CAAI Transactions on Intelligence Technology

Research Article

ISSN 2468-2322
Study on single-valued neutrosophic graph Received on 3rd June 2020
Revised on 25th June 2020
with application in shortest path problem Accepted on 14th July 2020
doi: 10.1049/trit.2020.0111
www.ietdl.org

Ruxiang Liu ✉
Department of Electronic and Information Engineering, Bozhou University, Bozhou, Anhui 236800, People’s Republic of China
✉ E-mail: [email protected]

Abstract: Fuzzy set and neutrosophic set are two efficient tools to handle the uncertainties and vagueness of any real-
world problems. Neutrosophic set is more useful than fuzzy set (intuitionistic fuzzy sets) to manage the uncertainties of
a real-life problem. This study introduces some new concepts of single-valued neutrosophic graph (SVNG). The
authors have discussed the definition of regular SVNG, complete SVNG and strong SVNG. The shortest path problem
is a well-known combinatorial optimisation problem in the field of graph theory due to its various applications.
Uncertainty is present in almost every application of shortest path problem which makes it very hard to decide the
edge weight properly. The main objective behind the work in this study is to determine an algorithmic technique for
shortest path problem which will be very easy and efficient for use in real-life scenarios. In this study, the authors
consider neutrosophic number to describe the edge weights of a neutrosophic graph for neutrosophic shortest path
problem. An algorithm is introduced to solve this problem. The uncertainties are incorporated in Bellman–Ford
algorithm for shortest path problem using neutrosophic number as arc length. They use one numerical example to
illustrate the effectiveness of the proposed algorithm.

1 Introduction graphs and presented many theorems in [16]. Parvathi et al.

[17–19] proposed some operations between two intuitionistic fuzzy
In 1965, Prof. Zadeh [1] described the idea of fuzzy set theory which graphs. In [20], Rashmanlou et al. proposed many products
has been expeditiously utilised to model the several decision-making operations such as lexicographic, direct product, strong product,
problems in which uncertainties may exist. Fuzzy set is a modified semi-strong product on intuitionistic fuzzy graphs. They have
version of simple set, where all the elements of the fuzzy set described the Cartesian production, join, composition and union
have changing degrees of membership values. The simple set on intuitionistic fuzzy graphs in their paper. For further study on
(crisp set) always have two truth values, either 0 (indicate false) or intuitionistic fuzzy graphs, please refer to [21–27]. Akram et al.
1 (indicate true). Crisp set is unable to handle the uncertainties of [28–32] have introduced the idea of pythagorean fuzzy graph.
the problems. However, the fuzzy set allows for its objects to have They have described the several applications of pythagorean fuzzy
the membership degree within 1 and 0 which provides more graph in their paper. Neutrosophic graph [33] is used to model
beneficial results, rather than considering only single value of many real-world problems which consist of inconsistent
either 1 or 0. The membership degree of a fuzzy set is a specific information. Recently, many scientists have researched on graph in
single value within 0 and 1. Experts are not able to handle with neutrosophic environment [34–41], for instance, Yang et al. [9],
the uncertainty of any decision-making problem properly using Arkam [12, 14], Ye [8], Naz et al. [10], Das and Edalatpanah
fuzzy set. Atanassov [2] has introduced the idea of intuitionistic [42], Dey et al. [43–47] and Broumi [48–51]. In 2020, Prof.
fuzzy set (IFS) [3, 4] by including a non-membership grade and a Smarandache introduced the idea of n-super hyper-graph [52] with
hesitancy grade of all the elements of the fuzzy set. IFS is present super-nodes and hyper-arcs for neutrosophic graph.
to describe the elements/objects of the fuzzy set from three The shortest path problem (SPP) is a well-known network optimisa-
different aspects of inferiority, superiority and hesitation, which tion problem in the area of operation research. In this problem, deci-
are generally modelled by the intuitionistic fuzzy numbers (IFNs). sion maker focuses on determining a shortest path between a
To handle more useful information of real-life problem under specified starting node and other nodes. The SPP has been considered
imprecise, vague and uncertain environment, Smarandache [5–7] to model in many real-life problems, e.g. economics, telecommunica-
has presented the novel idea of neutrosophic set, by generalising tions, transportation, scheduling, routing and supply chain manage-
the idea of IFS. The neutrosophic set can be used to capture the ment. Many researchers have studied intensively on the SPPs with
uncertainties due to inconsistent, vagueness and indeterminate data deterministic edge costs. These SPPs are referred to as standard
of any problem. It is nothing but an extended edition of simple SPPs. Decision maker can solve the standard SPPs efficiently using
classical set, fuzzy set and IFS. In neutrosophic set, each object several well-known algorithms introduced by some excellent research-
has three different types of membership grade: truth, false and ers. Although in standard SPP, the costs of the arcs are considered real
indeterminate. Those three membership grades of neutrosophic set numbers ([ R), most real-life scenarios, however, have many para-
are not dependent on each other and always within ]0, 1[. meters that may not be always precise (i.e. travelling demands, travel-
Graph is an efficient tool to model the real-life problems. ling costs, travelling capacities, travelling time etc.). Several types of
By modelling the graph, the objects and their relations are uncertainty are generally encountered in practical applications of SPP
symbolised by nodes and arcs. There exists many different types due to imperfect data, maintenance, failure or other reasons. In such
of information in real-life problems and we need several types of scenarios, the arc costs are non-deterministic in nature. Some research-
graphs to model those problems such as fuzzy graph, intuitionistic ers [53] use type-1 fuzzy numbers for handling the uncertainties in
fuzzy graphs and neutrosophic graph theory [8–15]. Shannon and standard SPP and this type of SPP is defined as fuzzy SPP (FSPP).
Atanassov [16] presented the concept of relationship between IFS. The FSPP cannot manage the several types of uncertainties because
Then they have introduced the concept of intuitionistic fuzzy the membership degree of type 1 fuzzy numbers is simply real

CAAI Trans. Intell. Technol., 2020, Vol. 5, Iss. 4, pp. 308–313

308 This is an open access article published by the IET, Chinese Association for Artificial Intelligence and
Chongqing University of Technology under the Creative Commons Attribution-NonCommercial-NoDerivs License (http://
creativecommons.org/licenses/by-nc-nd/3.0/)


number. To solve this problem, few researchers worked on intuitionis- fuzzy set on j and D = TD , ID , FD is a picture fuzzy set on E #
tic FSPP. In intuitionistic FSPP, the arc lengths are considered as V × V such that for each arc ij [ E
IFNs. It can work with uncertain information about the arc length
which consists of membership grade and non-membership grade sim-



TD i, j ≤ min TC (i), TC j ,
ultaneously. The main disadvantage of intuitionistic FSPP is that it



cannot handle the uncertain information of arc length if the sum of ID i, j ≤ min IC (i), IC j , (5)
non-membership and membership is bigger than 1. In such real-life



scenarios, an appropriate modelling technique may justifiably FD i, j ≥ max FC (i), FC j
employ the neutrosophic set, and so does the name neutrosophic
SPP [48–51] appears in the area of graph theory. The neutrosophic
SPP, involving addition operation and comparison operation of neu- Definition 5: A neutrosophic graph G = (C, D) is said to be regular
trosophic set, is different from the standard SPP (FSPP), which only neutrosophic graph if
uses crisp numbers (fuzzy number). In a neutrosophic SPP, the arc
length being neutrosophic numbers, the main objective of determining 

TD i, j = constant,
a path between two nodes being smaller than all the other paths is not i
easy, as the ranking of neutrosophic set as a comparison operation can i=j
be described in several ways. The Bellman–Ford algorithm is a 

common and efficient algorithm to solve the standard SPP. The clas- ID i, j = constant,
i (6)
sical Bellman–Ford algorithm is easy to implement for standard SPP. i=j
In this manuscript, an extended Bellman–Ford algorithm is designed 

to solve the neutrosophic SPP. In this algorithm, we need to address FD i, j = constant, ∀i, j [ E
two key issues to find the solution of SPP with neutrosophic para- i
i=j
meters. The first issue is how to find the summing operation of two
edges, i.e. neutrosophic numbers. It is needed to calculate the path
length. The second one is that how to compare the two different neu- Definition 6: A neutrosophic graph G = (C, D) is defined as strong
trosophic paths with their arc costs described by neutrosophic neutrosophic graph if
numbers. To solve these problems, the ranking method of neutro-
sophic set is adopted to extend the classical Bellman–Ford algorithm.


This research paper introduces some new concepts of TD i, j = TC (i) ^ TC j ,
single-valued neutrosophic graph (SVNG). We have discussed the


ID i, j = IC (i) ^ IC j , (7)
definition of regular SVNG, complete SVNG and strong SVNG.


In this manuscript, we consider neutrosophic number to describe FD i, j = FC (i) _ FC j ∀i, j [ E
the edge weights of a neutrosophic graph for neutrosophic SPP.
An algorithm is introduced to solve this problem. The
uncertainties are incorporated in Bellman–Ford algorithm for SPP Definition 7: A neutrosophic graph G = (C, D) is defined as
using neutrosophic number as arc length. We use one numerical complete neutrosophic graph if
example to illustrate the effectiveness of the proposed algorithm.



TD i, j = TC (i) ^ TC j ,



2 Preliminary ID i, j = IC (i) ^ IC j , (8)



In this section, we define neutrosophic graph and introduce different FD i, j = FC (i) _ FC j ∀i, j [ V
types of regular neutrosophic graph, strong neutrosophic graph,
complete neutrosophic graph and complement neutrosophic graph.
Definition 8: A path p in a neutrosophic graph G = (C, D) is a
Definition 1: Let U be a classical universal set. A neutrosophic set sequence of different vertices p0 , p1 , p2 , . . . , pk such that
[54] D on the U is described by three independent membership





functions: true membership function TD (x), indeterminate TD pi−1 pi ID pi−1 pi FD pi−1 pi . 0, i = 1, 2, . . . , k. (9)
membership function ID (x) and false membership function FD (x)
−
0 ≤ sup TD (x) + sup ID (x) + sup FD (x) ≤ 3+ (1) Here, k represents the length of path.

Definition 9: Let G = (C, D) be a neutrosophic graph. Then, G is said

Definition 2: Let U be a universal set. The single-valued to be connected neutrosophic graph if for every vertices i, j [ V ,
neutrosophic set [55] D on the universal U is denoted as follows: TB1 (i, j) . 0 or IB1 (i, j) . 0 or FB1 (i, j) , 1.
A = {kx:TD (x), ID (x), FD (x)|x [ U l} (2)
Definition 10: The complement
of a neutrosophic
 graph G = (C, D)
The functions TD (x) [ [0, 1], IA (x) [ [0, 1] and FA (x) [ [0, 1] are is a neutrosophic graph G′ = C′ , D′ if and only if follows the
named as degree of truth, indeterminacy and falsity membership of following equation:
x in A, satisfy the following condition:
(T ′C , I ′C , F ′C ) = (TC , IC , FC )
−
0 ≤ sup TD (x) + sup ID (x) + sup FD (x) ≤ 3 +



(3) T ′D i, j = TC (i) ^ TC j − TB i, j ,



 (10)
I ′D i, j = IC (i) ^ IC j − IB i, j ,




Definition 3: Let A = (TD , ID , FD ) be a single-valued neutrosophic F ′D i, j = FC (i) _ FC j − FB i, j ∀i, j [ V
set. A score function S [34] is defined as follows:
(1 + (TD − 2ID − FD )(2 − TD − FD )) Definition 11: A neutrosophic graph G is said to be
S(D) = (4)
2
(i) The G is self-complementary neutrosophic graph then G = G′ .
∗ (ii) The G is self-weak complement neutrosophic graph then G is
Definition 4: : Let G = (V, E ) be a simple graph.

 G = (C, D)
A pair
is a neutrosophic graph on G∗ where C = TC , IC , FC is a picture weak isomorphic to G′ .

CAAI Trans. Intell. Technol., 2020, Vol. 5, Iss. 4, pp. 308–313

This is an open access article published by the IET, Chinese Association for Artificial Intelligence and 309
Chongqing University of Technology under the Creative Commons Attribution-NonCommercial-NoDerivs License (http://
creativecommons.org/licenses/by-nc-nd/3.0/)
3 Operations on neutrosophic graph (i)
(a) TC (i) = TC1 (i) if i [ V 1 and i  V 2
In this section, we introduce six operations on neutrosophic graph, (b) TC (i) = TC2 (i) if i [ V 2 and i  V 1
viz., Cartesian product, composition, join, direct product, (c) TC (i) = TC1 (i) ^ TC2 (i) if i [ V 1 > V 2
lexicographic and strong product.



Definition 12: Let G1 = C1 , D1
and G2 = C2 , D2
are two  (ii)
neutrosophic graphs of G∗1 = V 1 , E 1 and G∗2 = V 2 , E 2 ,
respectively. The Cartesian product G1 × G2 of (a) IC (i) = IC1 (i) if i [ V 1 and i  V 2

neutrosophic
 (b) IC (i) = IC2 (i) if i [ V 2 and i  V 1
graph
G1 and G2 is defined by (C, D), where C = TC , I
C , FC and 
D = TD , ID
, FD
are two neutrosophic sets on (c) IC (i) = IC1 (i) ^ IC2 (i) if i [ V 1 > V 2

V =
V 1 × V 2 ,
and E = { i, i2 , i, j2 |i [ V 1 , i2 j2 [ E 2 }< { i1 , k , j1 , k |k [
V 2 , i1 j1 [ E 1 }, respectively, which satisfies the following:
(iii)
(a) FC (i) = FC1 (i) if i [ V 1 and i  V 2
(i) ∀(i1 , i2 ) [ V 1 × V 2 , (b) FC (i) = FC2 (i) if i [ V 2 and i  V 1
(c) FC (i) = FC1 (i) ^ IC2 (i) if i [ V 1 > V 2
(a) TC ((i1 , i2 )) = TC1 (i1 ) ^ TC2 (i2 )
(b) IC ((i1 , i2 )) = IC1 (i1 ) ^ IC2 (i2 )
(c) FC ((i1 , i2 )) = FC1 (i1 ) _ FC2 (i2 )
(iv)
(ii) ∀i [ V 1 and ∀(i2 , j2 ) [ E 2 , (a) TD (ij) = TD1 (ij) if ij [ E 1 and ij  E 2
(b) TD (ij) = TD2 (ij) if ij [ E 2 and ij  E 1
(a) TD ((i, i2 )(i, j2 )) = TC1 (i1 ) ^ TD2 (i2 j2 ) (c) TD (ij) = TD1 (ij) ^ TD2 (ij) if ij [ E 1 > E 2
(b) ID ((i, i2 )(i, j2 )) = IC1 (i1 ) ^ ID2 (i2 j2 )
(c) FD ((i, i2 )(i, j2 )) = FC1 (i1 ) _ FD2 (i2 j2 )
(v)
(iii) ∀k [ V 2 and ∀(i1 , i2 ) [ E 1 ,

(a) TD ((i1 , k)(j1 , k)) = TD1 (i1 j1 ) ^ TC2 (k) (a) ID (ij) = ID1 (ij) if ij [ E 1 and ij  E 2
(b) ID ((i1 , k)(j1 , k) = ID1 (i1 j1 ) ^ IC2 (k) (b) ID (ij) = ID2 (ij) if ij [ E 2 and ij  E 1
(c) FD ((i1 , k)(j1 , k)) = FD1 (i1 j1 ) _ FC2 (k) (c) ID (ij) = ID1 (ij) ^ ID2 (ij) if ij [ E 1 > E 2

Definition

13: The composition

 1 G2 of two neutrosophic graphs
G (vi)
G1 =
C1 , D1 and G2 = C
2 , D2 defined  as a pair (C, D), where
C = TC ,
IC , FC and  D = TD , ID
, FD
are two neutrosophic sets
(a) FD (ij) = FD1 (ij) if ij [ E 1 and ij  E 2
on
V =
V 1 ×  V 2 , and E = { i, i2
, i, j2
|i [ V  1 , i2 j2 [ E 2 }<
{ i1 , k , j1 , k |k [ V 2 , i1 j1 [ E 1 }<{ i1 , i2 , j1 , j2 |i2 j2 [ V 2 = j2 , (b) FD (ij) = FD2 (ij) if ij [ E 2 and ij  E 1
i1 j1 [ E 1 }, respectively, which satisfies the following: (c) FD (ij) = FD1 (ij) ^ FD2 (ij) if ij [ E 1 > E 2

(i) ∀(i1 , i2 ) [ V 1 × V 2 ,
Definition

15: The joining G
1 + G2  of two neutrosophic graphs
(a) TC ((i1 , i2 )) = TC1 (i1 ) ^ TC2 (i2 ) G1 = C1 , D
1 and  G2 = C2 , D2 is defined as (C, D),
(b) IC ((i1 , i2 )) = IC1 (i1 ) ^ IC2 (i2 ) where
C = TC , IC , FC is a neutrosophic set on V = V 1 < V 2 and
(c) FC ((i1 , i2 )) = FC1 (i1 ) _ FC2 (i2 )
D = TD , ID , FD is an another neutrosophic set on E =
(ii) ∀i [ V 1 and ∀(i2 , j2 ) [ E 2 , E 1 < E 2 < E ′ ((E ′ represents all edges joining the vertex of V 1
and V 2 ), which satisfies the following:
(a) TD ((i, i2 )(i, j2 )) = TC1 (i1 ) ^ TD2 (i2 j2 )
(b) ID ((i, i2 )(i, j2 )) = IC1 (i1 ) ^ ID2 (i2 j2 ) (i)
(c) FD ((i, i2 )(i, j2 )) = FC1 (i1 ) _ FD2 (i2 j2 )

(iii) ∀k [ V 2 and ∀(i1 , i2 ) [ E 1 , (a) TC (i) = TC1 (i) if i [ V 1 and i  V 2

(b) TC (i) = TC2 (i) if i [ V 2 and i  V 1
(a) TD ((i1 , k)(j1 , k)) = TD1 (i1 j1 ) ^ TC2 (k) (c) TC (i) = TC1 (i) ^ TC2 (i) if i [ V 1 > V 2
(b) ID ((i1 , k)(j1 , k) = ID1 (i1 j1 ) ^ IC2 (k)
(c) FD ((i1 , k)(j1 , k)) = FD1 (i1 j1 ) _ FC2 (k)
(ii)
(iv) ∀i2 j2 [ V 2 , i2 = j2 and ∀(i1 j1 ) [ E 1 ,

(a) TD ((i1 , i2 )(j1 , j2 )) = TC2 (i2 ) ^ TC2 (j2 ) ^ TD1 (i1 j1 ) (a) IC (i) = IC1 (i) if i [ V 1 and i  V 2
(b) ID ((i1 , i2 )(j1 , j2 )) = IC2 (i2 ) ^ IC2 (j2 ) ^ ID1 (i1 j1 ) (b) IC (i) = IC2 (i) if i [ V 2 and i  V 1
(c) FD ((i1 , i2 )(j1 , j2 )) = FC2 (i2 ) ^ FC2 (j2 ) ^ FD1 (i1 j1 ) (c) IC (i) = IC1 (i) ^ IC2 (i) if i [ V 1 > V 2

(iii)
Definition


G1 < G2 of two neutrosophic graph G1 =
 14: The union

C1 , D1 and  G2 = C2 , D2 is defined as (C, D), where C =

TC , IC , FC  is a neutrosophic set on V = V 1 < V 2 and D = (a) FC (i) = FC1 (i) if i [ V 1 and i  V 2
TD , ID , FD is an another neutrosophic set on E = E 1 < E 2 , (b) FC (i) = FC2 (i) if i [ V 2 and i  V 1
which satisfies the following: (c) FC (i) = FC1 (i) ^ IC2 (i) if i [ V 1 > V 2

CAAI Trans. Intell. Technol., 2020, Vol. 5, Iss. 4, pp. 308–313

310 This is an open access article published by the IET, Chinese Association for Artificial Intelligence and
Chongqing University of Technology under the Creative Commons Attribution-NonCommercial-NoDerivs License (http://
creativecommons.org/licenses/by-nc-nd/3.0/)




(iv) TD , ID , FD is an another


neutrosophic
 set on E = { i, i2
i, j2 |i [
(a) TD (ij) = TD1 (ij) if ij [ E 1 and ij  E 2
V 1 , i2 j2 [ E 2 } < { i1 , k j1 , k |k [ V 2 , i1 j1 [ E 1 } < { i1 , i2 ×
(b) TD (ij) = TD2 (ij) if ij [ E 2 and ij  E 1 j1 , j2 |i1 j1 [ E 1 , i2 j2 [ E 2 } which satisfies the following:
(c) TD (ij) = TD1 (ij) ^ TD2 (ij) if ij [ E 1 > E 2
(i) ∀(i1 , i2 ) [ V 1 × V 2
(a) TC (i1 , i2 ) = TC1 (i1 ) _ TC2 (i2 )
(v) (b) IC (i1 , i2 ) = IC1 (i1 ) _ IC2 (i2 )
(a) ID (ij) = ID1 (ij) if ij [ E 1 and ij  E 2 (c) FC (i1 , i2 ) = FC1 (i1 ) ^ FC2 (i2 )
(b) ID (ij) = ID2 (ij) if ij [ E 2 and ij  E 1
(c) ID (ij) = ID1 (ij) ^ ID2 (ij) if ij [ E 1 > E 2 (ii) ∀i [ V 1 , ∀(i2 j2 ) [ E 2
(a) TD (i, i2 )(i, j2 ) = TC1 (i) _ TD2 (i2 j2 )
(b) ID (i, i2 )(i, j2 ) = IC1 (i) _ ID2 (i2 j2 )
(vi) (c) FD (i, i2 )(i, j2 ) = FC1 (i) ^ FD2 (i2 j2 )
(a) FD (ij) = FD1 (ij) if ij [ E 1 and ij  E 2
(b) FD (ij) = FD2 (ij) if ij [ E 2 and ij  E 1 (iii) ∀i1 j1 [ E 1 , ∀(i2 j2 ) [ E 2
(c) FD (ij) = FD1 (ij) ^ FD2 (ij) if ij [ E 1 > E 2 (a) TD (i1 , i2 )(j1 , j2 ) = TD1 (i1 j1 ) _ TD2 (i2 j2 )
(b) ID (i1 , i2 )(j1 , j2 ) = ID1 (i1 j1 ) _ ID2 (i2 j2 )
(c) FD (i1 , i2 )(j1 , j2 ) = FD1 (i1 j1 ) ^ FD2 (i2 j2 )
(vii)
(iv) ∀i1 j1 [ E 1 , k [ V 2
(a) TD (ij) = TD1 (i) _ TD2 (j) if ij [ E ′
(b) ID (ij) = ID1 (i) _ ID2 (j) if ij [ E ′
(c) FD (ij) = FD1 (i) ^ FD2 (j) if ij [ E ′ (a) TD (i1 , k)(j1 , k) = TD1 (i1 j1 ) _ TC2 (k)
(b) ID (i1 , k)(j1 , k) = ID1 (i1 j1 ) _ IC2 (k)
(c) FD (i1 , k)(j1 , k) = FD1 (i1 j1 ) ^ FC2 (k)
Definition 16: The direct product G1 ∗ G2 of two neutrosophic

graph G1  and G2 is defined as a pair (C, D), where C =

TC , IC , FC  is a neutrosophic set on V = V 1 × V 2 and
D = 4 Proposed Bellman–Ford algorithm for

TD , ID , FD is an another neutrosophic set on E = { i1 , i2 neutrosophic SPP
j1 , j2 |i1 j1 [ E 1 , i2 j2 [ E 2 }, which satisfies the following:
Our proposed algorithmic approach is the modification of classical
(i) ∀(i1 , i2 ) [ V 1 × V 2 Bellman–Ford algorithm for neutrosophic SPP. In this algorithm,
(a) TC (i1 , i2 ) = TC1 (i1 ) _ TC2 (i2 ) we have incorporated the uncertainties in Bellman–Ford algorithm
(b) IC (i1 , i2 ) = IC1 (i1 ) _ IC2 (i2 ) using neutrosophic set as an edge weight. We have shown the
(c) FC (i1 , i2 ) = FC1 (i1 ) ^ FC2 (i2 ) pseudocode of our proposed algorithm for neutrosophic SPP in
Algorithm 1 (see Fig. 1). The flowchart of our proposed algorithm
(ii) ∀(i1 j1 ) [ E 1 , ∀(i2 j2 ) [ E 2 is given in Fig. 2. The proposed algorithm finds all possible
shortest paths between the source node and all other nodes in the
(a) TD (i1 , i2 )(j1 , j2 ) = TD1 (i1 j1 ) _ TD2 (i2 j2 ) neutrosophic graph G. Our proposed algorithm needs that the
(b) ID (i1 , i2 )(j1 , j2 ) = ID1 (i1 j1 ) _ ID2 (i2 j2 ) neutrosophic graph does not consist of any neutrosophic cycles of
(c) FD (i1 , i2 )(j1 , j2 ) = FD1 (i1 j1 ) ^ TD2 (i2 j2 ) negative neutrosophic length. However, if the graph contains any
neutrosophic cycle, then our proposed algorithm is able to find it.
The source is denoted, respectively, by source.
Definition 17:
The lexicographic

product G1 · G2 of two neutrosophic
graph G1 =
C1 , D1 and  G2 = C2 , D2 is defined as a pair (C, D),
where
C = TC , IC , FC is a neutrosophic set on V = V 1 ×
V 2 and  5 Numerical examples
D

=  TD , ID , FD is an another
neutrosophic

 set on E = { i, i2 ×
i, j2 |i [ V 1 , i2 j2 [ E 2 } < { i1 , i2 j1 , j2 |i1 j1 [ E 1 , i2 j2 [ E 2 }
A numerical example of neutrosophic SPP is used to describe our
which satisfies the following:
proposed Bellman–Ford algorithm. For this purpose, we use an
example neutrosophic graph, shown in Fig. 3, with five vertices
(i) ∀(i1 , i2 ) and eight edges. Our modified Bellman–Ford algorithm detects the
(a) TC (i1 , i2 ) = TC1 (i1 ) _ TC2 (i2 ) = IC1 (i1 ) _ IC2 (i2 ) = FC1 (i1 )^ shortest path between the starting vertex and all other nodes in the
FC2 (i2 ) neutrosophic graph with neutrosophic sets as edge weights. For
this neutrosophic SPP, we consider the starting vertex is vertex s.
(ii) ∀i [ V 1 , ∀(i2 j2 ) [ E 2 The following eight neutrosophic number, shown in Table 1, are
used as edge weights of the neutrosophic graph. Those
(a) TD (i, i2 )(i, j2 ) = TC1 (i) _ TD2 (i2 j2 ) neutrosophic sets are number from one to eight. For the
(b) ID (i, i2 )(i, j2 ) = IC1 (i) _ ID2 (i2 j2 ) neutrosophic graph, shown in Fig. 3, we assign those neutrosophic
(c) FD (i, i2 )(i, j2 ) = FC1 (i) ^ FD2 (i2 j2 ) sets to the edges of the neutrosophic graph randomly.
The steps of our algorithm are given below:
(iii) ∀i1 j1 [ E 1 , ∀(i2 j2 ) [ E 2
Step 1: Let the starting vertex be s. It is changed permanent node and
(a) TD (i1 , i2 )(j1 , j2 ) = TD1 (i1 j1 ) _ TD2 (i2 j2 ) the shortest path distance, i.e. dist_score[u] between (s) and (s) is 0.
(b) ID (i1 , i2 )(j1 , j2 ) = ID1 (i1 j1 ) _ ID2 (i2 j2 ) The accessible adjacent vertices between starting vertex s are v1 , v2
(c) FD (i1 , i2 )(j1 , j2 ) = FD1 (i1 j1 ) ^ FD2 (i2 j2 ) and v3 . The score values (distance) of the each adjacent edge of
starting vertex (s) are calculated by using (4). Among all the 3
nodes, the shortest one (s)  (v1 ) is taken out with score value 0.43.
Definition

18: The strong
product G1 G2 of two neutrosophic graphs Step 2: Now, our proposed algorithm moves the node v1 and the
G1 =
C1 , D1 and
 G2 = C2 , D2 is defined as a pair (C, D), where finding the shortest path started from the vertex (v1 ). The lowest
C = TC , IC , FC is a neutrosophic set on V = V 1 × V 2 and D = score (shortest distance) from (s)  (v1 ) to its adjacent is

CAAI Trans. Intell. Technol., 2020, Vol. 5, Iss. 4, pp. 308–313

This is an open access article published by the IET, Chinese Association for Artificial Intelligence and 311
Chongqing University of Technology under the Creative Commons Attribution-NonCommercial-NoDerivs License (http://
creativecommons.org/licenses/by-nc-nd/3.0/)
Fig. 1 Algorithm 1: neutrosophic Bellman–Ford algorithm

Fig. 3 Neutrosophic network with neutrosophic numbers as arc lengths for

example 1

Table 1 Arc lengths of the neutrosophic graph, represented as

neutrosophic number

Index SVNs

1 k(4.6, 5.5, 8.6)l

2 k(4.7, 6.9, 8.5)l
3 k(6.2, 7.6, 8.2)l
4 k(6.2, 8.9, 9.1)l
5 k(4.4, 5.9, 7.2)l
6 k(6.6, 8.8, 10)l
7 k(6.3, 7.5, 8.9)l
8 k(6.2, 7.6, 8.2)l
Fig. 2 Flowchart of neutrosophic Bellman–Ford algorithm

determined. Any one is minimum than the path (s)  (v2 ) by 6 Conclusion
comparing all the score values.
Step 3: The vertex v2 is changed to permanent and all the searching Graph theory has many real-life applications to the problems in
of shortest path begins with vertex (v1 ) and vertex (s). We find the operations research, computer network, economics, systems
score of the each adjacent of the path (s)  (v2 ). The shortest analysis, urban traffic planning and transportation. In real-life
score among all the unvisited path is (s)  (v3 ). scenarios, however, uncertainty may exist in almost every graph
Step 4: Similarity, we determine the path with lowest score (i.e. the theoretic problem. Neutrosophic set is a popular and useful tool to
shortest path) between the source vertex and every other vertex (t), work in uncertain environment. This paper presents some new
the neutrosophic shortest path is (s)  (v2 )  t. operation of SVNG model. We describe the definition of regular

CAAI Trans. Intell. Technol., 2020, Vol. 5, Iss. 4, pp. 308–313

312 This is an open access article published by the IET, Chinese Association for Artificial Intelligence and
Chongqing University of Technology under the Creative Commons Attribution-NonCommercial-NoDerivs License (http://
creativecommons.org/licenses/by-nc-nd/3.0/)
SVNG, complete SVNG and strong SVNG. In this manuscript, we [28] Akram, M., Naz, S.: ‘Energy of pythagorean fuzzy graphs with applications’,
consider neutrosophic number to describe the edge weights of a Mathematics, 2018, 6, (8), p. 136
[29] Akram, M., Habib, A., Ilyas, F., et al.: ‘Specific types of Pythagorean fuzzy
neutrosophic graph for neutrosophic SPP. An algorithm is graphs and application to decision-making’, Math. Comput. Appl., 2018, 23,
introduced to solve this problem. The uncertainties are (3), p. 42
incorporated in Bellman–Ford algorithm for SPP using [30] Dey, A., Pradhan, R., Pal, A., et al.: ‘The fuzzy robust graph coloring problem’.
neutrosophic number as arc length. We use one numerical example Proc. of the 3rd Int. Conf. on Frontiers of Intelligent Computing: Theory and
to illustrate the effectiveness of the proposed algorithm. The Applications (FICTA) 2014, India, 2015, pp. 805–813
[31] Dey, A., Pal, A.: ‘Prim’s algorithm for solving minimum spanning tree problem
SVNG can be utilised to model the social network, image in fuzzy environment’, Ann. Fuzzy Math. Inf., 2016, 12, (3), pp. 419–430
processing, telecommunication, expert systems and computer [32] Akram, M., Dar, J.M., Naz, S.: ‘Certain graphs under Pythagorean fuzzy
networks. environment’, Complex Intell. Syst., 2019, 5, (2), pp. 127–144
[33] Ghorai, G., Pal, M.: ‘Certain types of product bipolar fuzzy graphs’, Int. J. Appl.
Comput. Math., 2017, 3, (2), pp. 605–619
[34] Garg, H., Nancy: ‘An improved score function for ranking neutrosophic sets and
7 References its application to decision-making process’, Int. J. Uncertain. Quantif., 2016, 6,
(5), pp. 377–385
[1] Zadeh, L.A.: ‘Fuzzy sets’, Inf. Control, 1965, 8, (3), pp. 338–353 [35] Garg, H., Nancy: ‘Non-linear programming method for multi-criteria decision
[2] Atanassov, K.T.: ‘Intuitionistic fuzzy sets’, in ‘Intuitionistic fuzzy sets’ (Springer, making problems under interval neutrosophic set environment’, Appl. Intell.,
Heidelberg, Germany, 1999), pp. 1–137 2018, 48, (8), pp. 2199–2213
[3] Garg, H., Kumar, K.: ‘An advanced study on the similarity measures of
[36] Garg, H., Nancy: ‘Multi-criteria decision-making method based on prioritized
intuitionistic fuzzy sets based on the set pair analysis theory and their
muirhead mean aggregation operator under neutrosophic set environment’,
application in decision making’, Soft Comput., 2018, 22, (15), pp. 4959–4970
Symmetry, 2018, 10, (7), p. 280
[4] Singh, S., Garg, H.: ‘Distance measures between type-2 intuitionistic fuzzy sets
and their application to multicriteria decision-making process’, Appl. Intell., [37] Garg, H., Nancy: ‘Some new biparametric distance measures on single-valued
2017, 46, (4), pp. 788–799 neutrosophic sets with applications to pattern recognition and medical
[5] Smarandache, F.: ‘Neutrosophy: neutrosophic probability, set, and logic: analytic diagnosis’, Information, 2017, 8, (4), p. 162
synthesis & synthetic analysis’ (American Research Press, USA), 1998 [38] Chakraborty, A.: ‘A new score function of pentagonal neutrosophic number and
[6] Smarandache, F.: ‘A unifying field in logics: neutrosophic logic’, in ‘Philosophy’ its application in networking problem’, Int. J. Neutrosophic Sci., 2020, 1, (1),
(American Research Press, USA, 1999), pp. 1–141 pp. 40–51
[7] Smarandache, F.: ‘A unifying field in logics: neutrosophic logic. Neutrosophy, [39] Edalatpanah, S.: ‘A direct model for triangular neutrosophic linear
neutrosophic set, neutrosophic probability: neutrsophic logic. Neutrosophy, programming’, Int. J. Neutrosophic Sci., 2020, 1, (1), pp. 19–28
neutrosophic set, neutrosophic probability’ (Infinite Study, USA, 2005) [40] Alhabib, R., Salama, A.: ‘Using moving averages to pave the neutrosophic time
[8] Ye, J.: ‘Single-valued neutrosophic minimum spanning tree and its clustering series’, Int. J. Neutrosophic Sci., 2020, 3, (1), pp. 14–20
method’, J. Intell. Syst., 2014, 23, (3), pp. 311–324 [41] Garg, H., Nancy: ‘Algorithms for possibility linguistic single-valued
[9] Yang, H.L., Guo, Z.L., She, Y., et al.: ‘On single valued neutrosophic relations’, neutrosophic decision-making based on COPRAS and aggregation operators
J. Intell. Fuzzy Syst., 2016, 30, (2), pp. 1045–1056 with new information measures’, Measurement, 2019, 138, pp. 278–290
[10] Naz, S., Rashmanlou, H., Malik, M.A.: ‘Operations on single valued [42] Das, S.K., Edalatpanah, S.: ‘A new ranking function of triangular neutrosophic
neutrosophic graphs with application’, J. Intell. Fuzzy Syst., 2017, 32, (3), number and its application in integer programming’, Int. J. Neutrosophic Sci.,
pp. 2137–2151 2020, 4, (2), pp. 20–30
[11] Borzooei, R.A., Rashmanlou, H., Samanta, S., et al.: ‘Regularity of vague [43] Dey, A., Broumi, S., Bakali, A., et al.: ‘A new algorithm for finding minimum
graphs’, J. Intell. Fuzzy Syst., 2016, 30, (6), pp. 3681–3689 spanning trees with undirected neutrosophic graphs’, Granular Comput., 2019,
[12] Akram, M., Waseem, N., Dudek, W.A.: ‘Certain types of edge m-polar fuzzy 4, (1), pp. 63–69
graphs’, Iran. J. Fuzzy Syst., 2017, 14, (4), pp. 27–50 [44] Dey, A., Pradhan, R., Pal, A., et al.: ‘A genetic algorithm for solving fuzzy
[13] Akram, M.: ‘Certain bipolar neutrosophic competition graphs’, J. Indonesian shortest path problems with interval type-2 fuzzy arc lengths’, Malaysian
Math. Soc., 2017, 24, (1), pp. 1–25
J. Comput. Sci., 2018, 31, (4), pp. 255–270
[14] Akram, M., Siddique, S.: ‘Neutrosophic competition graphs with applications’,
[45] Broumi, S., Bakali, A., Talea, M., et al.: ‘Spanning tree problem with
J. Intell. Fuzzy Syst., 2017, 33, (2), pp. 921–935
neutrosophic edge weights’, Procedia Comput. Sci., 2018, 127, pp. 190–199
[15] Dey, A., Pal, A., Pal, T.: ‘Interval type 2 fuzzy set in fuzzy shortest path problem’,
Mathematics, 2016, 4, (4), p. 62 [46] Dey, A., Pal, A., Long, H.V., et al.: ‘Fuzzy minimum spanning tree with interval
[16] Shannon, A., Atanassov, K.: ‘On a generalization of intuitionistic fuzzy graphs’, type 2 fuzzy arc length: formulation and a new genetic algorithm’, Soft Comput.,
NIFS, 2006, 12, (1), pp. 24–29 2020, 24, (6), pp. 3963–3974
[17] Parvathi, R., Karunambigai, M.: ‘Intuitionistic fuzzy graphs’, in ‘Computational [47] Dey, A., Pal, A.: ‘Computing the shortest path with words’, Int. J. Adv. Intell.
intelligence, theory and applications’ (Springer, India, 2006), pp. 139–150 Paradigms, 2019, 12, (3–4), pp. 355–369
[18] Parvathi, R., Thamizhendhi, G.: ‘Domination in intuitionistic fuzzy graphs’, [48] Broumi, S., Talea, M., Bakali, A., et al.: ‘Shortest path problem in fuzzy,
Notes Intuitionistic Fuzzy Sets, 2010, 16, (2), pp. 39–49 intuitionistic fuzzy and neutrosophic environment: an overview’, Complex
[19] Parvathi, R., Karunambigai, M., Atanassov, K.T.: ‘Operations on intuitionistic Intell. Syst., 2019, 5, (4), pp. 371–378
fuzzy graphs’. 2009 IEEE Int. Conf. on Fuzzy Systems, Korea, 2009, [49] Broumi, S., Nagarajan, D., Bakali, A., et al.: ‘The shortest path problem in
pp. 1396–1401 interval valued trapezoidal and triangular neutrosophic environment’, Complex
[20] Rashmanlou, H., Samanta, S., Pal, M., et al.: ‘Intuitionistic fuzzy graphs with Intell. Syst., 2019, 5, (4), pp. 391–402
categorical properties’, Fuzzy Inf. Eng., 2015, 7, (3), pp. 317–334 [50] Broumi, S., Bakali, A., Talea, M., et al.: ‘Shortest path problem under interval
[21] Rashmanlou, H., Samanta, S., Pal, M., et al.: ‘Bipolar fuzzy graphs with valued neutrosophic setting’, Int. J. Adv. Trends Comput. Sci. Eng., 2019, 8,
categorical properties’, Int. J. Comput. Intell. Syst., 2015, 8, (5), pp. 808–818 (1), pp. 216–222
[22] Rashmanlou, H., Borzooei, R., Samanta, S., et al.: ‘Properties of interval valued [51] Broumi, S., Talea, M., Bakali, A., et al.: ‘Energy and spectrum analysis of
intuitionistic (s, t)–fuzzy graphs’, Pac. Sci. Rev. A, Nat. Sci. Eng., 2016, 18, (1), interval valued neutrosophic graph using Matlab’, Neutrosophic Sets Syst.,
pp. 30–37 2019, 24, (1), pp. 46–60
[23] Sahoo, S., Pal, M.: ‘Different types of products on intuitionistic fuzzy graphs’, [52] Smarandache, F.: ‘Extension of hypergraph to n-superhypergraph and to
Pac. Sci. Rev. A, Nat. Sci. Eng., 2015, 17, (3), pp. 87–96 plithogenic n-superhypergraph, and extension of hyperalgebra to n-ary
[24] Wang, J.: ‘A novel firefly algorithm for portfolio optimization problem’, IAENG (classical-/neutro-/anti-) hyperalgebra’, Neutrosophic Sets Syst., 2020, 33, (1),
Int. J. Appl. Math., 2019, 49, (1), pp. 45–50 pp. 290–296
[25] Akram, M., Akmal, R.: ‘Operations on intuitionistic fuzzy graph structures’,
[53] Dubois, D., Parde, H.: ‘Fuzzy sets and systems: theory and applications’, vol. 144
Fuzzy Inf. Eng., 2016, 8, (4), pp. 389–410
(Academic press, Newyork, 1980)
[26] Liu, Q.: ‘Resistance distance and Kirchhoff index of two edge-subdivision corona
graphs’, IAENG Int. J. Appl. Math., 2019, 49, (1), pp. 127–133 [54] Smarandache, F.: ‘Invisible paradox’, ‘Neutrosophy/neutrosophic probability,
[27] Prasertpong, R., Siripitukdet, M.: ‘Applying generalized rough set concepts to set, and logic’ (American Research Press, Rehoboth, 1998), pp. 22–23
approximation spaces of semigroups’, IAENG Int. J. Appl. Math., 2019, 49, [55] Wang, H., Smarandache, F., Zhang, Y., et al.: ‘Single valued neutrosophic sets’,
(1), pp. 51–60 Rev. Air Force Acad., 2010, 1, p. 10

CAAI Trans. Intell. Technol., 2020, Vol. 5, Iss. 4, pp. 308–313

This is an open access article published by the IET, Chinese Association for Artificial Intelligence and 313
Chongqing University of Technology under the Creative Commons Attribution-NonCommercial-NoDerivs License (http://
creativecommons.org/licenses/by-nc-nd/3.0/)