Neural Comput & Applic (2017) 28:1817–1834

DOI 10.1007/s00521-015-2154-y

ORIGINAL ARTICLE

Complex neutrosophic set

Mumtaz Ali1 • Florentin Smarandache2

Received: 6 July 2015 / Accepted: 21 December 2015 / Published online: 9 January 2016
 The Natural Computing Applications Forum 2016

Abstract Complex fuzzy sets and complex intuitionistic Keywords Fuzzy set  Intuitionistic fuzzy set  Complex
fuzzy sets cannot handle imprecise, indeterminate, incon- fuzzy set  Complex intuitionistic fuzzy set  Neutrosophic
sistent, and incomplete information of periodic nature. To set  Complex neutrosophic set
overcome this difficulty, we introduce complex neutro-
sophic set. A complex neutrosophic set is a neutrosophic
set whose complex-valued truth membership function, 1 Introduction
complex-valued indeterminacy membership function, and
complex-valued falsehood membership functions are the Fuzzy sets were first proposed by Zadeh in the seminal
combination of real-valued truth amplitude term in asso- paper [38]. This novel concept is used successfully in
ciation with phase term, real-valued indeterminate ampli- modeling uncertainty in many fields of real life. A fuzzy set
tude term with phase term, and real-valued false amplitude is characterized by a membership function l with the range
term with phase term, respectively. Complex neutrosophic [0,1]. Fuzzy sets and their applications have been exten-
set is an extension of the neutrosophic set. Further set sively studied in different aspects from the last few decades
theoretic operations such as complement, union, intersec- such as control [19, 38], reasoning [44], pattern recognition
tion, complex neutrosophic product, Cartesian product, [19, 44], and computer vision [44]. Fuzzy sets become an
distance measure, and d-equalities of complex neutro- important area for the research in medical diagnosis [29],
sophic sets are studied here. A possible application of engineering [19], etc. A large amount of the literature on
complex neutrosophic set is presented in this paper. fuzzy sets can be found in [8, 9, 15, 21, 30, 40–43]. In fuzzy
Drawbacks and failure of the current methods are shown, set, the membership degree of an element is single value
and we also give a comparison of complex neutrosophic set between 0 and 1. Therefore, it may not always be true that
to all such methods in this paper. We also showed in this the non-membership degree of an element in a fuzzy set is
paper the dominancy of complex neutrosophic set to all equal to 1 minus the membership degree because there is
current methods through the graph. some degree of hesitation. Thus, Atanassov [2] introduced
intuitionistic fuzzy sets in 1986 which incorporate the
hesitation degree called hesitation margin. The hesitation
margin is defining as 1 minus the sum of membership and
non-membership. Therefore, the intuitionistic fuzzy set is
& Mumtaz Ali characterized by a membership function l and non-mem-
[email protected]
bership function m with range [0,1]. An intuitionistic fuzzy
Florentin Smarandache set is the generalization of fuzzy set. Intuitionistic fuzzy sets
[email protected]
can successfully be applied in many fields such as medical
1
Department of Mathematics, Quaid-e-Azam University, diagnosis [29], modeling theories [11], pattern recognition
Islamabad 45320, Pakistan [31], and decision making [17].
2
University of New Mexico, 705 Gurley Ave., Gallup, Ramot et al. [23] proposed an innovative concept to the
NM 87301, USA extension of fuzzy sets by initiating the complex fuzzy sets

123
1818 Neural Comput & Applic (2017) 28:1817–1834

where the degree of membership l is traded by a complex- distance measure, Cartesian products, relations, projec-
valued of the form tions, and cylindric extensions. The complex fuzzy set has
pffiffiffiffiffiffiffi only one extra phase term, while complex intuitionistic
rs ð xÞ  ejxS ðxÞ ; j ¼ 1
fuzzy set has two additional phase terms. Several properties
where rS(x) and xS ðxÞ are both belongs to [0,1] and of complex intuitionistic fuzzy sets have been studied such
rs ð xÞ : ejxS ðxÞ has the range in complex unit disk. Complex as complement, union, intersection, T-norm, and S-norm.
fuzzy set is completely a different approach from the fuzzy Smarandache [28] in 1998 introduced Neutrosophy that
complex number discussed by Buckley [4–7], Nguyen et al. studies the origin, nature, and scope of neutralities and their
[21], and Zhang et al. [41]. The complex-valued mem- interactions with distinct ideational spectra. A neutrosophic
bership function of the complex fuzzy set has an amplitude set is characterized by a truth membership function T, an
term with the combination of a phase term which gives indeterminacy membership function I and a falsehood
wavelike characteristics to it. Depending on the phase term membership function F. Neutrosophic set is powerful
gives a constructive or destructive interference. Thus, mathematical framework which generalizes the concept of
complex fuzzy set is different from conventional fuzzy set classical sets, fuzzy sets [38], intuitionistic fuzzy sets [2],
[38], fuzzy complex set [23], type 2 fuzzy set [19], etc. due interval valued fuzzy sets [30], paraconsistent sets [28],
to the character of wavelike. The complex fuzzy set [23] dialetheist sets [28], paradoxist sets [28], and tautological
still preserves the characterization of uncertain information sets [28]. Neutrosophic sets handle the indeterminate and
through the amplitude term having value in the range of inconsistent information that exists commonly in our daily
[0,1] with the addition of a phase term. Ramot et al. [23, life. Recently neutrosophic sets have been studied by
24] discussed several properties of complex fuzzy sets such several researchers around the world. Wang et al. [33]
as complement, union, and intersection. with sufficient studied single-valued neutrosophic sets in order to use
amount of illustrative examples. Some more theory on them in scientific and engineering fields that give an
complex fuzzy sets can be seen in [10, 35]. Ramot et al. additional possibility to represent uncertainty, incomplete,
[24] also introduced the concept of complex fuzzy logic imprecise, and inconsistent data. Hanafy et al. [13, 14]
which is a novel framework for logical reasoning. The studied the correlation coefficient of neutrosophic set. Ye
complex fuzzy logic is a generalization of fuzzy logic, [35] studied the correlation coefficient of single-valued
based on complex fuzzy set. In complex fuzzy logic [24], neutrosophic sets. Broumi and Smaradache presented the
the inference rules are constructed and fired in such way correlation coefficient of interval neutrosophic set in [3].
that they are closely resembled to traditional fuzzy logic. Salama et al. [26] studied neutrosophic sets and neutro-
Complex fuzzy logic [24] is constructed to hold the sophic topological spaces. Some more literature on neu-
advantages of fuzzy logic while enjoying the features of trosophic sets can be found in [12–14, 18, 20, 25, 27, 32,
complex numbers and complex fuzzy sets. Complex fuzzy 34, 36, 37, 40].
logic is not only a linear extension to the conventional Pappis [22] studied the notion of ‘‘proximity measure,’’
fuzzy logic but rather a natural extension to those problems with an attempt to show that ‘‘precise membership values
that are very difficult or impossible to describe with one- should normally be of no practical significance.’’ Pappis
dimensional grades of membership. Complex fuzzy sets observed that the max–min compositional rule of inference
have found their place in signal processing [23], physics is preserved with respect to ‘‘approximately equal’’ fuzzy
[23], stock marketing [23] etc. sets. An important generalization of the work of Pappis
The concept of complex intuitionistic fuzzy set in short proposed by Hong and Hwang [15] which is mainly based
CIFS is introduced by Alkouri and Saleh in [1]. The that the max–min compositional rule of inference is pre-
complex intuitionistic fuzzy set is an extension of complex served with respect to ‘‘approximately equal fuzzy sets’’
fuzzy set by adding complex-valued non-membership and ‘‘approximately equal’’ fuzzy relation. But, Cai noticed
grade to the definition of complex fuzzy set. The complex that both the Pappis and Hong and Hwang approaches were
intuitionistic fuzzy sets are used to handle the information confined to fixed e. Therefore, Cai [8, 9] takes a different
of uncertainty and periodicity simultaneously. The com- approach and introduced d-equalities of fuzzy sets. Cai
plex-valued membership and non-membership function can proposed that if two fuzzy sets are equal to an extent of d,
be used to represent uncertainty in many corporal quanti- then they are said to be d-equal. The notions of d-equality
ties such as wave function in quantum mechanics, impe- are significance in both the fuzzy statistics and fuzzy rea-
dance in electrical engineering, complex amplitude, and soning. Cai [8, 9] applied them for assessing the robustness
decision-making problems. The novel concept of phase of fuzzy reasoning as well as in synthesis of real-time fuzzy
term is extend in the case of complex intuitionistic fuzzy systems. Cai also gave several reliability examples of d-
set which appears in several prominent concepts such as equalities [8, 9]. Zhang et al. [39] studied the d-equalities

123
Neural Comput & Applic (2017) 28:1817–1834 1819

of complex fuzzy set by following the philosophy of Ramot 2. Further, we studied set theoretic operations of complex
et al. [23, 24] and Cai [8, 9]. They mainly focus on the neutrosophic sets such as complement, union, inter-
results of Cai’s work [8, 9] to introduce d-equalities of section complex neutrosophic product, and Cartesian
complex fuzz sets, and thus, they systematically develop product.
distance measure, equality and similarity of complex fuzzy 3. We also introduced the novel concept ‘‘the game of
sets. Zhang et al. [39] then applied d-equalities of complex winner, neutral, and loser’’ for phase terms.
fuzzy sets in a signal processing application. 4. We studied a distance measure on complex neutro-
This paper is an extension of the work of Ramot et al. sophic sets which we have used in the application.
[23], Alkouri and Saleh [1], Cai [8, 9], and Zhang et al. 5. We introduced d-equalities of complex neutrosophic
[39] to neutrosophic sets. Basically, we follow the philos- set and studied their properties.
ophy of the work of Ramot et al. [23] to introduce complex 6. We also gave an algorithm for signal processing using
neutrosophic set. The complex neutrosophic is character- complex neutrosophic sets.
ized by complex-valued truth membership function, com- 7. Drawbacks and failures of the current methods
plex-valued indeterminate membership function, and presented in this paper.
complex-valued falsehood membership function. Further, 8. Finally, we gave the comparison of complex neutro-
complex neutrosophic set is the mainstream over all sophic sets to the current methods.
because it not only is the generalization of all the current
The organization of this paper is as follows. In
frameworks but also describes the information in a com-
Sect. 2, we presented some basic and fundamental
plete and comprehensive way.
concepts of neutrosophic sets, complex fuzzy sets, and
complex intuitionistic fuzzy sets. In the next section, we
1.1 Why complex neutrosophic set can handle
introduced complex neutrosophic sets and gave some
the indeterminate information in periodicity
interpretation of complex neutrosophic set for intuition.
We also introduced the basic set theoretic operations of
As we can see that uncertainty, indeterminacy, incom-
complex neutrosophic sets such as complement, union,
pleteness, inconsistency, and falsity in data are periodic in
intersection, complex neutrosophic product, and Carte-
nature, to handle these types of problems, the complex
sian product in the current section. Further, in this
neutrosophic set plays an important role. A complex neu-
section, the game of winner, neutral, and loser is
trosophic set S is characterized by a complex-valued truth
introduced for the phase terms in the case of union and
membership function TS(x), complex-valued indeterminate
intersection of two complex neutrosophic sets. It is
membership function IS(x), and complex-valued false
completely an innovative approach for the phase terms.
membership function FS(x) whose range is extended from
In Sect. 4, we introduced distance measure on complex
[0,1] to the unit disk in the complex plane. The complex
neutrosophic sets, d-equality on complex neutrosophic
neutrosophic sets can handle the information which is
sets and studied some of their properties. An application
uncertain, indeterminate, inconsistent, incomplete,
in signal processing is presented for the possible uti-
ambiguous, false because in TS(x), the truth amplitude term
lization of complex neutrosophic set in the Sect. 5. In
and phase term handle uncertainty and periodicity, in IS(x),
Sect. 6, we give the drawbacks of fuzzy sets, intu-
the indeterminate amplitude term and phase term handle
itionistic fuzzy sets, neutrosophic sets, complex fuzzy
indeterminacy and periodicity, and in FS(x), the false
sets, and complex intuitionistic fuzzy sets. We also give
amplitude term and phase term handle the falsity with
a comparison of different current methods to complex
periodicity. Complex neutrosophic set is an extension of
neutrosophic set in this section. Further, the dominancy
the neutrosophic set with three additional phase terms.
of complex neutrosophic sets over other existing meth-
Thus, the complex neutrosophic set deals with the
ods is shown in this section.
information/data which have uncertainty, indeterminacy,
We now review some basic concepts of neutrosophic
and falsity that are in periodicity while both the complex
sets, single-valued neutrosophic set, complex fuzzy sets,
fuzzy set and complex intuitionistic fuzzy sets cannot deal
and complex intuitionistic fuzzy sets.
with indeterminacy, inconsistency, imprecision, vagueness,
doubtfulness, error, etc. in periodicity.
The contributions of this paper are:
2 Literature review
1. We introduced complex neutrosophic set which deals
with uncertainty, indeterminacy, impreciseness, incon- In this section, we present some basic material which will
sistency, incompleteness, and falsity of periodic help in our later pursuit. The definitions and notions are
nature. taken from [1, 23, 28, 39].

123
1820 Neural Comput & Applic (2017) 28:1817–1834

Definition 2.1 [28] Neutrosophic set. Definition 2.4 [23] Intersection of neutrosophic sets.
Let X be a space of points and let x 2 X. A neutrosophic Let A and B be two complex neutrosophic sets in a
set S in X is characterized by a truth membership function universe of discourse X. Then, the intersection of A and B is
TS, an indeterminacy membership function IS, and a falsity denoted as A \ B, which is defined by
membership function FS. TS(x), IS(x), and FS(x) are real A \ B ¼ fðx; TA ð xÞ ^ TB ð xÞ; IA ð xÞ _ IB ð xÞ; FA ð xÞ _ FB ð xÞÞ : x 2 X g
standard or non-standard subsets of ]0-, 1?[, and
TS ; IS ; FS : X ! 0 ; 1þ ½. The neutrosophic set can be rep- for all x 2 X.
resented as The definitions and other notions of complex fuzzy sets
S ¼ fðx; TS ð xÞ; IS ð xÞ; FS ð xÞÞ : x 2 X g are given as follows.
There is no restriction on the sum of TS(x), IS(x), and Definition 2.5 [23] Complex fuzzy set.
FS(x), so 0  TS ð xÞ þ IS ð xÞ þ FS ð xÞ  3þ .
A complex fuzzy set S, defined on a universe of dis-
From philosophical point view, the neutrosophic set
course X, is characterized by a membership function gS(-
takes the value from real standard or non-standard subsets
x) that assigns any element x 2 X a complex-valued grade
of 0 ; 1þ ½. Thus, it is necessary to take the interval [0,1]
of membership in S. The values gS(x) all lie within the unit
instead of 0 ; 1þ ½ for technical applications. It is difficult
circle in the complex plane and thus all of the form
to apply 0 ; 1þ ½ in the real-life applications such as
pS ð xÞ:ej:lS ðxÞ where pS(x) and lS(x) are both real-valued and
engineering and scientific problems.
pS ð xÞ 2 ½0; 1. Here, pS(x) is termed as amplitude term and
A single-valued neutrosophic set [23] is characterized
by a truth membership function, TS(x), an indeterminacy ej:lS ðxÞ is termed as phase term. The complex fuzzy set may
membership function IS(x) and a falsity membership be represented in the set form as
function FS(x) with TS(x), IS(x), FS(x) 2 [0, 1] for all S ¼ fðx; gS ð xÞÞ : x 2 X g:
x 2 X. If X is continuous, then
Z The complex fuzzy set is denoted as CFS.
ðTS ð xÞ; IS ð xÞ; FS ð xÞÞ We now present set theoretic operations of complex
S¼ for all x 2 X:
x fuzzy sets.
X
Definition 2.6 [23] Complement of complex fuzzy set.
If X is discrete, then
X ðTS ð xÞ; IS ð xÞ; FS ð xÞÞ Let S be a complex fuzzy set on X, and gS ð xÞ ¼
S¼ for all x 2 X:
x pS ð xÞ:ej:lS ðxÞ its complex-valued membership function. The
X
complement of S denoted as c(S) and is specified by a
Actually, SVNS is an instance of neutrosophic set that function
can be used in real-life situations such as decision-making,
scientific, and engineering applications. We will use single- gcðSÞ ð xÞ ¼ pcðSÞ ð xÞ:ej:lcðSÞ ðxÞ ¼ ð1  pS ð xÞÞ:ejð2plS ðxÞÞ :
valued neutrosophic set to define complex neutrosophic set. Definition 2.7 [23] Union of complex fuzzy sets.
We now give some set theoretic operations of neutro-
sophic sets. Let A and B be two complex fuzzy sets on X, and
gA ð xÞ ¼ rA ð xÞ:ej:lA ðxÞ and gB ð xÞ ¼ rB ð xÞ:ej:lB ðxÞ be their
Definition 2.2 [33] Complement of neutrosophic set.
membership functions, respectively. The union of A and
The complement of a neutrosophic set S is denoted by B is denoted as A [ B, which is specified by a function
c(S) and is defined by gA[B ð xÞ ¼ rA[B ð xÞ:ej:lA[B ðxÞ ¼ ðrA ð xÞ _ rB ð xÞÞ:ejðlA ðxÞ_lB ðxÞÞ
TcðSÞ ð xÞ ¼ FS ð xÞ; IcðSÞ ð xÞ ¼ 1  IS ð xÞ; FcðSÞ ð xÞ
¼ TS ð xÞ for all x 2 X: where _ denote the max operator.

Definition 2.3 [23] Union of neutrosophic sets. Definition 2.8 [23] Intersection of complex fuzzy sets.

Let A and B be two complex neutrosophic sets in a Let A and B be two complex fuzzy sets on X, and gA ð xÞ ¼
universe of discourse X. Then, the union of A and B is rA ð xÞ:ej:lA ðxÞ and gB ð xÞ ¼ rB ð xÞ:ej:lB ðxÞ be their membership
denoted by A [ B, which is defined by functions, respectively. The intersection of A and B is
denoted as A \ B, which is specified by a function
A [ B ¼ fðx; TA ð xÞ _ TB ð xÞ; IA ð xÞ ^ IB ð xÞ; FA ð xÞ ^ FB ð xÞÞ : x 2 X g
gA\B ð xÞ ¼ rA\B ð xÞ:ej:lA\B ðxÞ ¼ ðrA ð xÞ ^ rB ð xÞÞ:ejðlA ðxÞ^lB ðxÞÞ
for all x 2 X, and _ denote the max operator and ^ denote
the min operator, respectively. where ^ denote the max operator.

123
Neural Comput & Applic (2017) 28:1817–1834 1821

We now give some basic definitions and set theoretic The union of A and B is denoted as A [ B, which is
operations of complex intuitionistic fuzzy sets. specified by a function
Definition 2.9 [39] Let A and B be two complex fuzzy gA[B ð xÞ ¼ pA[B ð xÞ:ej:lA[B ðxÞ ¼ ðpA ð xÞ _ pB ð xÞÞ:ejðlA ðxÞ_lB ðxÞÞ
sets on X, and gA ð xÞ ¼ rA ð xÞ:ej:lA ðxÞ and gB ð xÞ ¼ and
rB ð xÞ:ej:lB ðxÞ be their membership functions, respectively. fA[B ð xÞ ¼ rA[B ð xÞ:ej:xA[B ðxÞ ¼ ðrA ð xÞ ^ rB ð xÞÞ:ejðxA ðxÞ^xB ðxÞÞ
The complex fuzzy product of A and B, denoted as A  B,
and is specified by a function where _ and ^ denote the max and min operator,
l  respectively.
A ðxÞ:lB
ð xÞ
gAB ð xÞ ¼ rAB ð xÞ:ej:lAB ðxÞ ¼ ðrA ð xÞ:rB ð xÞÞ:ej2p 2p 2p :
Definition 2.14 [1] Intersection of complex intuitionistic
Definition 2.10 [39] Let A and B be two complex fuzzy fuzzy sets.
sets on X, and gA ð xÞ ¼ rA ð xÞ:ej:lA ðxÞ and gB ð xÞ ¼ Let A and B be two complex intuitionistic fuzzy sets on
rB ð xÞ:ej:lB ðxÞ be their membership functions, respectively. X, and gA ð xÞ ¼ pA ð xÞ:ej:lA ðxÞ , fA ð xÞ ¼ rA ð xÞ:ej:xA ðxÞ and
Then, A and B are said to be d equal if and only if
gB ð xÞ ¼ pB ð xÞ:ej:lB ðxÞ and fB ð xÞ ¼ rB ð xÞ:ej:xB ðxÞ be their
dðA; BÞ  1  d, where 0  d  1.
membership and non-membership functions, respectively.
For more literature on d-equality, we refer to [8, 9] and The intersection of A and B is denoted as A \ B, which is
[35]. specified by a function
Definition 2.11 [1] Complex intuitionistic fuzzy set. gA\B ð xÞ ¼ pA\B ð xÞ:ej:lA\B ðxÞ ¼ ðpA ð xÞ ^ pB ð xÞÞ:ejðlA ðxÞ^lB ðxÞÞ and
fA\B ð xÞ ¼ rA\B ð xÞ:ej:xA\B ðxÞ ¼ ðrA ð xÞ _ rB ð xÞÞ:ejðxA ðxÞ_xB ðxÞÞ
A complex intuitionistic fuzzy set S, defined on a uni-
verse of discourse X, is characterized by a membership
where ^ and _ denote the min and max operators,
function gS ð xÞ and non-membership function fS ð xÞ,
respectively.
respectively, that assign an element x 2 X a complex-val-
Next, the notion of complex neutrosophic set is
ued grade to both membership and non-membership in
introduced.
S. The values of gS ð xÞ and fS ð xÞ all lie with in the unit
circle in the complex plane and are of the form gS ð xÞ ¼
pS ð xÞ:ej:lS ðxÞ and fS ð xÞ ¼ rS ð xÞ:ej:xS ðxÞ , where pS ð xÞ; rS ð xÞ; 3 Complex neutrosophic set
lS ð xÞ, and xS ð xÞ are all real-valued and pS ð xÞ, rS ð xÞ 2
pffiffiffiffiffiffiffi
½0; 1 with j ¼ 1. The complex intuitionistic fuzzy set In this section, we introduced the innovative concept of
can be represented as complex neutrosophic set. The definition of complex neu-
trosophic set is as follows.
S ¼ fðx; gS ð xÞ; fS ð xÞÞ : x 2 U g:
Definition 3.1 Complex neutrosophic set.
It is denoted as CIFS.
A complex neutrosophic set S defined on a universe of
Definition 2.12 [1] Complement of complex intuitionis- discourse X, which is characterized by a truth membership
tic fuzzy set. function TS ðxÞ, an indeterminacy membership function
Let S be a complex intuitionistic fuzzy set, and gS ð xÞ ¼ IS ðxÞ, and a falsity membership function FS ðxÞ that assigns
pS ð xÞ:ej:lS ðxÞ and fS ð xÞ ¼ rS ð xÞ:ej:xS ðxÞ its membership and a complex-valued grade of TS ðxÞ, IS ðxÞ, and FS ðxÞ in S for
non-membership functions, respectively. The complement any x 2 X. The values TS ðxÞ, IS ðxÞ, FS ðxÞ and their sum
of S, denoted as c(S), is specified by a function may all within the unit circle in the complex plane and so is
of the following form,
gcðSÞ ð xÞ ¼ pcðSÞ ð xÞ:ej:lcðSÞ ðxÞ ¼ rS ð xÞ:ejð2plS ðxÞÞ and
TS ðxÞ ¼ pS ðxÞ:ejlS ðxÞ ; IS ðxÞ ¼ qS ðxÞ:ejmS ðxÞ and FS ðxÞ
fcðSÞ ð xÞ ¼ rcðSÞ ð xÞ:ej:xcðSÞ ðxÞ ¼ pS ð xÞ:ejð2pxS ðxÞÞ ¼ rS ðxÞ:ejxS ðxÞ
Definition 2.13 [1] Union of complex intuitionistic fuzzy where pS ðxÞ, qS ðxÞ, rS ðxÞ and lS ðxÞ, mS ðxÞ, xS ðxÞ are,
sets. respectively, real valued and pS ðxÞ; qS ðxÞ; rS ðxÞ 2 ½0; 1
Let A and B be two complex intuitionistic fuzzy sets on such that  0  pS ðxÞ þ qS ðxÞ þ rS ðxÞ  3þ .
X, and gA ð xÞ ¼ pA ð xÞ:ej:lA ðxÞ , fA ð xÞ ¼ rA ð xÞ:ej:xA ðxÞ and The complex neutrosophic set S can be represented in
gB ð xÞ ¼ pB ð xÞ:ej:lB ðxÞ and fB ð xÞ ¼ rB ð xÞ:ej:xB ðxÞ be their set form as
membership and non-membership functions, respectively. S ¼ fðx; TS ðxÞ ¼ aT ; IS ðxÞ ¼ aI ; FS ðxÞ ¼ aF Þ : x 2 X g;

123
1822 Neural Comput & Applic (2017) 28:1817–1834

where TS :X ! faT :aT 2 C; jaT j  1g; IS :X ! faI :aI 2 rS ðxÞ equal to cS ðxÞ with the false phase term equal to zero
C; jaI j  1g and FS :X ! faF :aF 2 C; jaF j  1g and for all x. Thus, it has seen that a complex neutrosophic set
can be easily transformed into a neutrosophic set. From this
jTS ðxÞ þ IS ðxÞ þ FS ðxÞj  3:
Throughout the paper, complex neutrosophic set refers discussion, it is concluded that the truth amplitude term is
to a neutrosophic set with complex-valued truth member- equivalent to the real-valued grade of truth membership
ship function, complex-valued indeterminacy membership function, the indeterminate amplitude term is equivalent to
function, and complex-valued falsity membership function the real-valued grade of indeterminate membership func-
while the term neutrosophic set with real-valued truth tion, and the false amplitude term is essentially equivalent
membership function, indeterminacy membership function, to the real-valued grade of false membership function. The
and falsity membership function. only distinguishing factors are truth phase term, indeter-
minate phase term, and false phase term. This differs the
3.1 Interpretation of complex neutrosophic set complex neutrosophic set from the ordinary neutrosophic
set. In simple words, if we omit all the three phase terms,
The concept of complex-valued truth membership function, the complex neutrosophic set will automatically convert
complex-valued indeterminacy membership function, and into neutrosophic set effectively. All this discussion is
complex-valued falsity membership function is not a sim- supported by the reality that pS ðxÞ, qS ðxÞ, and rS ðxÞ have
ple task in understanding. In contrast, real-valued truth range [0,1] which is for real-valued grade of truth mem-
membership function, real-valued indeterminacy member- bership, real-valued grade of indeterminate membership,
ship function, and real-valued falsity membership function and real-valued grade of false membership.
in the interval [0,1] can be easily intuitive. It should also be noted that complex neutrosophic sets
The notion of complex neutrosophic set can be easily are the generalization of complex fuzzy sets, conventional
understood from the form of its truth membership function, fuzzy sets, complex intuitionistic fuzzy sets and intuition-
indeterminacy membership function, and falsity member- istic fuzzy sets, classical sets. This means that complex
ship function which appears in above Definition 3.1. neutrosophic set is an advance generalization to all the
existence methods and due to this feature, the concept of
TS ðxÞ ¼ pS ðxÞ:ejlS ðxÞ ; IS ðxÞ ¼ qS ðxÞ:ejmS ðxÞ and FS ðxÞ complex neutrosophic set is a distinguished and novel one.
¼ rS ðxÞ:ejxS ðxÞ
3.2 Numerical example of a complex neutrosophic
It is clear that complex grade of truth membership
set
function is defined by a truth amplitude term pS ðxÞ and a
truth phase term lS ðxÞ. Similarly, the complex grade of
Example 3.2 Let X ¼ fx1 ; x2 ; x3 g be a universe of dis-
indeterminacy membership function is defined as an inde-
course. Then, S be a complex neutrosophic set in X as given
terminate amplitude term qS ðxÞ and an indeterminate phase
below:
term mS ðxÞ, while the complex grade of falsity membership
function is defined by false amplitude term rS ðxÞ and a false  
0:6ej:0:8 ; 0:3:ej: 4 ; 0:5:ej:0:3
3p

phase term xS ðxÞ, respectively. It should be noted that the S¼

 x1 
truth amplitude term pS ðxÞ is equal to jTS ðxÞj, the amplitude
0:7ej:0 0:2:ej:0:9 ; 0:1:ej: 3
2p

of TS ðxÞ. Also, the indeterminate amplitude term qS ðxÞ is þ

x2
equal to jIS ðxÞj and the false amplitude term rS ðxÞ is equal ð0:9ej:0:1 ; 0:4:ej:p ; 0:7:ej:0:7 Þ
to jFS ðxÞj. þ :
x3
Complex neutrosophic sets are the generalization of
neutrosophic sets. It is a easy task to represent a neutro- 3.3 Set theoretic operations on complex neutrosophic
sophic set in the form of complex neutrosophic set. For set
instance, the neutrosophic set S is characterized by a real-
valued truth membership function aS ðxÞ, indeterminate Ramot et al. [23], calculated the complement of member-
membership function bS ðxÞ, and falsehood membership ship phase term lS ð xÞ by several possible method such as
function cS ðxÞ. By setting the truth amplitude term pS ðxÞ lcS ð xÞ ¼ lS ð xÞ; 2p  lS ð xÞ. Zhang [39] defined the com-
equal to aS ðxÞ, and the truth phase term lS ðxÞ equal to zero plement of the membership phase term by taking the
for all x and similarly we can set the indeterminate rotation of lS ð xÞ by p radian as lcS ð xÞ ¼ lS ð xÞ þ p.
amplitude term qS ðxÞ equal to bS ðxÞ and the indeterminate To define the complement of a complex neutrosophic
phase term equal to zero, while the false amplitude term set, we simply take the neutrosophic complement [29] for

123
Neural Comput & Applic (2017) 28:1817–1834 1823

the truth amplitude term pS ð xÞ, indeterminacy amplitude where the truth membership function TA[B ð xÞ, the inde-
term qS ð xÞ, and falsehood amplitude term rS ð xÞ. For phase terminacy membership function IA[B ð xÞ, and the falsehood
terms, we take the complements defined in [23]. We now membership function FA[B ð xÞ are defined by
proceed to define the complement of complex neutrosophic
TA[B ð xÞ ¼ ½ðpA ð xÞ _ pB ð xÞÞ:ej:lTA[B ðxÞ ;
set.
IA[B ð xÞ ¼ ½ðqA ð xÞ ^ qB ð xÞÞ:ej:mIA[B ðxÞ ;
Definition 3.3 Complement of complex neutrosophic set.
FA[B ð xÞ ¼ ½ðrA ð xÞ ^ rB ð xÞÞ:ej:xFA[B ðxÞ ;
Let S ¼ fðx; TS ðxÞ; IS ðxÞ; FS ðxÞÞ : x 2 X g be a complex
neutrosophic set in X. Then, the complement of a complex where _ and ^ denote the max and min operators,
neutrosophic set S is denoted as c(S) and is defined by respectively. To calculate phase the terms ej:lA[B ðxÞ , ej:mA[B ðxÞ ,
   and ej:xA[B ðxÞ , we define the following:
cðSÞ ¼ x; TSc ðxÞ; ISc ðxÞ; FSc ðxÞ : x 2 X ;
    Definition 3.7 Let A and B be two complex neutrosophic
where TSc ðxÞ ¼ c pS ð xÞ:ej:lS ðxÞ , ISc ðxÞ ¼ c qS ð xÞ:ej:mS ðxÞ ,
    sets in X with complex-valued truth membership functions
and FSc ðxÞ ¼ c rS ð xÞ:ej:xS ðxÞ in which c pS ð xÞ:ej:lS ðxÞ ¼ TA ð xÞ and TB ð xÞ, complex-valued indeterminacy member-
c
cðpS ð xÞÞ:ej:ls ðxÞ is such that cðpS ð xÞÞ ¼ rS ðxÞ and mcS ð xÞ ¼ ship functions IA ð xÞ and IB ð xÞ, and complex-valued false-
mS ð xÞ; 2p  mS ð xÞ or mS ð xÞ þ p. Similarly, cðrS ð xÞ: hood membership functions FA ð xÞ and FB ð xÞ, respectively.
c
ej:lS ðxÞ Þ ¼ cðrS ð xÞÞ:ej:xs ðxÞ , where cðqS ð xÞÞ ¼ 1  qS ð xÞ and The union of the complex neutrosophic sets A and B is
mcS ð xÞ ¼ mS ð xÞ; 2p  mS ð xÞ or mS ð xÞ þ p. denoted by A [N B which is associated with the function:
  c
Finally, c rS ð xÞ:ej:lS ðxÞ ¼ cðrS ð xÞÞ:ej:xs ðxÞ , where h : fðaT ;aI ;aF Þ : aT ;aI ;aF 2 C; jaT þaI þaF j3; jaT j; jaI j; jaF j1g
cðrS ð xÞÞ ¼ pS ð xÞ and xcS ð xÞ ¼ xS ð xÞ; 2p  xS ð xÞ or  fðbT ;bI ;bF Þ : bT ;bI ;bF 2 C; jbT þbI þbF j3; jbT j; jbI j; jbF j1g
xS ð xÞ þ p. ! fðdT ;dI ;dF Þ : dT ;dI ;dF 2 C; jdT þdI þdF j3; jdT j; jdI j; jdF j1g;

Proposition 3.4 Let A be a complex neutrosophic set on

where a; b; d, a0 ; b0 ; d0 , and a00 ; b00 ; d00 are the complex truth
X. Then, cðcð AÞÞ ¼ A:
membership, complex indeterminacy membership, and
Proof By definition 3.1, we can easily prove it. complex falsity membership of A, B, and A [ B,
respectively.
Proposition 3.5 Let A and B be two complex neutro-
sophic sets on X. Then, cðA \ BÞ ¼ cð AÞ [ cðBÞ. A complex value is assigned by h, that is, for all x 2 X,
Definition 3.6 Union of complex neutrosophic sets. hðTA ð xÞ; TB ð xÞÞ ¼ TA[B ð xÞ ¼ dT ;
hðIA ð xÞ; IB ð xÞÞ ¼ IA[B ð xÞ ¼ dI and
Ramot et al. [23] defined the union of two complex
fuzzy sets A and B as follows. hðFA ð xÞ; FB ð xÞÞ ¼ FA[B ð xÞ ¼ dF :
Let lA ð xÞ ¼ rA ð xÞ:ej:xA ðxÞ and lB ð xÞ ¼ rB ð xÞ:ej:xB ðxÞ be This function h must obey at least the following axio-
the complex-valued membership functions of A and B, matic conditions.
respectively. Then, the membership union of A [ B is given For any a; b; c; d; a0 ; b0 ; c0 ; d0 ; a00 ; b00 ; c00 ; d 00 2 fx : x 2 C;
by lA[B ð xÞ ¼ ½rA ð xÞ  rB ð xÞ:ej:xA[B ðxÞ . Since rA ð xÞ and j xj  1g:
rB ð xÞ are real-valued and belong to ½0; 1, the operators max
and min can be applied to them. For calculating phase term • Axiom 1: jhT ða; 0Þj ¼ jaj; jhI ða0 ; 1Þj ¼ ja0 j and
xA[B ð xÞ, they give several methods. jhF ða00 ; 1Þj ¼ ja00 j (boundary condition).
Now we define the union of two complex neutrosophic • Axiom 2: hT ða; bÞ ¼ hT ðb; aÞ; hI ða0 ; b0 Þ ¼ hI ðb0 ; a0 Þ and
sets as follows: hF ða00 ; b00 Þ ¼ hF ðb00 ; a00 Þ (commutativity condition).
Let A and B be two complex neutrosophic sets in X, • Axiom 3: if jbj  jd j, then jhT ða; bÞj  jhT ða; d Þj and if
where jb0 j  jd 0 j, then jhI ða0 ; b0 Þj  jhI ða0 ; d0 Þj and if
jb00 j  jd00 j, then jhF ða00 ; b00 Þj  jhF ða00 ; d00 Þj (monotonic
A ¼ fðx; TA ð xÞ; IA ð xÞ; FA ð X ÞÞ : x 2 X g and condition).
B ¼ fðx; TB ð xÞ; IB ð xÞ; FB ð X ÞÞ : x 2 X g: • Axiom 4: hT ðhT ða; bÞ; cÞ ¼ hT ða; hT ðb; cÞÞ, hI ðhI ða0 ;
b0 Þ; c0 Þ ¼ hI ða0 ; hI ðb0 ; c0 ÞÞ and hF ðhF ða00 ; b00 Þ; c00 Þ ¼
Then the union of A and B is denoted as A [N B and is hF ða00 ; hF ðb00 ; c00 ÞÞ (associative condition).
given as
It may be possible in some cases that the following are
A [N B ¼ fðx; TA[B ð xÞ; IA[B ð xÞ; FA[B ð xÞÞ : x 2 X g also hold:

123
1824 Neural Comput & Applic (2017) 28:1817–1834

 
• Axiom 5: h is continuous function (continuity). 0:6ej:0:8 ; 0:3:ej: 4 ; 0:5:ej:0:3
3p

• Axiom 6: jhT ða; aÞj [ jaj, jhI ða0 ; a0 Þj\ja0 j and A¼

 x1 
jhF ða00 ; a00 Þj\ja00 j (superidempotency). 0:7ej:0 0:2:ej:0:9 ; 0:1:ej: 3
2p

• Axiom 7: jaj  jcj and jbj  jdj, then jhT ða; þ

x2
bÞj  jhT ðc; dÞj, also ja0 j jc0 j and jb0 j jd0 j, then ð0:9ej:0:1 ; 0:4:ej:p ; 0:7:ej:0:7 Þ
jhI ða0 ; b0 Þj jhI ðc0 ; d 0 Þj and ja00 j jc00 j and jb00 j  jd 00 j, þ ;
x3
then jhF ða00 ; b00 Þj jhF ðc0 ; d0 Þj (strict monotonicity).
and
The phase term of complex truth membership function,  p

complex indeterminacy membership function, and complex 0:8ej:0:9 ; 0:1:ej:4 ; 0:4:ej:0:5
B¼
falsity membership function belongs to ð0; 2pÞ. To define  x1 
0:6ej:0:1 ; 1:ej:0:6 ; 0:01:ej: 3
4p

the phase terms, we suppose that lTA[B ð xÞ ¼ lA[B ð xÞ, þ

mIA[B ð xÞ ¼ mA[B ð xÞ, and xFA[B ð xÞ ¼ xA[B ð xÞ. Now we take  x2 
those forms which Ramot et al. presented in [23] to define 0:2ej:0:3 ; 0:ej:2p ; 0:5:ej:0:5
þ ;
the phase terms of TA[B ð xÞ, IA[B ð xÞ, and FA[B ð xÞ, x3
respectively. Then
(a) Sum:  
0:8:ej:0:9 ; 0:3:ej: 4 ;0:5:ej:0:5
3p

A [N B ¼ ;
lA[B ð xÞ ¼ lA ð xÞ þ lB ð xÞ; x1
 
0:6:ej:0:1 ;0:2:ej:0:9 ;0:01;ej: 3 ð0:2:ej:0:3 ; 0:ej:2p ;0:5:ej:0:7 Þ
4p
mA[B ð xÞ ¼ mA ð xÞ þ mB ð xÞ;
; :
xA[B ð xÞ ¼ xA ð xÞ þ xB ð xÞ: x2 x3

(b) Max: Definition 3.9 Intersection of complex neutrosophic sets.

lA[B ð xÞ ¼ maxðlA ð xÞ; lB ð xÞÞ; Let A and B be two complex neutrosophic sets in X,
mA[B ð xÞ ¼ maxðmA ð xÞ; mB ð xÞÞ; where
xA[B ð xÞ ¼ maxðxA ð xÞ; xB ð xÞÞ: A ¼ fðx; TA ð xÞ; IA ð xÞ; FA ð X ÞÞ : x 2 X g and
(c) Min: B ¼ fðx; TB ð xÞ; IB ð xÞ; FB ð X ÞÞ : x 2 X g:

lA[B ð xÞ ¼ minðlA ð xÞ; lB ð xÞÞ; Then, the intersection of A and B is denoted as A \N B

mA[B ð xÞ ¼ minðmA ð xÞ; mB ð xÞÞ; and is defined as
xA[B ð xÞ ¼ minðxA ð xÞ; xB ð xÞÞ: A \N B ¼ fðx; TA\B ð xÞ; IA\B ð xÞ; FA\B ð xÞÞ : x 2 X g;

(d) ‘‘The game of winner, neutral, and loser’’: where the truth membership function TA\B ð xÞ, the inde-
terminacy membership function IA\B ð xÞ, and the falsehood

lA ðxÞ if pA [ pB membership function FA\B ð xÞ are given as:
lA[B ð xÞ ¼ ;
lB ðxÞ if pB [ pA TA\B ð xÞ ¼ ½ðpA ð xÞ ^ pB ð xÞÞ:ej:lTA\B ðxÞ ;

mA ðxÞ if qA \qB
mA[B ð xÞ ¼ ; IA\B ð xÞ ¼ ½ðqA ð xÞ _ qB ð xÞÞ:ej:mIA\B ðxÞ ;
mB ðxÞ if qB \qA
 FA\B ð xÞ ¼ ½ðrA ð xÞ _ rB ð xÞÞ:ej:xFA\B ðxÞ ;
xA ðxÞ if rA \rB
xA[B ð xÞ ¼ :
xB ðxÞ if rB \rA where _ and ^ denote the max and min operators,
respectively. We calculate phase terms ej:lA\B ðxÞ , ej:mA\B ðxÞ ,
The game of winner, neutral, and loser is a novel con- and ej:xA\B ðxÞ after define the following:
cept, and it is the generalization of the concept ‘‘winner
Definition 3.10 Let A and B be two complex neutro-
take all’’ introduced by Ramot et al. [23] for the union of
phase terms. sophic sets in X with complex-valued truth membership
functions TA ð xÞ and TB ð xÞ, complex-valued indeterminacy
Example 3.8 Let X ¼ fx1 ; x2 ; x3 g be a universe of dis- membership functions IA ð xÞ and IB ð xÞ, and complex-valued
course. Let A and B be two complex neutrosophic sets in falsehood membership functions FA ð xÞ and FB ð xÞ,
X as shown below: respectively. The intersection of the complex neutrosophic

123
Neural Comput & Applic (2017) 28:1817–1834 1825

sets A and B is denoted by A \N B which is associated with Proof Here we only prove part 1. Let A; B; C be three
the function: complex neutrosophic sets in X and TA ð xÞ; IA ð xÞ,FA ð xÞ,
/ : fðaT ;aI ;aF Þ : aT ;aI ;aF 2 C; jaT þaI þaF j3; jaT j; jaI j; jaF j1g TB ð xÞ; IB ð xÞ; FB ð xÞ and TC ð xÞ; IC ð xÞ; FV ð xÞ, respectively,
 fðbT ;bI ;bF Þ : bT ;bI ;bF 2 C; jbT þbI þbF j3; jbT j; jbI j; jbF j1g be their complex-valued truth membership function, com-
! fðdT ;dI ;dF Þ : dT ;dI ;dF 2 C; jdT þdI þdF j3; jdT j; jdI j; jdF j1g;
plex-valued indeterminate membership function, and
complex-valued falsehood membership functions. Then,
where a;b;d, a0 ;b0 ;d0 , and a00 ;b00 ;d 00 are the complex truth we have
membership, complex indeterminacy membership, and TðA[BÞ\C ð xÞ ¼ pðA[BÞ\C ðxÞ:ej:lðA[BÞ\C ðxÞ
complex falsity membership of A, B, and A\B, respec- ¼ minðpA[B ðxÞ; pC ð xÞÞ:ej: minðlA[B ðxÞ;lC ðxÞÞ ;
tively. / assigned a complex value, that is, for all x 2 X;
¼ minðmaxðpA ðxÞ; pB ðxÞÞ; pC ð xÞÞ
/ðTA ð xÞ; TB ð xÞÞ ¼ TA\B ð xÞ ¼ dT ; : ej: minðmaxðlA ðxÞ;lB ðxÞÞ;lC ðxÞÞ ;
/ðIA ð xÞ; IB ð xÞÞ ¼ IA\B ð xÞ ¼ dI and ¼ maxðminðpA ðxÞ; pc ðxÞÞ; minðpB ð xÞ; pC ðxÞÞÞ
/ðFA ð xÞ; FB ð xÞÞ ¼ FA\B ð xÞ ¼ dF : : ej:maxðminðlA ðxÞ;lC ðxÞÞ;minðlB ðxÞ;lC ðxÞÞÞ ;
¼ maxðpA\C ðxÞ; pB\C ð xÞÞ: ej:maxðlA\C ðxÞ;lB\C ðxÞÞ ;
/ must satisfy at least the following axiomatic ¼ pðA\CÞ[ðB\CÞ ðxÞ:ej:lðA\CÞ[ðB\CÞ ðxÞ ¼ TðA\CÞ[ðB\CÞ ðxÞ:
conditions.
For any a; b; c; d; a0 ; b0 ; c0 ; d0 ; a00 ; b00 ; c00 ; d 00 2 fx : x 2 C; Similarly, on the same lines, we can show it for
j xj  1g: IðA[BÞ\C ð xÞ and FðA[BÞ\C ð xÞ, respectively. h

• Axiom 1: If jbj ¼ 1, then j/T ða; bÞj ¼ jaj. If jb0 j ¼ 0, Proposition 3.12 Let A and B be two complex neutro-
then j/I ða0 ; b0 Þj ¼ ja0 j and if j/F ða00 ; b00 Þj ¼ ja00 j sophic sets in X. Then,
(boundary condition).
1. ðA [ BÞ \ A ¼ A;
• Axiom 2: /T ða; bÞ ¼ /T ðb; aÞ; /I ða0 ; b0 Þ ¼ /I ðb0 ; a0 Þ,
2. ðA \ BÞ [ A ¼ A:
and /F ða00 ; b00 Þ ¼ /F ðb00 ; a00 Þ (commutative condition).
• Axiom 3: if jbj  jdj, then j/T ða; bÞj  j/T ða; dÞj and if
Proof We prove it for part 1. Let A and B be two complex
jb0 j  jd 0 j, then j/I ða0 ; b0 Þj  j/I ða0 ; d0 Þj and if
neutrosophic sets in X and TA ð xÞ; IA ð xÞ,FA ð xÞ and
jb00 j  jd00 j, then j/F ða00 ; b00 Þj  j/F ða00 ; d 00 Þj (monotonic
TB ð xÞ; IB ð xÞ; FB ð xÞ, respectively, be their complex-valued
condition).
truth membership function, complex-valued indeterminate
• Axiom 4: /T ð/T ða; bÞ; cÞ ¼ /T ða; /T ðb; cÞÞ, /I ð/I ða0 ;
membership function, and complex-valued falsehood
b0 Þ; c0 Þ ¼ /I ða0 ; /I ðb0 ; c0 ÞÞ, and /F ð/F ða00 ; b00 Þ; c00 Þ ¼
membership functions. Then,
/F ða00 ; /F ðb00 ; c00 ÞÞ (associative condition).
TðA[BÞ\A ð xÞ ¼ pðA[BÞ\A ð xÞ: ej:lðA[BÞ\A ðxÞ
The following axioms also hold in some cases.
¼ minðpA[B ð xÞ; pA ð xÞÞ
• Axiom 5: / is continuous function (continuity).
: ei:minðlA[B ðxÞ;lA ðxÞÞ ;
• Axiom 6: j/T ða; aÞj [ jaj, j/I ða0 ; a0 Þj \ja0 j, and
j/F ða00 ; a00 Þj\ja00 j (superidempotency). ¼ minðmaxðpA ð xÞ; pB ð xÞÞ; pA ð xÞÞ
• Axiom 7: jaj  jcj and jbj  jd j, then j/T ða; : ei:minðmaxðlA ðxÞ;lB ðxÞÞ;lA ðxÞÞ
bÞj  j/T ðc; d Þj, also ja0 j jc0 j and jb0 j jd0 j, then ¼ TA ð xÞ:
j/I ða0 ; b0 Þj j/I ðc0 ; d0 Þj and ja00 j jc00 j and jb00 j  jd 00 j,
then j/F ða00 ; b00 Þj j/F ðc00 ; d00 Þj (strict monotonicity). Similarly, we can show it for IðA[BÞ\A ð xÞ and
FðA[BÞ\A ð xÞ, respectively. h
We can easily calculate the phase terms ej:lA\B ðxÞ ,
ej:mA\B ðxÞ , and ej:xA\B ðxÞ on the same lines by winner, neutral, Definition 3.13 Let A and B be two complex neutro-
and loser game. sophic sets on X, and TA ð xÞ ¼ pA ð xÞ:ej:lA ðxÞ , IA ð xÞ ¼
qA ð xÞ:ej:mA ðxÞ , FA ð xÞ ¼ rA ð xÞ:ej:xA ðxÞ and TB ð xÞ ¼ pB ð xÞ:
Proposition 3.11 Let A; B; C be three complex neutro-
ej:lB ðxÞ , IB ð xÞ ¼ qB ð xÞ:ej:mB ðxÞ , FB ð xÞ ¼ rB ð xÞ:ej:xB ðxÞ ,
sophic sets on X. Then,
respectively, be their complex-valued truth membership
1. ðA [ BÞ \ C ¼ ðA \ C Þ [ ðA \ BÞ; function, complex-valued indeterminacy membership
2. ðA \ BÞ [ C ¼ ðA [ C Þ \ ðA [ BÞ: function, and complex-valued falsity membership function.

123
1826 Neural Comput & Applic (2017) 28:1817–1834

The complex neutrosophic product of A and B denoted as IA1 A2 AN ð xÞ ¼ qA1 A2 AN ð xÞ:ej:mA1 A2 AN ðxÞ
A  B and is specified by the functions,
¼ maxðqA1 ðx1 Þ; qA2 ðx2 Þ; . . .; qAN ðxN ÞÞ
TAB ð xÞ ¼ pAB ð xÞ:ej:lAB ðxÞ ¼ ðpA ð xÞ:pB ð xÞÞ:ej:2pð 2p : 2p Þ ;
lA ðxÞ lB ð xÞ

: ejmaxðmA1 ðx1 Þ;mA2 ðx2 Þ;...;mAN ðxN ÞÞ ;

I ð xÞ ¼ q ð xÞ:ej:lAB ðxÞ ¼ ðq ð xÞ:q ð xÞÞ:ej:2pð 2p : 2p Þ ;
mA ðxÞ mB ðxÞ

AB AB A B
and
¼ ðrA ð xÞ:rB ð xÞÞ:ej:2pð Þ:
xA ðxÞ lB ð xÞ
j:lAB ð xÞ :
FAB ð xÞ ¼ rAB ð xÞ:e 2p 2p

FA1 A2 AN ð xÞ ¼ rA1 A2 AN ð xÞ:ej:xA1 A2 AN ðxÞ
Example 3.14 Let X ¼ fx1 ; x2 ; x3 g and let ¼ maxðrA1 ðx1 Þ; rA2 ðx2 Þ; . . .; rAN ðxN ÞÞ
 
0:6ej1:2p ; 0:3ej0:5p ; 1:0ej0:1p : ejmaxðxA1 ðx1 Þ;xA2 ðx2 Þ;...;xAN ðxN ÞÞ ;
A¼
x1 where x ¼ ðx1 ; x2 ; . . .; xN Þ 2 |fflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflffl}
X  X    X :
ð1:0e ; 0:2ej:3p ; 0:5ej0:4p Þ
j2p
N
þ
x2
 j1:6p

0:8e ; 0:1ej1:2 ; 0:6ej0:1p 4 Distance measure and d-equalities of complex
þ ;
x3 neutrosophic sets
ð0:6ej1:2p ; 0:1ej0:4p ; 1:0ej0:1p Þ
B¼ In this section, we introduced distance measure and other
x1
 j1:2p
 operational properties of complex neutrosophic sets.
1:0e ; 0:3ej:2p ; 0:7ej0:5p
þ
x2 Definition 4.1 Let CNð X Þ be the collection of all com-
 
0:2e j1:6p
; 0:2ej1:3p ; 0:7ej0:1p plex neutrosophic sets on X and A; B 2 CNð X Þ. Then, A
þ B if and only if TA ð xÞ  TB ð xÞ such that the amplitude
x3
terms pA ð xÞ  pB ð xÞ and the phase terms lA ð xÞ  lB ð xÞ,
Then and IA ð xÞ IB ð xÞ such that the amplitude terms
 
0:36ej0:72p ; 0:3ej0:1p ; 1:0ej0:0025p qA ð xÞ qB ð xÞ and the phase terms mA ð xÞ mB ð xÞ whereas
AB¼ ;
x1 FA ð xÞ FB ð xÞ such that the amplitude terms rA ð xÞ rB ð xÞ
ð1:0ej1:2p ; 0:06ej3p ; 0:35ej0:1p Þ and the phase terms xA ð xÞ xB ð xÞ.
;
x3 Definition 4.2 Two complex neutrosophic sets A and
 
0:16ej1:28p ; 0:02ej0:78p ; 0:42ej0:005p B are said to equal if and only if pA ð xÞ ¼ pB ð xÞ,
x3 qA ð xÞ ¼ qB ð xÞ, and rA ð xÞ ¼ rB ð xÞ for amplitude terms and
lA ð xÞ ¼ lB ð xÞ, mA ð xÞ ¼ mB ð xÞ, xA ð xÞ ¼ xB ð xÞ for phase
Definition 3.15 Let An be N complex neutrosophic sets on
terms (arguments).
X ðn ¼ 1; 2; . . .; N Þ, and TAn ð xÞ ¼ pAn ð xÞ:ej:lAn ðxÞ ,
IAn ð xÞ ¼ qAn ð xÞ:ej:mAn ðxÞ
, and FAn ð xÞ ¼ rAn ð xÞ:ej:xAn ðxÞ be their Definition 4.3 A distance of complex neutrosophic sets is
complex-valued membership function, complex-valued inde- a function dCNS :CNð X Þ  CNð X Þ ! ½0; 1 such that for any
terminacy membership function and complex-valued non- A; B; C 2 CNð X Þ
membership function, respectively. The Cartesian product of 1. 0  dCNS ðA; BÞ  1;
An , denoted as A1  A2      AN , specified by the function 2. dCNS ðA; BÞ ¼ 0 if and only if A ¼ B;
TA1 A2 AN ð xÞ ¼ pA1 A2 AN ð xÞ:ej:lA1 A2 AN ðxÞ 3. dCNS ðA; BÞ ¼ dCNS ðB; AÞ;
¼ minðpA1 ðx1 Þ; pA2 ðx2 Þ; . . .; pAN ðxN ÞÞ 4. dCNS ðA; BÞ  dCNS ðA; CÞ þ dCNS ðC; BÞ:

: ejminðlA1 ðx1 Þ;lA2 ðx2 Þ;...;lAN ðxN ÞÞ ; Let dCNS :CNð X Þ  CNð X Þ ! ½0; 1 be a function which
is defined as

maxðsupx2X jpA ð xÞ  pB ð xÞj; supx2X jqA ð xÞ  qB ð xÞj; supx2X jrA ð xÞ  rB ð xÞjÞ;

dCNS ðA; BÞ ¼ max
maxð2p1 supx2X jlA ð xÞ  lB ð xÞj; 2p1 supx2X jmA ð xÞ  mB ð xÞj; 2p1 supx2X jxA ð xÞ  xB ð xÞjÞ

123
Neural Comput & Applic (2017) 28:1817–1834 1827

Theorem 4.4 The function dCNS ðA; BÞ defined above is a 7. If A ¼ ðd1 ÞB and B ¼ ðd2 ÞC; then A ¼ ðdÞC; where
distance function of complex neutrosophic sets on X. d ¼ d1 d2 :
Proof The proof is straightforward. h
Proof 4.7 Properties 1–4, 6 can be proved easily. We only
Definition 4.5 Let A and B be two complex neutrosophic prove 5 and 7.
sets on X, and TA ð xÞ ¼ pA ð xÞ:ej:lA ðxÞ , IA ð xÞ ¼ qA ð xÞ:ej:mA ðxÞ , 5. Since A ¼ ðda ÞB for all a 2 J; we have
FA ð xÞ ¼ rA ð xÞ:ej:xA ðxÞ and TB ð xÞ ¼ pA ð xÞ:ej:lB ðxÞ , IB ð xÞ ¼
qB ð xÞ:ej:mB ðxÞ ; FB ð xÞ ¼ rB ð xÞ:ej:xB ðxÞ are their complex-val-
ued truth membership, complex-valued indeterminacy

maxðsupx2X jpA ð xÞ  pB ð xÞj; supx2X jqA ð xÞ  qB ð xÞj; supx2X jrA ð xÞ  rB ð xÞjÞ;

dCNS ðA; BÞ ¼ max  1  da
maxð2p1 supx2X jlA ð xÞ  lB ð xÞj; 2p1 supx2X jmA ð xÞ  mB ð xÞj; 2p1 supx2X jxA ð xÞ  xB ð xÞjÞ

membership, and complex-valued falsity membership Therefore,

functions, respectively. Then, A and B are said to be d-
supjpA ð xÞ  pB ð xÞj  1  sup da ;
equal, if and only if dCNS ðA; BÞ  1  d, where 0  d  1. It x2X a2J
is denoted by A ¼ ðdÞB. supjqA ð xÞ  qB ð xÞj  1  sup da ;
x2X a2J
Proposition 4.6 For complex neutrosophic sets A, B, and
supjrA ð xÞ  rB ð xÞj  1  sup da ; and
C, the following holds. x2X a2J

1. A ¼ ð0ÞB, 1
supjl ð xÞ  lB ð xÞj  1  sup da ;
2. A ¼ ð1ÞB if and only if A ¼ B; 2p x2X A a2J
3. If A ¼ ðdÞB if and only if B ¼ ðdÞA; 1
supjmA ð xÞ  mB ð xÞj  1  sup da
4. A ¼ ðd1 ÞB and d2  d1 , then A ¼ ðd2 ÞB; 2p x2X a2J
5. If A ¼ ðda ÞB, then A ¼ ðsupa2J da ÞB for all a 2 J, 1
where J is an index set, supjxA ð xÞ  xB ð xÞj  1  sup da :
2p x2X a2J
6. If A ¼ ðd0 ÞB and there exist a unique d such that A ¼
ðdÞB; then d0  d for all A; B Thus,

0 1
maxðsupx2X jpA ð xÞ  pB ð xÞj; supx2X jqA ð xÞ  qB ð xÞj; supx2X jrA ð xÞ  rB ð xÞjÞ;
dCNS ðA; BÞ ¼ max@ 1 1 1 A
max supjlA ð xÞ  lB ð xÞj; supjmA ð xÞ  mB ð xÞj; supjxA ð xÞ  xB ð xÞj
2p x2X 2p x2X 2p x2X
 1  sup da
a2J

123
1828 Neural Comput & Applic (2017) 28:1817–1834

Hence, A ¼ ðsupa2J da ÞB. which implies

7. Since A ¼ ðd1 ÞB, we have

maxðsupx2X jpA ð xÞ  pB ð xÞj; supx2X jqA ð xÞ  qB ð xÞj; supx2X jrA ð xÞ  rB ð xÞjÞ;

dCNS ðA; BÞ ¼ max  1  d1
maxð2p1 supx2X jlA ð xÞ  lB ð xÞj; 2p1 supx2X jmA ð xÞ  mB ð xÞj; 2p1 supx2X jxA ð xÞ  xB ð xÞjÞ

which implies supjpB ð xÞ  pC ð xÞj 1  d2 ;

x2X
supjpA ð xÞ  pB ð xÞj  1  d1 ;
x2X supjqB ð xÞ  qC ð xÞj 1  d2 ;
x2X
supjqA ð xÞ  qB ð xÞj  1  d1 ;
x2X supjrB ð xÞ  rC ð xÞj  1  d1 and
x2X
supjrA ð xÞ  rB ð xÞj  1  d1 and 1
x2X 1
supjl ð xÞ  lC ð xÞj 1  d2 ; supjm ð xÞ  mC ð xÞj  1  d2 ;
1 1 2p x2X B 2p x2X B
supjl ð xÞ  lB ð xÞj  1  d1 ; supjm ð xÞ  mB ð xÞj  1  d1
2p x2X A 2p x2X A 1
1
supjxB ð xÞ  xC ð xÞj 1  d2 :
supjx ð xÞ  xB ð xÞj  1  d1 : 2p x2X
2p x2X A

Now,
Also we have B ¼ ðd2 ÞC, so

0 1
maxðsupx2X jpB ð xÞ  pC ð xÞj; supx2X jqB ð xÞ  qC ð xÞj; supx2X jrB ð xÞ  rC ð xÞjÞ;
dCNS ðB; CÞ ¼ max@ 1 A  1  d2
max 2p1 supx2X jlB ð xÞ  lC ð xÞj; 2p1 supx2X jmB ð xÞ  mC ð xÞj; supjxB ð xÞ  xC ð xÞj
2p x2X

123
Neural Comput & Applic (2017) 28:1817–1834 1829

maxðsupx2X jpA ð xÞ  pC ð xÞj; supx2X jqA ð xÞ  qC ð xÞj; supx2X jrA ð xÞ  rC ð xÞjÞ;

dCNS ðA; C Þ ¼ max
maxð2p1 supx2X jlA ð xÞ  lC ð xÞj; 2p1 supx2X jmA ð xÞ  mC ð xÞj; 2p1 supx2X jxA ð xÞ  xC ð xÞjÞ
0 1
maxðsupx2X jpA ð xÞ  pB ð xÞj; supx2X jqA ð xÞ  qB ð xÞj; supx2X jrA ð xÞ  rB ð xÞjÞþ
B maxðsup jp ð xÞ  p ð xÞj; sup jq ð xÞ  q ð xÞj; sup jr ð xÞ  r ð xÞjÞ; C
B x2X B C x2X B C x2X B C C
 maxB C
@ maxð2p1 supx2X jlA ð xÞ  lB ð xÞj; 2p1 supx2X jmA ð xÞ  mB ð xÞj; 2p1 supx2X jxA ð xÞ  xB ð xÞjÞþ A
maxð2p1 supx2X jlB ð xÞ  lC ð xÞj; 2p1 supx2X jmB ð xÞ  mC ð xÞj; 2p1 supx2X jxB ð xÞ  xC ð xÞjÞ
 maxðð1  d1 Þ þ ð1  d2 Þ; ð1  d1 Þ þ ð1  d2 ÞÞ ¼ ð1  d1 Þ þ ð1  d2 Þ ¼ 1  ðd1 þ d2  1Þ;

From Definition 4.3, dCNS ðA; CÞ  1. Therefore, dCNS sampled N times. Suppose that Sl0 ðk0 Þ denote the k0 th of the
ðA; C Þ  1  d1 d2 ¼ 1  d; where d ¼ d1 d2 : Thus, l0 -th signal, where 1  k0  N and 1  l0  L0 . Now we form
A ¼ ðdÞC. h the following algorithm for this application.
Theorem 4.8 If A ¼ ðdÞB, then cð AÞ ¼ ðdÞcðBÞ; where
5.1 Algorithm
cð AÞ and cðBÞ are the complement of the complex neu-
trosophic sets A and B.
Step 1. Write the discrete Fourier transforms of the L0
Proof Since signals in the form of complex neutrosophic set,

0   1
max supx2X pCð AÞ ðxÞ  pcðBÞ ðxÞ ; supx2X qcð AÞ ðxÞ  qcðBÞ ð xÞ ; supx2X rcð AÞ ð xÞ  rcðBÞ ð xÞ ;
B 0 1 C
B C
dCNS ðcð AÞ; cðBÞÞ ¼ maxB 2p supx2X lcð AÞ ð xÞ  lcðBÞ ð xÞ ; 2p supx2X mcð AÞ ð xÞ  mcðBÞ ð xÞ ;
1 1
C
@ max@ A A
1
2p sup x2X x cð A Þ ð x Þ  x c ð B Þ ð x Þ
0 1
maxðsupx2X jrA ðxÞ  rB ðxÞj; supx2X jð1  qA ðxÞÞ  ð1  qB ð xÞÞj; supx2X jpA ðxÞ  pB ðxÞjÞ;
B supx2X jð2p  lA ðxÞÞ  ð2p  lB ðxÞÞj; 2p1 supx2X jð2p  mA ð xÞÞ  ð2p  mB ð xÞÞj; C
¼ max@ 1
A
max 2p1
2p sup x2X j ð2p  x A ð x Þ Þ  ð 2p  x B ð x ÞÞ j
0 1
maxðsupx2X jpA ðxÞ  pB ð xÞj; supx2X jqA ðxÞ  qB ðxÞj; supx2X jrA ð xÞ  rB ð xÞjÞ;
B supx2X jlA ð xÞ  lB ð xÞj; 2p1 supx2X jmA ðxÞ  mB ð xÞj; C
¼ max@ 1
A ¼ dCNS ðA; BÞ  1  d
max 2p1
2p supx2X jxA ð xÞ  xB ð xÞj

5 Application of complex neutrosophic set

1 XN   2pjðn1Þðk0 1Þ
in signal processing Sl 0 ð k 0 Þ ¼ : Cl0 ;n ; Dl0 ;n ; El0 ;n :e N ð1Þ
N n¼1
The complex neutrosophic set and d-equalities of complex where Cl0 ;n ; Dl0 ;n ; El0 ;n are the complex-valued Fourier
neutrosophic sets are applied in signal processing appli- coefficients of the signals and 1  n  N.
cation which demonstrates to point out a particular signal
The above sum may be written as
of interest out of a large number of signals that are received
by a digital receiver. This is the example which Ramot 1 XN   jð2pðn1Þðk0 1Þþal0 ;n Þ
Sl 0 ð k 0 Þ ¼ : Ul0 ;n ; Vl0 ;n ; Wl0 ;n :e N ð2Þ
et al. [23] discussed for complex fuzzy set. We now apply N n¼1
complex neutrosophic set to this example.
Suppose that there are L0 different signals, where Cl0 ;n ¼ Ul0 ;n :ejal0 ;n , Dl0 ;n ¼ Vl0 ;n :ejal0 ;n , and El0 ;n ¼
S1 ðtÞ; S2 ðtÞ; . . .; SL0 ðtÞ. These signals have been detected Wl0 ;n :ejal0 ;n with Ul0 ;n , Vl0 ;n , Wl0 ;n 0, and al0 ;n are real-valued
and sampled by a digital receiver and each of which is for all n.

123
1830 Neural Comput & Applic (2017) 28:1817–1834

The purpose of this application is to point out the ref- The similarity between two signals can be measured by
erence signals R0 of the L0 signals. This reference signal this method. By this method, we can find the right signals
R\prime has been also sampled N times ð1  n  N Þ. which have not only uncertain but also indeterminate,
Step 2. Write the Fourier coefficient of R0 in the form of inconsistent, false because when the signals are received by
complex neutrosophic set, a digital receiver, there is a chance for the right signals,
chance for the indeterminate signals, and the chance that
1 XN   j2pðn  1Þðk0  1Þ
R0 ð k 0 Þ ¼ : CR0 ;n ; DR0 ;n ; ER0 ;n :e the signals are not the right one. Thus, by using a complex
N n¼1 N neutrosophic set, we can find the correct reference signals
ð3Þ by taking all the chances, while the complex fuzzy set and
complex intuitionistic fuzzy set cannot find the correct
where CR0 ;n ; DR0 ;n ; ER0 ;n are the complex Fourier coefficients
reference signals if we take all the chances because they
of the reference signals.
are not able to deal with the chance of indeterminacy.
The above expression can be rewritten as
This method can be effectively used for any application
1 XN   jð2pðn  1Þðk0  1Þ þ aR0 ;n Þ in signal analysis in which the chance of indeterminacy is
R0 ðk0 Þ ¼ : UR0 ;n ; VR0 ;n ; WR0 ;n :e
N n¼1 N important.
ð4Þ
where CR0 ;n ¼ UR0 ;n :ejaR0 ;n , DR0 ;n ¼ VR0 ;n :ejaR0 ;n , and ER0 ;n ¼ 6 Drawbacks of the current methods
WR0 ;n :ejaR0 ;n with UR0 ;n , VR0 ;n , WR0 ;n 0, and aR0 ;n are real-
valued for all n. The complex fuzzy sets [23] are used to represents the infor-
Step 3 Since the sum of truth amplitude term, indeter- mation with uncertainty and periodicity simultaneously. The
minate amplitude term, and falsity amplitude term (in the novelty of complex fuzzy sets appears in the phase term with
case when they are crisp numbers, not sets) is not neces- membership term (amplitude term). The main problem with
sarily equal to 1, the normalization is not required and we complex fuzzy set is that it can only handle the problems of
can keep them un-normalized. But if the normalization is uncertainty with periodicity in the form of amplitude term
needed, we can normalize the amplitude terms of Sl0 ðk0 Þ (real-valued membership function) which handle uncertainty
and R0 ðk0 Þ, respectively, as follows: and an additional term called phase term to represent peri-
Ul0 ;n Vl0 ;n odicity, but the complex fuzzy set cannot deal with inconsis-
U ¼ ; V ¼ ; tent, incomplete, indeterminate, false etc. information which
0
l ;n Ul0 ;n þ Vl0 ;n þ Wl0 ;n l0 ;n Ul0 ;n þ Vl0 ;n þ Wl0 ;n
Wl0 ;n UR0 ;n appears in a periodic manner in our real life. For example, in
W ¼ and U ¼ ; quantum mechanics, a wave particle such as an electron can be
0
l ;n Ul0 ;n þ Vl0 ;n þ Wl0 ;n 0
R ;n UR0 ;n þ VR0 ;n þ WR0 ;n
in two different positions at the same time. Thus, the complex
VR0 ;n WR0 ;n
V ¼ ; W ¼ : fuzzy set is not able to deal with this phenomenon.
R0 ;n UR0 ;n þ VR0 ;n þ WR0 ;n R0 ;n UR0 ;n þ VR0 ;n þ WR0 ;n
Complex intuitionistic fuzzy set [1] represents the infor-
mation involving two or more answers of type: yes, no, I do not
Step 4 Calculate the similarity/distances between the
know, I am not sure, and so on, which is happening repeatedly
signals R0 ðk0 Þ and the signals Sl0 ðk0 Þ as follows.
over a period of time. CIFS can represent the information on
people’s decision which happens periodically. In CIFS, the

0 1
B max sup Ul0 ;n  UR0 ;n ; sup Vl0 ;n  VR0 ;n ; sup Wl0 ;n  WR0 ;n ; C
0 0 0 B x2X x2X x2X C
dCNS ðSl0 ðk Þ; R ðk ÞÞ ¼ maxB 0 C
@ 1 ð2pðn  1Þðk  1Þ þ a 0
l ;n Þ ð2pðn  1Þðk0  1Þ þ aR0 ;n Þ A
max sup 
2p x2X N N

Step 5 In order to identify Sl0 as R0 , compare 1  novelty also appears in the phase term but for both membership
dCNS ðSl0 ; R0 ðk0 ÞÞ to a threshold d, where 1  l0  L0 and non-membership functions in some inherent concepts in
 
If 1  dCNS S0l ðk0 Þ; R0 ðk0 Þ exceeds the threshold, iden- contrast to CFS which is only characterized by a membership
tify Sl0 as R0 . function. The complex fuzzy set [23] has only one additional

123
Neural Comput & Applic (2017) 28:1817–1834 1831

phase term, but in CIFS [1], we have two additional phase We now apply all these mentioned methods in the
terms. This confers more range values to represent the uncer- table one by one to show that which method is suitable to
tainty and periodicity semantics simultaneously, and to define describe the situation of above mentioned voting process
the values of belongingness and non-belongingness for any best and what is the failure of the rest of the methods. It is
object in these complex-valued functions. The failure of the clear that the fuzzy set cannot handle this situation because it
CIFS appears in the inconsistent, incomplete, indeterminate only represents the membership 0.4 voters while it fails to tell
information which happening repeatedly. about the non-membership 0.3 and indeterminate member-
The current research (complex fuzzy set) cannot solve this ship 0.3 simultaneously in first voting process. Similar is the
problem because the complex fuzzy set is not able to deal with situation in second voting process. Now when we apply
indeterminate, incomplete, and inconsistent date which is in intuitionistic fuzzy set to both the voting process, it tells us
periodicity. The weaknesses of complex fuzzy set are that it only about the membership 0.4 and non-membership 0.3 in
deals only with uncertainty, but indeterminacy and falsity are first voting process, but cannot tell anything about the 0.3
far away from the scope of complex fuzzy sets. Similarly, the undecided voters in first process. Thus, intuitionistic fuzzy
complex intuitionistic fuzzy set cannot handle the inconsis- set also fails to handle this situation. We now apply neutro-
tent, indeterminate, incomplete data in periodicity simulta- sophic set. The neutrosophic set tells about the membership
neously. Thus, both the approaches are unable to deal with 0.4 voters, non-membership 0.3 voters, and indeterminate
inconsistent, indeterminate, and incomplete data of periodic membership or undecided 0.3 voters in the first round, and
nature. For example, both the methods fail to deal with the similarly, it tells about the second round but neutrosophic set
information which is true and false at the same time or neither cannot describe both the voting process simultaneously. By
true nor false at the same time. applying complex fuzzy set to both the voting process, if we
It is a fundamental fact that some information has not only a set that the amplitude term represents the membership 0.4 in
certain degree of truth, but also a falsity degree as well as first voting process and the phase term represents 0.5 voters
indeterminacy degree that are independent from each other. in second process which form complex-valued membership
This indeterminacy exits both in a subjective and an objective function to represent in both the voting process for an attri-
sense in a periodic nature. What should we do if we have the bute q. But complex fuzzy set remains unsuccessful to
following situation? For instance [16], a 20 temperature describe the non-membership and indeterminacy in both the
means a cool day in summer and a warm day in winter. But if process. The complex intuitionistic fuzzy set only handle
we assume this situation as in the following manner, a 20 complex-valued membership and complex-valued non-
temperature means cool day in summer and a warm day in membership in both the process by setting 0.4 and 0.3 as
winter but neither cool nor warm day in spring. The question is amplitude membership and amplitude non-membership in
that why we ignore this situation? How we can handle it? Why process one and setting 0.5 and 0.3 as phase terms in second
the current methods fail to handle it? We cannot ignore this process. But clearly it fails to identify the indeterminacy
kind of situation of daily life. This phenomenon indicates that (undecidedness) in both the voting process. Finally, by
information is not only of semantic uncertainty and period- applying the complex neutrosophic set to both the voting
icity but also of semantic indeterminacy and periodicity. process by considering the votes in process one as amplitude
terms of membership, non-membership and indeterminate
membership, and setting the second process vote as phase
7 Discussion terms of membership, non-membership, and indeterminacy.
Therefore, the amplitude term of membership in first process
In the Table 1, we showed comparison of different current and the phase term in second process forms complex-valued
approaches to complex neutrosophic sets. In the Table 1, truth membership function. Similarly, the amplitude term of
from 1, we mean that the corresponding method can handle non-membership in process one and the phase term of non-
the uncertain, false, indeterminate, uncertainty with period- membership in second process form complex-valued falsity
icity, falsity with periodicity, and indeterminacy with peri- membership function. Also, the amplitude term of unde-
odicity, while from 0, we mean the corresponding method cidedness in first process and the phase term of indetermi-
fails. It is clear from the Table 1 that how complex neutro- nacy in second process form the complex-valued
sophic sets are dominant over all the current methods. indeterminate membership function. Thus, both the voting
Consider two voting process for some attribute q. In the process forms a complex neutrosophic set as whole which is
first voting process, 0.4 voters say ‘‘yes,’’ 0.3 say ‘‘no,’’ and shown below:
0.3 are undecided. Similarly, in second voting process, 0.5 n
voters say ‘‘yes,’’ 0.3 say ‘‘no’’ and 0.2 are undecided for the S¼ q; TS ðqÞ ¼ 0:4:ej2pð0:5Þ ; IS ðqÞ ¼ 0:3:ej2pð0:3Þ ;
o
same attribute q. These two voting processes held on two FS ðqÞ ¼ 0:3:ej2pð0:2Þ
different dates.

123
1832 Neural Comput & Applic (2017) 28:1817–1834

6 Complex Neutrosophic sets A

given universe of discourse
Complex unit interval
5
Complex Intuitionistic Fuzzy
sets A given universe of
4
discourse Complex unit
interval
Complex Fuzzy sets A given
3
universe of discourse Complex
unit interval
2
Neutrosophic sets A given
universe of discourse Real unit
1 interval
Intuitionistic Fuzzy sets A
0 given universe of discourse
Real unit interval
Fuzzy sets A given universe of
discourse Real unit interval

Fig. 1 Dominancy of complex neutrosophic sets to all other current approaches

Therefore, complex neutrosophic set represent both the amplitude term, and the additional term is called phase
situations in a single set simultaneously, whereas all the term. Thus, in this way, the truth amplitude term represents
other mentioned methods in the table are not able to handle uncertainty and the phase term represents periodicity in the
this situation as whole. uncertainty. Thus, a complex-valued truth membership
The graphical representation in Fig. 1 shows the domi- function represents uncertainty with periodicity as a whole.
nancy of the complex neutrosophic set to all other existing Similarly, complex-valued indeterminate membership
methods. The highest value indicates the ability of the function represents indeterminacy with periodicity and
approach to handle all type of uncertain, incomplete, complex-valued falsehood membership function represents
inconsistent, imprecise information or data in our real-life falsity with periodicity. Further, we presented an interpre-
problems. Each value on the left vertical line shows the tation of complex neutrosophic set and also discussed some
value of the ability of the corresponding method on the of the basic set theoretic properties such as complement,
horizontal line in the graph. union, intersection, complex neutrosophic product, Carte-
sian product in this paper. We also presented d-equalities
of complex neutrosophic set and then using these d-
8 Conclusion equalities in the application of signal processing. Draw-
backs of the current methods are discussed and a com-
An extended form of complex fuzzy set and complex parison of all these methods to complex neutrosophic sets
intuitionistic fuzzy set is presented in this paper, so-called was presented in this paper.
complex neutrosophic set. Complex neutrosophic set can This paper is an introductory paper of complex neutro-
handle the redundant nature of uncertainty, incomplete- sophic sets, and indeed, much research is still needed for
ness, indeterminacy, inconsistency, etc. A complex neu- the full comprehension of complex neutrosophic sets. The
trosophic set is defined by a complex-valued truth complex neutrosophic set presented in this paper is an
membership function, complex-valued indeterminate entire general concept which is not limited to a specific
membership function, and a complex-valued falsehood application.
membership function. Therefore, a complex-valued truth Acknowledgments We are very thankful to Prof. Dr. Jie Lu of
membership function is a combination of traditional truth University of Technology Sydney, UTS Australia, for her valuable
membership function with the addition of an extra term. comments and suggestions. We are also very thankful to the reviewers
The traditional truth membership function is called truth for their comments and suggestions which improved this work.

123
Neural Comput & Applic (2017) 28:1817–1834 1833

Appendix

Comparison of complex neutrosophic sets to fuzzy sets, intu-

itionistic fuzzy sets, neutrosophic sets, complex fuzzy sets, and
complex intuitionistic fuzzy sets is listed below (Table 1).

Table 1 Comparison of complex neutrosophic sets to the current approaches

Sets/logics Domain Co-domain Uncertainty Falsity Indeterminacy Uncertainty Falsity with Indeterminacy
with periodicity with periodicity
periodicity

Fuzzy sets A given Real unit 1 0 0 0 0 0

universe of interval
discourse
Intuitionistic A given Real unit 1 1 0 0 0 0
fuzzy sets universe of interval
discourse
Neutrosophic sets A given Real unit 1 1 1 0 0 0
universe of interval
discourse
Complex fuzzy A given Complex 1 0 0 1 0 0
sets universe of unit
discourse interval
Complex A given Complex 1 1 0 1 1 0
intuitionistic universe of unit
fuzzy sets discourse interval
Complex A given Complex 1 1 1 1 1 1
neutrosophic universe of unit
sets discourse interval

References 12. Guo Y, Cheng DH (2009) New neutrosophic approach to image

segmentation. Pattern Recognit 42:587–595
1. Alkouri A and Salleh A (2012) Complex intuitionistic fuzzy sets. 13. Hanafy MI, Salama AA, Mahfouz K (2012) Correlation of neu-
In: International conference on fundamental and applied sciences, trosophic data. Int Refereed J Eng Sci 1(2):39–43
14. Hanafy K, Salama AA, Mahfouz MK (2013) Correlation coeffi-
AIP conference proceedings, vol 1482, pp 464–470
2. Atanassov TK (1986) Intuitionistic fuzzy sets. Fuzzy Sets Syst cients of neutrosophic sets by centroid method. Int J Probab Stat
20:87–96 2(1):9–12
3. Broumi S, Smarandache F (2013) Correlation coefficient of 15. Hong HD, Hwang YS (1994) A note on the value similarity of
fuzzy systems variables. Fuzzy Sets Syst 66(3):383–386
interval neutrosophic set. Appl Mech Mater 436:511–517
4. Buckley JJ (1987) Fuzzy complex numbers. In: Proceedings of 16. Jun M, Zhang G, Lu J (2012) A method for multiple periodic
ISFK, Guangzhou, China, pp 597–700 factor prediction problems using complex fuzzy sets. IEEE Trans
5. Buckley JJ (1989) Fuzzy complex numbers. Fuzzy Sets Syst Fuzzy Syst 20(1):32–45
33(3):333–345 17. Lin L, Yuan XH, Xia QZ (2007) Multicriteria fuzzy decision-
6. Buckley JJ (1991) Fuzzy complex analysis I: definition. Fuzzy making methods based on intuitionistic fuzzy sets. J Comput Syst
Sets Syst 41(2):269–284 Sci 73(1):84–88
7. Buckley JJ (1992) Fuzzy complex analysis II: integration. Fuzzy 18. Lupiáñez GF (2008) On neutrosophic topology. Kybernetes
Sets Syst 49(2):171–179 37(6):797–800
8. Cai YK (1995) d-Equalities of fuzzy sets. Fuzzy Sets Syst 19. Mendel MJ (1995) Fuzzy logic systems for engineering: a tuto-
76(1):97–112 rial. Proc IEEE 83:345–377
20. Majumdar P, Samanta KS (2014) On similarity and entropy of
9. Cai YK (2001) Robustness of fuzzy reasoning and delta-equali-
ties of fuzzy sets. IEEE Trans Fuzzy Syst 9(5):738–750 neutrosophic sets. J Intell Fuzzy Syst 26(3):1245–1252
10. Chen Z, Aghakhani S, Man J, Dick S (2011) ANCFIS: a neuro- 21. Nguyen TH, Kandel A, Kreinovich V (2000) Complex fuzzy sets:
fuzzy architecture employing complex fuzzy sets. IEEE Trans towards new foundations. In: The ninth IEEE international con-
Fuzzy Syst 19(2):305–322 ference on fuzzy systems, vol 2, pp 1045–1048. doi:10.1109/
11. Deschrijive G, Kerre EE (2007) On the position of intuitionistic FUZZY.2000.839195
fuzzy set theory in the framework of theories modelling impre- 22. Pappis PC (1991) Value approximation of fuzzy systems vari-
cision. Inf Sci 177(8):1860–1866 ables. Fuzzy Sets Syst 39(1):111–115

123
1834 Neural Comput & Applic (2017) 28:1817–1834

23. Ramot D, Milo R, Friedman M, Kandel A (2002) Complex fuzzy 34. Ye J (2013) Multicriteria decision-making method using the
sets. IEEE Trans Fuzzy Syst 10(2):171–186 correlation coefficient under single-valued neutrosophic envi-
24. Ramot D, Friedman M, Langholz G, Kandel A (2003) Complex ronment. Int J Gen Syst 42(4):386–394
fuzzy logic. IEEE Trans Fuzzy Syst 11(4):450–461 35. Ye J (2013) Similarity measures between interval neutrosophic
25. Salama AA, Alblowi AS (2012) Generalized neutrosophic set and sets and their applications in multicriteria decision making. J In-
generalized neutrosophic topological spaces. J Comput Sci Eng tell Fuzzy Syst 26(1):165–172
2(7):29–32 36. Ye J (2014) Single valued neutrosophic cross-entropy for multi-
26. Salama AA, Alblowi AS (2012) Neutrosophic set and neutro- criteria decision making problems. Appl Math Model
sophic topological space. ISORJ Math 3(4):31–35 38(3):1170–1175
27. Salama AA, El-Ghareeb AH, Manie AM, Smarandache F (2014) 37. Ye J (2014) Vector similarity measures of simplified neutro-
Introduction to develop some software programs for dealing with sophic sets and their application in multicriteria decision making.
neutrosophic sets. Neutrosophic Sets Syst 3:53–54 Int J Fuzzy Syst 16(2):204–211
28. Smarandache F (1999) A unifying field in logics neutrosophy: 38. Zadeh AL (1965) Fuzzy sets. Inf Control 8:338–353
neutrosophic probability, set and logic. American Research Press, 39. Zhang G, Dillon ST, Cai YK, Ma J, Lu J (2009) Operation
Rehoboth properties and d-equalities of complex fuzzy sets. Int J Approx
29. Szmidt E, Kacprzyk J (2004) Medical diagnostic reasoning using Reason 50:1227–1249
a similarity measure for intuitionistic fuzzy sets. Note IFS 40. Zhang M, Zhang L, Cheng DH (2010) A neutrosophic approach
10(4):61–69 to image segmentation based on watershed method. Signal Pro-
30. Turksen I (1986) Interval valued fuzzy sets based on normal cess 90(5):1510–1517
forms. Fuzzy Sets Syst 20:191–210 41. Zhang QZ (1992) Fuzzy limit theory of fuzzy complex numbers.
31. Vlachos KL, Sergiadis DG (2007) Intuitionistic fuzzy informa- Fuzzy Sets Syst 46(2):227–235
tion—applications to pattern recognition. Pattern Recognit Lett 42. Zhang QZ (1992) Fuzzy distance and limit of fuzzy numbers.
28(2):197–206 Fuzzy Syst Math 6(1):21–28
32. Wang H, Smarandache F, Zhang QY, Sunderraman R (2005) 43. Zhang QZ (1991) Fuzzy continuous function and its properties.
Interval neutrosophic sets and logic: theory and applications in Fuzzy Sets Syst 43(2):159–175
computing. Hexis, Phoenix 44. Zimmermann AL (1991) Fuzzy set theory—and its applications,
33. Wang H, Smarandache F, Zhang QY, Sunderraman R (2010) 2nd edn. Kluwer, Boston
Single valued neutrosophic sets. Multispace Multistruct
4:410–413

123