨ ϕ j ((F ϒ G × F ϒ H ) (γ, β)).
Let γ ∈ V H and α, β ∈ E G . Then for each j = 1, 2, 3, . . . , m,
∴ ϕ j ((F Γ G × F Γ H ) ((α, γ) (β, γ))) ≤ ϕ j ((F ϒ G × F ϒ H ) (α, γ)) ∧ ϕ j ((F ϒ G × F ϒ G ) (β, γ)).
Let
ϕ j ((F Γ G × F Γ H ) ((α, β) (γ, δ))) = 0 ≤ ϕ j ((F ϒ G × F ϒ H ) (α, β)) ∧ ϕ j ((F ϒ G × F ϒ H ) (γ, δ)).
This completes the proof. □
∀ (α, β) ∈ V
G
× V
H
, ϕ
j
((T
ϒ
G
∘ T
ϒ
H
) (α, β)) = ϕ
j
(T
ϒ
G
(α)) ∧ ϕ
j
(T
ϒ
H
(β)); ϕ
j
((I
ϒ
G
∘ I
ϒ
H
) (α, β)) = ϕ
j
(I
ϒ
G
(α)) ∧ ϕ
j
(I
ϒ
H
(β)); ϕ
j
((F
ϒ
G
∘ F
ϒ
H
) (α, β)) = ϕ
j
(F
ϒ
G
(α)) ∨ ϕ
j
(F
ϒ
H
(β)); ∀γ ∈ V
G
and ∀ (α, β) ∈ E
H
, ϕ
j
((T
Γ
G
∘ T
Γ
H
) ((γ, α) (γ, β))) = ϕ
j
(T
ϒ
G
(γ)) ∧ ϕ
j
(T
Γ
H
(α, β)); ϕ
j
((I
Γ
G
∘ I
Γ
H
) ((γ, α) (γ, β))) = ϕ
j
(I
ϒ
G
(γ)) ∧ ϕ
j
(I
Γ
H
(α, β)); ϕ
j
((F
Γ
G
∘ F
Γ
H
) ((γ, α) (γ, β))) = ϕ
j
(F
ϒ
G
(γ)) ∨ ϕ
j
(F
Γ
H
(α, β)); ∀γ ∈ V
H
and ∀αβ ∈ E
G
, ϕ
j
((T
Γ
G
∘ T
Γ
H
) ((α, γ) (β, γ))) = ϕ
j
(T
Γ
G
(αβ)) ∧ ϕ
j
(T
ϒ
H
(γ)); ϕ
j
((I
Γ
G
∘ I
Γ
H
) ((α, γ) (β, γ))) = ϕ
j
(I
Γ
G
(αβ)) ∧ ϕ
j
(I
ϒ
H
(γ)); ϕ
j
((F
Γ
G
∘ F
Γ
H
) ((α, γ) (β, γ))) = ϕ
j
(F
Γ
G
(αβ)) ∨ ϕ
j
(F
ϒ
H
(γ)); ∀ (α, β) (γ, δ) ∈ E0 - E, ϕ
j
((T
Γ
G
∘ T
Γ
H
) ((α, β) (γ, δ))) = ϕ
j
(T
Γ
G
(αγ)) ∧ ϕ
j
(T
ϒ
H
(β)) ∧ ϕ
j
(T
ϒ
H
(δ)); ϕ
j
((I
Γ
G
∘ I
Γ
H
) ((α, β) (γ, δ))) = ϕ
j
(I
Γ
G
(αγ)) ∧ ϕ
j
(I
ϒ
H
(β)) ∧ ϕ
j
(I
ϒ
H
(δ)); ϕ
j
((F
Γ
G
∘ F
Γ
H
) ((α, β) (γ, δ))) = ϕ
j
(F
Γ
G
(αγ)) ∨ ϕ
j
(F
ϒ
H
(β)) ∨ ϕ
j
(F
ϒ
H
(δ));
ϕ
j
((T
Γ
G
∘ T
Γ
H
) ((α, β) (γ, δ))) =0; ϕ
j
((I
Γ
G
∘ I
Γ
H
) ((α, β) (γ, δ))) =0; ϕ
j
((F
Γ
G
∘ F
Γ
H
) ((α, β) (γ, δ))) =0.
Here E = {((α, γ) (β, γ)) : αβ ∈ E
G
, γ ∈ V
H
} ∪ {((α, γ) (α, δ)) : α ∈ V
G
, γδ ∈ E
H
} and E0 = E ∪ {((α, γ) (β, δ)) : αγ ∈ E
G
γ ≠ δ ∈ V
H
} .
Compute G1 ∘ G2. Then we have the following:
(T Γ G ∘ T Γ H ) ((x1, y1) (x1, y2)) = <0.4, 0.5, 0.4 >,
(I Γ G ∘ I Γ H ) ((x1, y1) (x1, y2)) = <0.3, 0.3, 0.1 >,
(F Γ G ∘ F Γ H ) ((x1, y1) (x1, y2)) = <0.4, 0.5, 0.6 >.
Similarly, we get
(T ϒ G ∘ T ϒ H ) (x1, y1) = <0.6, 0.5, 0.4 >,
(I ϒ G ∘ I ϒ H ) (x1, y1) = <0.4, 0.4, 0.5 >,
(F ϒ G ∘ F ϒ H ) (x1, y1) = <0.5, 0.4, 0.5 >. See Fig. 2.

Composition of two 3-polar SVNG’s.
γ ∈ V
H
} ∪ {((α, γ) (α, δ)) : α ∈ V
G
, γδ ∈ E
H
} and E0 = E ∪ {((α, γ) (β, δ)) : αγ ∈ E
G
γ ≠ δ ∈ V
H
} . Let γ ∈ V
G
and αβ ∈ E
H
. Then for each j = 1, 2, 3, . . . , m,
Let γ ∈ V
H
and αβ ∈ E
G
. Then for each j = 1, 2, 3, . . . , m,
Let (α, β) (γ, δ) ∈ E0 - E. Then for each j = 1, 2, 3, . . . , m,
Let
ϕ j ((T Γ G ∘ T Γ H ) ((α, β) (γ, δ))) =0 ≤ ϕ j ((T ϒ G ∘ T ϒ H ) (α, β)) ∧ ϕ j ((T ϒ G ∘ T ϒ H ) (γ, δ)).
Let γ ∈ V G and αβ ∈ E H . Then for each j = 1, 2, 3, . . . , m,
∴ ϕ j ((I Γ G ∘ I Γ H ) ((γ, α) (γ, β))) ≤ ϕ j ((I ϒ G ∘ I ϒ H ) (γ, α)) ∧ ϕ j ((I ϒ G ∘ I ϒ H ) (γ, β)).
Let γ ∈ V
H
and αβ ∈ E
G
. Then for each j = 1, 2, 3, . . . , m,
Let
ϕ j ((I Γ G ∘ I Γ H ) ((α, β) (γ, δ))) =0 ≤ ϕ j ((I ϒ G ∘ I ϒ H ) (α, β)) ∧ ϕ j ((I ϒ G ∘ I ϒ H ) (γ, δ)).
Let (α, β) (γ, δ) ∈ E0 - E. Then for each j = 1, 2, 3, . . . , m,
Let γ ∈ V
G
and αβ ∈ E
H
. Then for each j = 1, 2, 3, . . . , m,
Let γ ∈ V
H
and αβ ∈ E
G
. Then for each j = 1, 2, 3, . . . , m,
Let (α, β) (γ, δ) ∈ E0 - E. Then for each j = 1, 2, 3, . . . , m,
Let
Then ϕ j ((F Γ G ∘ F Γ H ) ((α, β) (γ, δ))) =0 ≤ ϕ j ((F ϒ G ∘ F ϒ H ) (α, β)) ∨ ϕ j ((F ϒ G ∘ F ϒ H ) (γ, δ)).
This completes the proof. □
ϕ
j
((T
Γ
G
∪ T
Γ
H
) (ζη)) =0, ϕ
j
((I
Γ
G
∪ I
Γ
H
) (ζη)) =0, ϕ
j
((F
Γ
G
∪ F
Γ
H
) (ζη)) =0.
if ξ ∈ V
G
∪ V
H
, ϕ
j
((T
ϒ
G
+ T
ϒ
H
) (ξ)) = ϕ
j
((T
ϒ
G
∪ T
ϒ
H
) (ξ)); ϕ
j
((I
ϒ
G
+ I
ϒ
H
) (ξ)) = ϕ
j
((I
ϒ
G
∪ I
ϒ
H
) (ξ)); ϕ
j
((F
ϒ
G
+ F
ϒ
H
) (ξ)) = ϕ
j
((F
ϒ
G
∪ F
ϒ
H
) (ξ)) ; if ξς ∈ E
G
∪ E
H
, ϕ
j
((T
Γ
G
+ T
Γ
H
) (ξς)) = ϕ
j
((T
Γ
G
∪ T
Γ
H
) (ξς)); ϕ
j
((I
Γ
G
+ I
Γ
H
) (ξς)) = ϕ
j
((I
Γ
G
∪ I
Γ
H
) (ξς)); ϕ
j
((F
Γ
G
+ F
Γ
H
) (ξς)) = ϕ
j
((F
Γ
G
∪ F
Γ
H
) (ξς)) ; E
a
=Collection of all edges adding the elements of V
G
and V
H
and assuming that V
G
∩ V
H
= ϕ and if ξς ∈ E
a
, ϕ
j
((T
Γ
G
+ T
Γ
H
) (ξ, ς)) = ϕ
j
(T
ϒ
G
(ξ) ∧ ϕ
j
(T
ϒ
H
(ς)); ϕ
j
((I
Γ
G
+ I
Γ
H
) (ξ, ς)) = ϕ
j
(I
ϒ
G
(ξ) ∧ ϕ
j
(I
ϒ
H
(ς)); ϕ
j
((F
Γ
G
+ F
Γ
H
) (ξ, ς)) = ϕ
j
(F
ϒ
G
(ξ) ∨ ϕ
j
(F
ϒ
H
(ς)); ϕ
j
((T
Γ
G
+ T
Γ
H
) (ξς)) =0, ϕ
j
((I
Γ
G
+ I
Γ
H
) (ξς)) =0, ϕ
j
((F
Γ
G
+ F
Γ
H
) (ξς)) =0.
Compute G + H. Then we have the following:
(T Γ G + T Γ H ) (x1x2) = <0.3, 0.2, 0.4 >,
(I Γ G + I Γ H ) (x1x2) = <0.4, 0.4, 0.3 >,
(F Γ G + F Γ H ) (x1x2) = <0.4, 0.3, 0.5 >.
Similarly, we get
(T ϒ G + T ϒ H ) (x1) = <0.6, 0.5, 0.4 >,
(I ϒ G + I ϒ H ) (x1) = <0.5, 0.4, 0.5 >,
(F ϒ G + F ϒ H ) (x1) = <0.4, 0.3, 0.6 >. See Fig. 3.

Join of two 3-polar SVNG’s.
∀ϑ ∈ V
G
, ϕ
j
(T
ϒ
G
(ϑ)) ≤ ϕ
j
(T
ϒ
H
(θ (ϑ))), (b) ϕ
j
(I
ϒ
G
(ϑ)) ≤ ϕ
j
(I
ϒ
H
(θ (ϑ))) and (c) ϕ
j
(F
ϒ
G
(ϑ)) ≤ ϕ
j
(F
ϒ
H
(θ (ϑ))), ϕ
j
(T
Γ
G
(ζη)) ≤ ϕ
j
(T
Γ
H
(θ (ζη))), (b) ϕ
j
(I
Γ
G
(ζη)) ≤ ϕ
j
(I
Γ
H
(θ (ζη))) and (c) ϕ
j
(F
Γ
G
(ζη)) ≤ ϕ
j
(F
Γ
H
(θ (ζη))).
∀ϑ ∈ V
G
, ϕ
j
(T
ϒ
G
(ϑ)) = ϕ
j
(T
ϒ
H
(θ (ϑ))), ϕ
j
(I
ϒ
G
(ϑ)) = ϕ
j
(I
ϒ
H
(θ (ϑ))) and (c) ϕ
j
(F
ϒ
H
(ϑ)) = ϕ
j
(F
ϒ
H
(θ (ϑ))), If G is isomorphic to H, then we can write G ≅ H.
θ is a homomorphism and ∀ϑ ∈ V
G
, ϕ
j
(T
ϒ
G
(ϑ)) = ϕ
j
(T
ϒ
H
(θ (ϑ))), ϕ
j
(I
ϒ
G
(ϑ)) = ϕ
j
(I
ϒ
H
(θ (ϑ))) and ϕ
j
(F
ϒ
G
(ϑ)) = ϕ
j
(F
ϒ
H
(θ (ϑ))).
A weak isomorphism allways preserve the weights of the vertices, but not necessarily the weights of the edges.
θ is a homomorphism, and
ϕ
j
(T
Γ
G
(αβ)) = ϕ
j
(T
ϒ
G
(α)) ∧ ϕ
j
(T
ϒ
G
(β)), ϕ
j
(I
Γ
G
(αβ)) = ϕ
j
(I
ϒ
G
(α)) ∧ ϕ
j
(I
ϒ
G
(β)), ϕ
j
(F
Γ
G
(αβ)) = ϕ
j
(F
ϒ
G
(α)) ∨ ϕ
j
(F
ϒ
G
(β)).
ϕ j (T Γ G (αβ)) < ϕ j (T ϒ G (α)) ∧ ϕ j (T ϒ G (β)) and
ϕ j (T Γ H (γδ)) < ϕ j (T ϒ H (γ)) ∧ ϕ j (T ϒ H (δ)).
Without loss of generality, we assume that
Let
Again, ϕ j ((T ϒ G × T ϒ H ) (ξ, γ)) = ϕ j (T ϒ G (ξ)) ∧ ϕ j (T ϒ H (γ)) and ϕ j ((T ϒ G × T ϒ H ) (ξ, δ)) = ϕ j (T ϒ G (ξ)) ∧ ϕ j (T ϒ H (δ))
∴ ϕ j ((T ϒ G × T ϒ H ) (ξ, γ)) ∧ ϕ j ((T ϒ G × T ϒ H ) (ξ, δ)) = ϕ j (T ϒ G (ξ)) ∧ ϕ j (T ϒ H (γ)) ∧ ϕ j (T ϒ H (δ))
∴ ϕ j ((T Γ G × T Γ H ) ((ξ, γ) (ξ, δ))) = ϕ j (T ϒ G (ξ)) ∧ ϕ j (T Γ H (γδ)) < ϕ j ((T ϒ G × T ϒ H ) (ξ, γ)) ∧ ϕ j ((T ϒ G × T ϒ H ) (ξ, δ)).
This implies that G × H is not a strong m-polar SVNG, but G × H is strong SVNG. By contrapositively, at least one of G or H must be a strong SVNG.
This completes the proof. □
Complement of m-polar SVNG
Then,
Let If for j = 1, 2, . . . , m, ϕ
j
T
Γ
G
(αβ)) =0, Let If for j = 1, 2, . . . , m, ϕ
j
(I
Γ
G
(αβ)) =0, Let If for j = 1, 2, . . . , m, ϕ
j
(F
Γ
G
(αβ)) =0,
This completes the proof. □
or, ϕ j (T Γ G (αβ)) = ϕ j (T ϒ G (θ (α))) ∧ ϕ j (T ϒ G (θ (β))) - ϕ j (T Γ G (θ (α) θ (β))),
(By taking ∑α≠β on both sides)
or, ∑α≠βϕ j (T Γ G (αβ)) + ∑α≠βϕ j (T Γ G (θ (α) θ (β))) = ∑α≠βϕ j (T ϒ G (θ (α))) ∧ ϕ j (T ϒ G (θ (β))),
or, 2∑α≠βϕ j (T Γ G (αβ)) = ∑α≠βϕ j (T ϒ G (α)) ∧ ϕ j (T ϒ G (β)),
or,
(ii) Let
or,
(By taking ∑α≠β on both sides)
or, ∑α≠βϕ j (I Γ G (αβ)) + ∑α≠βϕ j (I Γ G (θ (α) θ (β))) = ∑α≠βϕ j (I ϒ G (θ (α))) ∧ ϕ j (I ϒ G (θ (β)))
or, 2∑α≠βϕ j (I Γ G (αβ)) = ∑α≠βϕ j (I ϒ G (α)) ∧ ϕ j (I ϒ G (β))
or,
(iii) Let
or,
or, ∑α≠βϕ j (F Γ G (αβ)) + ∑α≠βϕ j (F Γ G (θ (α) θ (β))) = ∑α≠βϕ j (F ϒ G (θ (α))) ∨ ϕ j (F ϒ G (θ (β)))
or, 2∑α≠βϕ j (F Γ G (αβ)) = ∑α≠βϕ j (F ϒ G (α)) ∨ ϕ j (F ϒ G (β))
or,
This completes the proof. □
ϕ
j
(T
ϒ
G
(ξ)) = ϕ
j
(T
ϒ
H
(θ (ξ))), ∀ξ ∈ V
G
and ϕ
j
(T
Γ
G
(ξη)) = ϕ
j
(T
Γ
H
(θ (ξ) θ (η)), ϕ
j
(I
ϒ
G
(ξ)) = ϕ
j
(I
ϒ
H
(θ (ξ))), ∀ξ ∈ V
G
and ϕ
j
(I
Γ
G
(ξη)) = ϕ
j
(I
Γ
H
(θ (ξ) , θ (η)), ϕ
j
(F
ϒ
G
(ξ)) = ϕ
j
(F
ϒ
H
(θ (ξ))), ∀ξ ∈ V
G
and ϕ
j
(F
Γ
G
(ξη)) = ϕ
j
(F
Γ
H
(θ (ξ) θ (η)), Let 0 < ϕ
j
(T
Γ
H
(θ (α) θ (β)) ≤1. ∴ Let 0 < ϕ
j
(I
Γ
H
(θ (α) θ (β)) ≤1. ∴ Let 0 < ϕ
j
(F
Γ
H
(θ (α) θ (β)) ≤1. ∴
So
Conversely, let
Let or, ϕ
j
(T
ϒ
H
(ψ (α))∧ ϕ
j
(T
ϒ
H
(ψ (β)) = ϕ
j
(T
ϒ
H
(ψ (α))) ∧ϕ
j
(T
ϒ
H
(ψ (β))) - ϕ
j
(T
Γ
H
(ψ (α) ψ (β))) , or, ϕ
j
(T
Γ
H
(ψ (α) ψ (β))) =0. ∴ ϕ
j
(T
Γ
H
(ψ (α) ψ (β))) =0 = ϕ
j
(T
Γ
G
(αβ)), for each j = 1, 2, . . . , m. If for each j = 1, 2, . . . m, 0 < ϕ
j
(T
Γ
G
(αβ)) ≤1, then So, we have,
Let or, ϕ
j
(I
ϒ
H
(ψ (α))∧ ϕ
j
(I
ϒ
H
(ψ (β)) = ϕ
j
(I
ϒ
H
(ψ (α))) ∧ ϕ
j
(I
ϒ
H
(ψ (β))) - ϕ
j
(I
Γ
H
(ψ (α) ψ (β))) , or, ϕ
j
(T
Γ
H
(ψ (α) ψ (β))) =0. ∴ ϕ
j
(I
Γ
H
(ψ (α) ψ (β))) =0 = ϕ
j
(I
Γ
G
(αβ)), for each j = 1, 2, . . . , m. If for each j = 1, 2, . . . m, 0 < ϕ
j
(I
Γ
G
(αβ)) ≤1, then So, we have
Let or, ϕ
j
(F
ϒ
H
(ψ (α))∨ ϕ
j
(F
ϒ
H
(ψ (β)) = ϕ
j
(F
ϒ
H
(ψ (α))) ∨ϕ
j
(F
ϒ
H
(ψ (β))) - ϕ
j
(F
Γ
H
(ψ (α) ψ (β))) , or, ϕ
j
(F
Γ
H
(ψ (α) ψ (β))) =0. ∴ ϕ
j
(F
Γ
H
(ψ (α) ψ (β))) =0 = ϕ
j
(F
Γ
G
(αβ)), for each j = 1, 2, . . . , m. If for each j = 1, 2, . . . , m, 0 < ϕ
j
(F
Γ
G
(αβ)) ≤1, then So, we have
∴ G ≅ H.
This completes the proof. □
Different type of product of m-polar SVNG’s
∀ (ξ, η) ∈ V
G
× V
H
and for each j = 1, 2, . . . , m, ϕ
j
((T
ϒ
G
⊓ T
ϒ
H
) (ξ, η)) = ϕ
j
(T
ϒ
G
(ξ)) ∧ ϕ
j
(T
ϒ
H
(η)), ϕ
j
((I
ϒ
G
⊓ I
ϒ
H
) (ξ, η)) = ϕ
j
(I
ϒ
G
(ξ)) ∧ ϕ
j
(I
ϒ
H
(η)), ϕ
j
((F
ϒ
G
⊓ F
ϒ
G
) (ξ, η)) = ϕ
j
(F
ϒ
G
(ξ)) ∨ ϕ
j
(F
ϒ
H
(η)), ∀ ((α, γ) (β, δ) ∈ E2 and for each j = 1, 2, . . . , m, ϕ
j
((T
Γ
G
⊓ T
Γ
H
) ((α, γ) (β, δ)) = ϕ
j
(T
Γ
G
(αβ)) ∧ ϕ
j
(T
Γ
H
(γδ)), ϕ
j
((I
Γ
G
⊓ I
Γ
H
) ((α, γ) (β, δ)) = ϕ
j
(I
Γ
G
(αβ)) ∧ ϕ
j
(I
Γ
H
(γδ)), ϕ
j
((F
Γ
G
⊓ F
Γ
H
) ((α, γ) (β, δ)) = ϕ
j
(F
Γ
G
(αβ)) ∨ ϕ
j
(F
Γ
H
(γδ)).
ϕ
j
(T
Γ
G
(αβ)) = ϕ
j
(T
ϒ
G
(α)) ∧ ϕ
j
(T
ϒ
G
(β)) and ϕ
j
(T
Γ
H
(γδ)) = ϕ
j
(T
ϒ
H
(γ)) ∧ ϕ
j
(T
ϒ
H
(δ)) , ϕ
j
(I
Γ
G
(αβ)) = ϕ
j
(I
ϒ
G
(α)) ∧ ϕ
j
(I
ϒ
G
(β)) and ϕ
j
(I
Γ
H
(γδ)) = ϕ
j
(I
ϒ
H
(γ)) ∧ ϕ
j
(I
ϒ
H
(δ)), ϕ
j
(F
Γ
G
(αβ)) = ϕ
j
(F
ϒ
G
(α)) ∨ ϕ
j
(F
ϒ
G
(β)) and ϕ
j
(F
Γ
H
(γδ)) = ϕ
j
(F
ϒ
H
(γ)) ∨ ϕ
j
(F
ϒ
H
(δ)).
Now, ∀ ((ξ, η) (ς, υ)) ∈ E2 and for each j = 1, 2, . . . , m,
Hence G ⊓ H is a strong m-polar SVNG.
This completes the proof.□
∀ (ξ, η) ∈ V
G
× V
H
and for each j = 1, 2, . . . , m, ϕ
j
((T
ϒ
G
• T
ϒ
H
) (ξ, η)) = ϕ
j
(T
ϒ
G
(ξ)) ∧ ϕ
j
(T
ϒ
H
(η)), ϕ
j
((I
ϒ
G
• I
ϒ
H
) (ξ, η)) = ϕ
j
(I
ϒ
G
(ξ)) ∧ ϕ
j
(I
ϒ
H
(η)), ϕ
j
((F
ϒ
G
• F
ϒ
H
) (ξ, η)) = ϕ
j
(F
ϒ
G
(ξ)) ∨ ϕ
j
(F
ϒ
H
(η)), ∀ ((α, γ) (β, δ) ∈ E2, ∀ ((ξ, η) (ξ, υ)) ∈ (E3 - E2) and for each j = 1, 2, . . . , m, ϕ
j
((T
Γ
G
• T
Γ
H
) ((ξ, η) (ξ, υ)) = ϕ
j
(T
ϒ
G
(ξ)) ∧ ϕ
j
(T
Γ
H
(ηυ)) and ϕ
j
((T
Γ
G
• T
Γ
H
) ((α, γ) (β, δ)) = ϕ
j
(T
Γ
G
(αβ)) ∧ ϕ
j
(T
Γ
H
(γδ)), ϕ
j
((I
Γ
G
• I
Γ
H
) ((ξ, η) (ξ, υ)) = ϕ
j
(I
ϒ
G
(ξ)) ∧ ϕ
j
(I
Γ
H
(ηυ)) and ϕ
j
((I
Γ
G
• I
Γ
H
) ((α, γ) (β, δ)) = ϕ
j
(I
Γ
G
(αβ)) ∧ ϕ
j
(I
Γ
H
(γδ)), ϕ
j
((F
Γ
G
• F
Γ
H
) ((ξ, η) (ξ, υ)) = ϕ
j
(F
ϒ
G
(ξ)) ∨ ϕ
j
(F
Γ
H
(ηυ)) and ϕ
j
((F
Γ
G
• F
Γ
H
) ((α, γ) (β, δ)) = ϕ
j
(F
Γ
G
(αβ)) ∨ ϕ
j
(F
Γ
H
(γδ)).
ϕ
j
(T
Γ
G
(αβ)) = ϕ
j
(T
ϒ
G
(α)) ∧ ϕ
j
(T
ϒ
G
(β)) and ϕ
j
(T
Γ
H
(γδ)) = ϕ
j
(T
ϒ
H
(γ)) ∧ ϕ
j
(T
ϒ
H
(δ)), ϕ
j
(I
Γ
G
(αβ)) = ϕ
j
(I
ϒ
G
(α)) ∧ ϕ
j
(I
ϒ
H
(β)) and ϕ
j
(I
Γ
H
(γδ)) = ϕ
j
(I
ϒ
G
(γ)) ∧ ϕ
j
(I
ϒ
H
(δ)), ϕ
j
(F
Γ
G
(αβ)) = ϕ
j
(F
ϒ
G
(α)) ∨ ϕ
j
(F
ϒ
G
(β)) and ϕ
j
(F
Γ
H
(γδ)) = ϕ
j
(F
ϒ
H
(γ)) ∨ ϕ
j
(F
ϒ
H
(δ)).
Now, ∀ ((ξ, η) (ς, υ)) ∈ E2, ∀ ((α, β) (α, γ)) ∈ E3 - E2 and for each j = 1, 2, . . . , m
Hence G • H is a semi strong m-polar SVNG.
This completes the proof. □
∀ (ξ, η) ∈ V
G
× V
H
and for each j = 1, 2, . . . , m, ϕ
j
((T
ϒ
G
⨂ T
ϒ
H
) (ξ, η)) = ϕ
j
(T
ϒ
G
(ξ)) ∧ ϕ
j
(T
ϒ
H
(η)) ϕ
j
((I
ϒ
G
⨂ I
ϒ
H
) (ξ, η)) = ϕ
j
(I
ϒ
G
(ξ)) ∧ ϕ
j
(I
ϒ
H
(η)) ϕ
j
((F
ϒ
G
⨂ F
ϒ
H
) (ξ, η)) = ϕ
j
(F
ϒ
G
(ξ)) ∨ ϕ
j
(F
ϒ
H
(η)). ∀ ((α, γ) (β, δ) ∈ E2, ∀ ((ξ, η) (ξ, υ)) ∈ E, ((α, β) (γ, β)) ∈ E and for each j = 1, 2, . . . , m, ϕ
j
((T
Γ
G
⨂ T
Γ
H
) ((ξ, η) (ξ, υ))) = ϕ
j
(T
ϒ
G
(ξ)) ∧ ϕ
j
(T
Γ
H
(ηυ)), ϕ
j
((T
Γ
G
⨂ T
Γ
H
) ((α, β) (γ, β))) = ϕ
j
(T
Γ
G
(αγ)) ∧ ϕ
j
(T
ϒ
H
(β)) and ϕ
j
((T
Γ
G
⨂ T
Γ
H
) ((α, γ) (β, δ))) = ϕ
j
(T
Γ
G
(αβ)) ∧ ϕ
j
(T
Γ
H
(γδ)), ϕ
j
((I
Γ
G
⨂ I
Γ
H
) ((ξ, η) (ξ, υ))) = ϕ
j
(I
ϒ
G
(ξ)) ∧ ϕ
j
(I
Γ
H
(ηυ)), ϕ
j
((I
Γ
G
⨂ I
Γ
H
) ((α, β) (γ, β))) = ϕ
j
(I
Γ
G
(αγ)) ∧ ϕ
j
(I
ϒ
H
(β)) and ϕ
j
((I
Γ
G
⨂ I
Γ
H
) ((α, γ) (β, δ))) = ϕ
j
(I
Γ
G
(αβ)) ∧ ϕ
j
(I
Γ
H
(γδ)), ϕ
j
((F
Γ
G
⨂ F
Γ
H
) ((ξ, η) (ξ, υ))) = ϕ
j
(F
ϒ
G
(ξ)) ∨ ϕ
j
(F
Γ
H
(ηυ)), ϕ
j
((F
Γ
G
⨂ F
Γ
H
) ((α, β) (γ, β))) = ϕ
j
(F
Γ
G
(αγ)) ∨ ϕ
j
(F
ϒ
H
(β)) and ϕ
j
((F
Γ
G
⨂ F
Γ
H
) ((α, γ) (β, δ))) = ϕ
j
(F
Γ
G
(αβ)) ∨ ϕ
j
(F
Γ
H
(γδ)).
Application
The 1-polar SVNG is a special type of SVNG. The SVNG is used for modeling several real life applications, e.g., telephone network planning [12], safest path finding in an airline travel[5], multi-criteria decision making, choosing of location and gas terminal, Baltic sea [9], cellular network provider Companies [26], and so on. The bipolar SVNG (2-polar SVNG) is an example of m-polar SVNG where the value of m = 2. Bipolar SVNG has used in many real life applications such as organizational designations, radio channels [2], decision making problems [6], and brands competition [1] etc. Additionally, m-polar SVNG’s (m > 2) are very efficient to handle the uncertainties in many real life situations, e.g., when a country like India elects its prime minister or If a set of persons decides which film to watch. For e.g., we can consider a real life application of a mobile company. In a mobile company, a group of managing director decides which mobile to manufacture. This type of real life application can be modeled by m-polar SVNG. Here, different mobile can be considered as vertices and two nodes are connected by an arc if there exists any relationship between the two vertices.
The membership degree of each vertex/mobile describes the preference of the mobile which collects from a group of mobile users. The preference degree of a mobile (within [0, 1]) is expressed by a single user. Thus, each vertex has multi-preference degrees corresponding to a mobile. Similarly, the relationship degree between two nodes determines the relationship value of the arc. Between two mobiles, one mobile may have a better camera, may be in very high RAM and processor, may be battery, and so on. In such real life scenarios, there exists multipolar information between two mobiles. This network is a classical case of m-polar SVNG’s. For any mobile company, it is very significant to decide which mobile to produce so that the company can make the profit and band as much as possible. A banded and good mobile is readily purchased by several customers if the price of the mobile is also very much inexpensive. The decision of which mobile to produce is a decision making problem. By making the proper decision, a mobile company can sell their mobile throughout the world. The company can keep in their mind that the mobile is excellent quality, in very demand, easily accessible, and so on. Therefore, we can use m-polar SVNG’s to model many real life application in many field of engineering and science, electronic engineering, computer science, mathematics, artificial intelligence and artificial neural networks.
Here, we introduce a model of m-polar SVNG graph G = (V G , ϒ G , Γ G ) to evaluate the performance of professors by the feedback of the students of the computer science department in a college during the session 2018-2019. Here, the vertices are used to represent the professors of the computer science department and arcs are used to represent the relationship among the two professors. Suppose the computer science department has six professors represented as V = {t1, t2, t3, t4, t5}. The value of membership of each vertex presents the corresponding professors feedback response of the students based regularity, updated information, presentation style, generation of interest, lecture quality, and promoting future study among students. All the features of a professor according to the student is uncertain, therefore 4-polar single-valued neutrosophic set is used to represent the node set V.
In the Table 1, the truth-membership values, indeterminatistic membership values of the professor’s are described which is based on the rating of the students.
4-Polar SVNS A of V
G
4-Polar SVNS A of V G
Arc cost is used to represent the relationship among the professor’s which can be computed as follows. ϕ
j
(T
Γ
G
(xy))≤ ϕ
j
(T
ϒ
G
(x)) ∧ ϕ
j
(T
ϒ
G
(y)) ; ϕ
j
(I
Γ
G
(xy))≤ ϕ
j
(I
ϒ
G
(x)) ∧ ϕ
j
(I
ϒ
G
(y)) ; ϕ
j
(F
Γ
G
(xy)) ≥ ϕ
j
(F
ϒ
G
(x)) ∨ ϕ
j
(F
ϒ
G
(y)),
The membership degrees are shown in the Table 2.
4-Polar SVNR on V G
Score Value of The Teacher’s
The professor’s performance are ranked according to the following characteristics: Professor’s performance value is <60%, professor’s performance is Professor’s performance value is ≥60% and <65%, professor’s performance is Professor’s performance value is ≥65% and <70%, Professor’s performance value is Professor’s performance value is ≥70%, Professor’s performance value is
In the Table 1, the performance of t2, t3, t5 are very good and the performance of the professor t4 is excellent. Among all the professors, professor t4 is the best professor according the response value of the students of the computer science department. We compute the score value of each professor according to the formula
where T =average of truth membership,
I =average of indeterminastic membership and
F = average of truth of falsity membership.
Neutrosophic graph is an efficient method for modeling many uncertain decision optimization problems. Multi-polar information generally exists in almost every decision making making. This type of information cannot be properly modeled using simple neutrosophic graph. We can use the m- polar neutrosophic set/graph to deal this type of uncertainty. The main contribution of this study is to define the m- polar neutrosophic graph model and also present several operations, such as union, join, composition and ring sum on m-polar neutrosophic graph. We describe some new products (strong product, semi strong product, complete product, direct product, Cartesian product and lexicographic product) for m-polar neutrosophic graph. Finally, an application of m-polar neutrosophic graph is describe the evaluation of the teacher’s performance of a college.
In our future work, we will focus on transportation networks, production management, artificial neural networks and database theory of m-polar neutrosophic graph which are very much efficient in many branches of engineering, science and medicine and we will also try is to extend this work to m-polar fuzzy interval graphs, m-polar fuzzy intersection graphs and m-polar fuzzy hypergraphs and so on.
