Accepted Manuscript

Three-way decisions based on neutrosophic sets and AHP-QFD

framework for supplier selection problem

Mohamed Abdel-Basset, M. Gunasekaran, Mai Mohamed, Naveen Chilamkurti

PII: S0167-739X(18)31085-9
DOI: https://doi.org/10.1016/j.future.2018.06.024
Reference: FUTURE 4286

To appear in: Future Generation Computer Systems

Received date : 7 May 2018

Revised date : 25 May 2018
Accepted date : 10 June 2018

Please cite this article as: M. Abdel-Basset, M. Gunasekaran, M. Mohamed, N. Chilamkurti,

Three-way decisions based on neutrosophic sets and AHP-QFD framework for supplier selection
problem, Future Generation Computer Systems (2018),
https://doi.org/10.1016/j.future.2018.06.024

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form.
Please note that during the production process errors may be discovered which could affect the
content, and all legal disclaimers that apply to the journal pertain.
Three-Way Decisions Based on Neutrosophic Sets and AHP-
QFD Framework for Supplier Selection Problem

Mohamed Abdel-Basset1, M. Gunasekaran2 ,Mai Mohamed1 and Naveen Chilamkurti 3

1
Department of Operations Research, Faculty of Computers and Informatics, Zagazig University,
Sharqiyah, Egypt.
1
E-mail: analyst_mohamed@yahoo.com; E-mail: mmgaafar@zu.edu.eg
2
University of California, Davis, United States
E-mail: gmanogaran@ucdavis.edu
3
Department of Computer Science and Computer Engineering, LaTrobe University,
Melbourne, Australia
E-mail: n.chilamkurti@latrobe.edu.au

ABSTRACT
The neutrosophic set is an excellent tool for dealing with vague and inconsistent
information effectively. Consequently, by studying the concept of three-way decisions
based on neutrosophic set, we can find a suitable manner to take a reasonable
decision. In this article, we suggest two rules of three-way decisions based on three
membership degrees of neutrosophic set. A new evaluation function is presented to
calculate weights of alternatives, for choosing the best one. We also study a supplier
selection problem (selecting suppliers to obtain the indispensable materials for
assisting the outputs of companies). The best suppliers need to be selected to enhance
quality, service, to reduce cost, and to control time. The most widely used technique
for determining the requirements of a company is the Quality Function Deployment
(QFD). Since traditional QFD technique does not prioritize stakeholders' requirements
and fails to deal with vague and inconsistent information, this research also integrates
it with Analytic Hierarchy Process (AHP) depending on neutrosophic environment. A
case study is presented to illustrate the effectiveness of the proposed model.
KEYWORDS Three-Way Decisions; Neutrosophic Analytic Hierarchy Process;
Quality Function Deployment; Neutrosophic Set; Supplier Selection; Stakeholders.
1. Introduction
The three-way decision theory is proposed based on the classifications of a finite set
of objects into three parts, called positive, negative and boundary regions. These
regions are viewed as acceptance region, rejection region and not sure region.
According to these regions, one can determine rules for acceptance, rejection and
uncertainty. The three-way decisions are used in real life fields, such as management
[1,2], medical and engineering decision making [3-5], social judgments [6], or
statistics, especially in testing hypothesis [7]. The applications of three-way decisions
were discussed in different researches [8-36]. There are various models of three-way
decisions; all of them being constructed to deal with uncertainty. Pawlak [37] in 1982
propounded the theory of rough set, where the lower and upper approximations are
presented by dividing a finite set of objects into positive, negative and boundary
regions. An object belongs to the positive region of the set when its equivalence class
is totally included in the set, and this makes positive region little. To overcome this
problem, a generalization of Pawlak rough sets [37-42] has been produced. This
generalization is decision-theoretic rough sets. Three-way decision theory is used to
illustrate the three regions of rough sets [43-45]. Zadeh [46] introduced fuzzy set in
1965, illustrating, by its membership degree, the inclusion of objects to a set. The
membership degree is determined as a real number in unit interval. Although fuzzy set
is employed in many fields, it fails to represent the boundary and negative information
of knowledge. To overcome this drawback of fuzzy set, Atanassov [47] propounded
the concept of intuitionistic fuzzy set. In intuitionistic fuzzy set, two membership
degrees were used to represent uncertain information. Smarandache [48], in 1995,
introduced the neutrosophic sets, in which three membership degrees were used to
represent uncertain and inconsistent information of knowledge. These membership
degrees are truth-membership (acceptance), falsity-membership (rejection) and
indeterminacy-membership (uncertainity) degrees, their summation having no
constrains. Thus neutrosophic set represent the positive, negative and indeterminate
information efficiently and more flexible than fuzzy and intuitionistic fuzzy sets. The
truth-membership degree, the falsity-membership degree and the indeterminacy
membership degree in neutrosophic set are compatible with the positive, negative and
boundary regions in three-way decisions. It means that the neutrosophic set is the best
model for representing three-way decisions, also for handling uncertain, inconsistent
information, and can be combined for enhancing decision making process. For real
life applications, such as engineering and scientific applications, Smarandache [48]
and Wang et. al. [49] proposed a case of neutrosophic set, called the single valued
neutrosophic set. Here, the range of truth-membership, falsity-membership and
indeterminacy-membership degree is between unit interval [0, 1]. Single valued
neutrosophic set can be considered as a generalization of classical set, fuzzy set and
intuitionistic fuzzy set. Another important part of this research is to study supplier's
selection problem in neutrosophic environment. The Suppliers Selection process (SS),
performed by the purchasing department of a business company, is one of the most
stringent tasks. Nowadays, business companies are strongly dependent on their
suppliers, so the estimation and the buying processes of suppliers are very important.
Due to the increased competition between regional markets and globalization, the
numbers of suppliers and criteria to be taken into consideration in the selection
processes of suppliers have extremely grown. Therefore, the accommodation of any
service or manufacturing company depends on an efficient and effective process of
supplier selection. The selection process of suppliers can be considered as a multi-
criteria decision making problem, and it contains both qualitative and quantitative
criteria. It is very important to make a trade-off between the palpable and impalpable
criteria, which are sometimes in conflict, in order to select the best suppliers. The
most important criteria for supplier's selection are price, delivery, quality, cost and
location [50]. There are two parts of supplier's selection process:
 Criteria determination process of supplier.
 Performance measure process of supplier.
The selected supplier should meet company requirements, so in this research we use
the Quality Function Deployment (QFD) to determine effectively company
requirements. The traditional QFD cannot prioritize the requirements, consequently
we integrated it with AHP for better prioritizing the process of requirements. Because
company requirements are always conflicting, vague and inconsistent in nature, the
integrated AHP-QFD framework is analyzed by employing the neutrosophic set.
1.1 Research contribution
We proposed a novel model of three-way decisions based on neutrosophic set. Many
researchers have combined three-way decisions with fuzzy and intuitionistic fuzzy
sets, but it is not the ideal process for representing three-way decisions, since the
fuzzy set possess only a membership function, corresponding to the positive region in
the three-way decisions, though ignoring the boundary region and the negative region.
Although the intuitionistic fuzzy set represents both positive and negative regions of
three-way decisions, still the neutrosophic set asseverates the ideal representation,
since it contains three membership degrees, namely the truth, the indeterminacy an the
falsity, corresponding to the positive, boundary and negative regions of three-way
decisions. In many cases, in real life applications, the indeterminacy exists in large
percentage, and, in other cases, its value does not depend on the acceptance or
rejection of things.
In this research, we apply for the first time the three-way decisions on neutrosophic
set, and implement it in incomplete information systems. The second part of our
research illustrates the efficient and effective rule of neutrosophic set in supplier
selection process. In order to introduce a practical solution of supplier selection
problem, an integration of neutrosophic analytic hierarchy process with quality
function deployment is presented. It is the first time to introduce the integration of
AHP-QFD in neutrosophic environment. Since the criteria for supplier selection
process are always vague, we overcome this drawback by employing the neutrosophic
set, which deals efficiently with inconsistent information, by taking into account all
aspects of the decision making process. In addition, the scale of traditional quality
function deployment has two drawbacks:
 It cannot prioritize company requirements.
 The group of experts should be in consensus on suggested weights, and that is
a very difficult process in real situations.
Our proposed framework for supplier selection problem solved the previous
drawbacks by introducing the integrated AHP-QFD in neutrosophic environment.
Also, the scale of traditional QFD is enhanced, becoming more accurate by using
neutrosophic AHP to calculate it. The proposed framework is simple and accurate, it
can be applied to a wide range of problems, it can help companies to succeed and
achieve customer satisfaction, as well estimating different alternatives.
The structure of this research is as follows: In section 2, the literature review on
different techniques used in supplier's selection process is covered. In section 3, the
three-way decisions is examined based on neutrosophic set. In section 4, a detailed
explanation of our proposed framework for supplier selection problem is presented. In
section 5, a real case study is parsed to validate the applicability of our framework,
and a comparative analysis between different multi-criteria decision making
approaches with QFD is presented. Section 6 concludes the importance of our
research and discusses future directions of study.
2. Literature review
In previous section, we reviewed the three-way decision theory, its applications and
models. Further, we focus on supplier selection problem. More than 63% of supplier
selection researches are found in multi-criteria environment [51]. A purchasing
decision framework and a discussion of complexity and importance included in
purchasing decisions were presented by De Boer et al. [52]. Lee et al. [53] proposed a
supplier selection and management system to ensure the effectiveness of supplier
selection management. William Ho et al. [54] claim that DEA is the most general
technique for supplier selection problem and AHP-GP is the most widespread
integrated technique to solve supplier selection problem. They also noted that the
most common criteria for supplier selection are not price or cost, but rather quality
and delivery. A fuzzy suitability index was proposed by Bevilacqua et al. [55] to rank
pertinent criteria of supplier selection. To select the best suppliers in supply chain, Ha
and Krishnan [56] propounded an integrated technique of AHP, DEA and NN. Based
on the Case Based Reasoning technique in fuzzy surrounding and mathematical
programming, Faez et al. [60] developed an integrated model for selection process of
item vendor. To deal with numerous and conflicting criteria to rank and select best
suppliers of value chain framework, Bhattacharya et al. [61] integrated AHP with
QFD and cost factor measure (CFM). The structural equation modeling is employed
by Punniyamoorthy et al. [62] to calculate weight of criteria and fuzzy AHP, used to
calculate weight of alternatives corresponding to each criterion for supplier's selection
problem. A three multi-criteria decision making technique was proposed by Mihat
Zeydan et al. [63] for determining efficiency of suppliers. The first technique was
fuzzy AHP and used for determining weight of criteria; another technique consisted in
transforming qualitative values to one quantitative value; and finally DEA, used for
ranking of suppliers. The QFD and data mining method have been used by Ni et al.
[64] for selecting and ranking suppliers. Bevilacque et al. [65] made use of fuzzy
numbers to QFD for supplier's selection problem. The analytic network process was
put forth by Gencer and Gurpinar for supplier's selection problem [66]. Gencer and
Gurpinar also availed of fuzzy QFD for selection process of internet service provider
[67]. A case study of a washing machine company in Turkey was discussed by
Kilincci and Onal using fuzzy AHP [68]. Zouggari and Benyoucef published a model
for supplier selection problem depending on fuzzy TOPSIS [69]. The integration of
QFD and fuzzy AHP was tested by Bojana and Boris for selecting suppliers of
electronic components [70]. There is no integration of QFD and AHP in neutrosophic
environment, consequently it is the first application of neutrosophic integration of
QFD and AHP framework in suppliers selection problem.
2.1 Gaps of previous researches
The gaps of previous theories used in three-way decisions were presented in the
previous section. Therefore, we list here only the gaps of techniques used in suppliers
selection problem. The previous researches failed to deal effectively and efficiently
with vague and inconsistent information usually existing in real world situations.The
most popular technique that have been used in suppliers selection problem fails to
take into considerations the requirements of company's stakeholders and it ignores
ways to achieve these requirements. The scale of some techniques is not accurate,
such as traditional QFD scale.
3. Three-way decisions based on neutrosophic sets
In this section, the three-way decision theory is studied based on neutrosophic sets.
The basic idea of the three-way decision is to assess the universe into positive,
negative and boundary areas, by using an evaluation function and thresholds [68]. As
we annunciated above, in neutrosophic sets the truth-membership, the falsity-
membership and the indeterminacy-membership degrees correspond to the positive,
negative and boundary regions in three-way decisions, respectively. It means that the
neutrosophic sets and three-way decisions could be integrated for enhancing their
performance in decision-making process.
3.1 Neutrosophic sets
In this sub-section, the important definitions of neutrosophic sets are clearly stated
[69].
Definition 1. Let be a space of points and ∈ . A neutrosophic set in is
recognized by a truth-membership function ( ), an indeterminacy-membership
function and a falsity-membership function , where ( ), and
are subsets of ]-0, 1+[. ( ): →]-0, 1+[, : →]-0, 1+[ and : →]-0,
1+[. There is no restriction on summation of membership functions. Therefore, 0− ≤
sup ( ) + sup + sup ≤3+.
Definition 2. A single valued neutrosophic set over is an object taking the form
={〈 , ( ), , 〉: ∈ }, where ( ): →[0,1], : →[0,1] and
: →[0,1] with 0≤ ( )+ + ≤3 for all ∈ . For convenience, a SVN
number is represented by = ( , b, c), where , , ∈ [0, 1] and + + ≤3.
Definition 3. A single valued triangular neutrosophic number,
 =〈 , , ; , , 〉 is a neutrosophic set on the real line set ,
with , , , , and , , ∈ [0, 1]. The truth-membership, the
indeterminacy-membership and the falsity-membership functions of this neutrosophic
number are defined as:

(1)

0 otherwise ,

1 2

2 3 2

1 otherwise ,



3


1 otherwise .
where , and represent the maximum degree of truth-membership, the
minimum degree of indeterminacy and the falsity-memberships, respectively. A
single valued triangular neutrosophic number =〈 , , ; , , 〉 may express
an indeterminate quantity about ,which is approximately equal to .
Definition 4. Let =〈 , , ; , , 〉 and =〈 , , ; , , 〉 be two
single valued triangular neutrosophic numbers and  0 be any real number. Then,
 Addition of two triangular neutrosophic numbers
+ =〈 , , ; ˄ , ˅ , ˅ 〉

 Subtraction of two triangular neutrosophic numbers

=〈 , , ; ˄ , ˅ , ˅ 〉

 Inverse of a triangular neutrosophic number

〈 , , ; , , 〉,Where 0
 Multiplication of triangular neutrosophic number by constant value

〈  , , ; , , 〉  0
 =
〈  , , ; , , 〉  0
 Division of triangular neutrosophic number by constant value

1 2 3
〈 , , ; , , 〉  0
  
 3 2 1
〈 , , ; , , 〉  0
  

 Division of two triangular neutrosophic numbers

〈 , , ; ˄ , ˅ , ˅ 〉 0, 0

〈 , , ; ˄ , ˅ , ˅ 〉 0, 0

〈 , , ; ˄ , ˅ , ˅ 〉 0, 0

 Multiplication of two triangular neutrosophic numbers

〈 , , ; ˄ , ˅ , ˅ 〉 0, 0
〈 , , ; ˄ , ˅ , ˅ 〉 0, 0
〈 , , ; ˄ , ˅ , ˅ 〉 0, 0

3.2 Neutrosophic sets in incomplete information systems

A system with missing values is called an incomplete information system. Let
, be an incomplete information system, where is a universe of discourse
containing a finite set of objects and a set of attributes such that : → where
∈ and is the value set of . In some cases, the values of some attributes are
missing. When at least one attribute exists, ∈ has a null value of and is an
incomplete information system. We symbolize null value by *. To determine the
binary similarity relation of a subset of attributes R ⊆ , use the following equation
[70]:
SIM = , ∈ │∀r ∈ R, r a r b OR r a ∗ OR r b ∗ (4)
Then, the objects that are similar to regarding to are denoted by the following:
∈ │ , ∈ SIM (5)
For ⊆ , ⊆ and using the similarity relation [70], the lower and upper
approximations of are defined as follows:
∈ │ ⊆ ∈ │ ⊆ (6)
where is the set of objects contained in with certainty.
 ∈ │ ∩ ∅ ∪ │ ∈ (7)
where is the set of objects that have the possibility to be contained in , ~ is
the impossibility of belonging a set of objects to , and the boundary set BR is the
set of objects that might be or might be not contained in , denoted by BR =
.
The single valued neutrosophic set in incomplete system denoted by
A*= 〈 , , , 〉│ ∈ , (8)
where, represents the truth-membership degree of a ∈ , stands for
the falsity-membership degree of a∉ , and represents the indeterminacy
membership degree of a ∈ BR which implies that may or may not be in . Thus,
by given an object , we can detect all its degrees of truth, falsity and indeterminacy
membership functions, and obtain the neutrosophic sets.
We can conclude the law of truth-membership degree from the previous steps as
follows:
is the truth-membership of object , and it means the set of objects contained
in with certainty and similar to with respect to . It can be written as
∩
│ | |
(9)

where | | stands for cardinality of set .

The falsity-membership of object
is similar to with respect to
to . It can be written as:
|~
~ │
is determined through giving a set of objects that
and with impossibility of belonging these objects
∩ |
| |
, (10)

According to definition 2, : → 0,1 , : → 0,1 and

+ 3, then A* is a single valued neutrosophic set on . Since in
neutrosophic set the value of indeterminacy does not depend on the value of truth and
falsity, the value of is detected via expert's opinions.

3.3 Evaluation functions and decision rules for three-way decisions

based on neutrosophic sets
Since neutrosophic sets have three degrees, we construct the rules of three-way
decision based on these degrees. If the truth-membership degree of an
object belonging to neutrosophic set is large and the falsity-membership degree
of an object belonging to neutrosophic set is small, then the decision belongs to
positive areas. If the truth-membership degree of an object belonging to
neutrosophic set is small and the falsity-membership degree of an object
belonging to neutrosophic set is large, then the decision belongs to negative areas.
If the truth-membership degree and falsity-membership of an object are small, then
the indeterminacy-membership degree is large and we decide that belongs to the
boundary areas. We can construct rules of three-way decision by comparing the three
neutrosophic degrees as follows:
Rule 1. If , then ∈ positive areas;
If , then ∈ negative areas;
Otherwise, then ∈ boundary areas.
Since the evaluation functions for three-way decisions in theoretic rough set depend
on condition probabilities, a pair of threshold ( , ) should be selected. Here, in
evaluation function for three-way decisions based on neutrosophic sets, the two-
thresholds should be chosen( , ). Let, 0 1, 1 0 and for a convenient
representation we suggest that 0.5 and 0.5 .
Rule 2. If , then ∈ positive areas;
If , then ∈ negative areas;
If , then ∈ boundary areas.
In this research, instead of condition probabilities of evolution function, we will use
the membership degrees as in rule 1.
Example 1. There are three alternatives of cars , , and a set of attributes such
as: price, speed, comfortable and engine power. According to these attributes, there
exist three classes of decisions as follows: good car, poor car and
excellent car. By giving the membership degrees of car attributes according to
each decision as in Table 1, we can classify each car according to its class.

Cars
( )

1 0 0 0.1 0.1 0.9 0.15 0.10 0.85

1 0 0 0 0 1 0.1 0.1 0.9

 0.3 0.75 0.7 0.4 0.3 0.6 0.4 0.65 0.6

Table 1. The degrees of truth-membership, indeterminacy-membership and falsity-membership.

By using rule 1, fFor :

1
1 1
, 1 1
, Decide ∈ positive ( );
1 1 1 1

1
0.9 1
, 1 1
, Decide ∈ negative ( );
2 2 2 2

1
0.85 1
, 1 1
, Decide ∈ negative ( ).
3 3 3 3

Therefore, is a good car, not a poor car, nor an excellent car.

For , we decide is a good car, not a poor, nor an excellent, for the same rule.
For ,

3
0.75 3
, 3 3
, Decide ∈ boundary ( );
1 1 1 1

3
0.6 3
, 3 3
, Decide ∈ negative( );
2 2 2 2

3
0.65 3
, 3 3
, Decide ∈ boundary ( ).
3 3 3 3

Therefore, we are not certain that is a good car, but we are certain that is not a
poor car. Also, it is not certain that is an excellent car.
Until now, we only assigned objects to its class, by dividing a universe of discourse
into three areas, namely positive, negative, boundary areas, corresponding to truth-
membership, falsity-membership and indeterminacy membership degrees respectively
in neutrosophic sets. What about obtaining the optimal decision?
Combining the concept of neutrosophic set to three-way decisions, we can determine
the three areas plus the optimal final decision (i.e. in the previous example we classify
each car to its decision state, but we did not say that is the best car for our
selection). Also in the previous example, there are two cars with the same
classification value (i.e. , , are good cars, not poor cars and also not excellent
cars), and this shows that the final decision of car selection is not accurate when we
want to select only one car (i.e. we need to determine the weight for good cars to
select only one). In order to extend the definition of three-way decisions by using
neutrosophic sets and obtain a final decision effectively and efficiently, we accredit a
value to each attribute. This value is a triangular neutrosophic number. Beside this
value, the three areas are defined according to this attribute. To calculate the value of
each decision attribute, we use the following evaluation function:
Let =〈 , , 〉 be a triangular neutrosophic number with three-way
decisions , , , which correspond to truth-membership, indeterminacy-
membership and falsity-membership degrees, then the evaluation function of this
number according to three-way decisions is as follows:

3 (11)

where , , is the lower, median and upper value of the triangular number. In
evaluation function, | | means the absolute value of . By using the evaluation
function of three-way decisions based on neutrosophic environment, we can obtain a
specific, accurate and effective decision. Therefore, the neutrosophic set is an efficient
and effective model of three-way decisions by considering all aspects of uncertainty,
vague and inconsistent information. The optimal final decision is also reasonably
obtained.
If we return to the first example and intend to apply the evaluation function in order
determine the crisp value of each class of decision, then see the next example:
Example 2. As in Table 1, the degree of membership functions of car attributes
according to each decision is presented, then let us determine the triangular value of
each car attribute according to each decision as in table 2.
Cars
( )

6 7 8 2 3 4 1 2 3
2 3 4 6 7 8 1 1 1
1 1 1 2 3 4 1 1 1
Table 2. The triangular value of each car attribute according to each decision class.

, stands for the lower bound of car attribute according to decision class,

denotes the median value, and designates the upper value.

From Table 1, we found that:

is a good car , not a poor car, nor an excellent car, and also is a good car , not a
poor car and also not an excellent car. We are not certain that is a good car, but we
are certain that is not a poor car. Moreover, it is not certain that is an excellent
car. In order to select only the best car between and , we calculate their weight
using Eq. 11, as follows:
Evaluation function of according to by using Table 1 and Table 2 values:
 6 7 8 1 0 0 8,
3

Evaluation function of according to by using Table 1 and Table 2 values:

2 3 4 0.1 0.1 0.9 5.4,

3

Evaluation function of according to by using Table 1 and Table 2 values:

1 2 3 0.15 0.1 0.85 1.2,

3

After then, we make a normalization process to calculate the weights of each decision
class:
Weight of according to ;
8
0.55,
14.6
Weight of according to ;
5.4
0.36,
14.6
Weight of according to ;
1.2
0.08
14.6
The same steps for C .
Evaluation function of according to by using Table 1 and Table 2 values:

2 3 4 1 0 0 4,
3

Evaluation function of according to by using Table 1 and Table 2 values:

6 7 8 0 0 1 6,
3

Evaluation function of according to by using Table 1 and Table 2 values:

1 1 1 0.1 0.1 0.9 0.1.

3

After then, we make a normalization process to calculate the weights of each decision
class:
Weight of according to ;
4
0.39,
10.1
Weight of according to ;
6
0.59,
10.1
Weight of according to ;
0.1
0.009,
10.1
Because , are good cars, but we want to select only one of them, we analyze the
weights values. Since 0.55 and 0.39, then the first car, , is
best.
4. The neutrosophic QFD, AHP and proposed methodology
In the previous section, we introduced the three-way decisions in incomplete
information system, and used it in classification process of objects. We also
introduced an evaluation function if more than one object has the same decision class.
We noted that the three regions of the three-way decisions can be considered as a
reflection of truth, indeterminacy and falsity membership degrees of neutrosophic set.
In this section, the proposed methodology for solving supplier selection problem in
neutrosophic environment is presented with details.
4.1 Neutrosophic QFD
Yoji Ako, proposed the QFD in 1970s, as tool for developing quality [71]. It is a very
important technique for considering the stakeholders' requirements. A QFD
transforms the quality requirements for stakeholders or customers of firms, suppliers
and employees to quality characteristics. A QFD is named the customers' voice
technique. Nowadays, QFD has a widespread usage in various fields. The traditional
components of QFD are as follows (see also Figure 1):
 The central part of the figure is the house of quality (HOQ), which shows the
relationship between quality requirements (WHATs) and quality
characteristics (HOWs).
 The left part of HOQ consists of the quality requirements (WHATs) or
organizations' requirements and weights of WHATs.
 The top of HOQ reflects the quality characteristics (HOWs) or how to achieve
organizations' requirements, and the correlation between WHATs and HOWs.
 The right part of HOQ includes a matrix of competitors.
 The bottom part shows the degree of weights of HOWs, target and evaluation
process.
 Correlation
Matrix

Quality characteristics
(HOWs)

Quality Weights of Relationship matrix Matrix of

requirements quality competitors
requirements (HOQ)
(WHATs)

Weights of
Quality characteristics

Target

Evaluation
`
Figure 1. The framework of quality function deployment.
The determination process of quality requirements weights is a critical step in QFD,
since a group of experts asks to enter a scale system either 1-5 or 1-9 of each quality
requirement criteria. The previous scale has some drawbacks, because it fails to
prioritize quality requirements, considered as a subjective value, and all experts
should be in consensus, which is not an easy matter. The previous scale also fails to
deal with vague and inconsistent information, which exist, usually, in real world
situations. To treat the previous drawbacks, we apply QFD in neutrosophic
environment for better dealing with vague information, and we integrate the QFD
with neutrosophic AHP to calculate weights of quality requirements. We illustrate
neutrosophic AHP with details in the next sub-section. We treat the relationship
matrix values of QFD as neutrosophic numbers, since strong relationship takes value
9 , moderate relationship takes value 3 and weak relationship takes value 1. The
correlation matrix of the QFD contains these symbols: ++, which means a strong
positive correlation between HOWs, +, denoting a positive correlation, –, designing a
negative correlation and – –, which means a strong negative correlation. The weight
of each quality characteristic (HOWs) is calculated through multiplying each
relationship value by its corresponding weight of quality requirement (WHATs), and
the mathematical representation of weight for HOWs is as follows:
(HOW) = the relationship value of (HOW)i1 * (WHAT1)+…+the relationship
value of (HOW)in*w(WHATn) (12)
All neutrosophic numbers used in this research are triangular neutrosophic numbers:
4.2 Neutrosophic AHP
The Analytic Hierarchy Process (AHP) was founded by Saaty in the late 1970s. It is a
vastly used method in diverse multi-criteria decision analysis problems. The AHP can
deal with qualitative and quantitative criteria. The scale in AHP is called the nine-
point scale [72]. The AHP dissects complex problems into simple and easy sub-
problems. The classical AHP does not deal with vague information; the fuzzy theory
was embedded to the classical AHP. Since the back stone of fuzzy programming is
the membership function and decision makers assumed it according to their
experience, it fails to deal with indeterminacy function which exists usually in real life
situations and affect the quality of decisions. The scale in Fuzzy AHP cannot reflect
accurately the perceptions of the decision maker. In order to solve the previous
drawbacks, we apply AHP in neutrosophic environment to deal with truth, falsity and
indeterminacy degrees, and then reflect the perceptions of the decision maker
accurately and efficiently. The steps are as follows:
 Dissect a complex problem into sub-problems, easier to solve.
 Construct the hierarchy of the problem, where the first level is the goal, second
level stands for criteria and sub-criteria, and in the final level of hierarchy place
the alternatives.
 Let experts construct the decision matrix by depending on nine-point scale, each
number in decision matrix being a triangular neutrosophic number. The scale of
neutrosophic AHP is presented in Table 3:

Saaty scale Explanation Neutrosophic Triangular Scale

1 Equally significant
3 Slightly significant
5 Strongly significant
7 very strongly significant
9 Absolutely significant
1 〈 1, 1, 1 ; 0.50 ,0.50, 0.50〉
3 〈 2, 3, 4 ; 0.30 ,0.75, 0.70〉
5 〈 4, 5, 6 ; 0.80 ,0.15, 0.20〉
7 〈 6, 7, 8 ; 0.90 ,0.10, 0.10〉
9 〈 9, 9, 9 ; 1.00 ,0.00, 0.00〉

2 2 〈 1, 2, 3 ; 0.40 ,0.65, 0.60〉

4 4 〈 3, 4, 5 ; 0.60 ,0.35, 0.40〉
sporadic values between two
6 6 〈 5, 6, 7 ; 0.70 ,0.25, 0.30〉
close scales
8 8 〈 7, 8, 9 ; 0.85 ,0.10, 0.15〉
Table 3. Linguistic terms and the identical triangular neutrosophic numbers.

The decision matrix form is as follows:

̃ ̃ ⋯ ̃
= ⋮ ⋱ ⋮ (13)
̃ ̃ ⋯ ̃
The element ̃ is a triangular neutrosophic number
( ̃ 〈 , , ; , , 〉) , where , , are lower, median and upper
bound of neutrosophic number, , , are the truth (positive), indeterminacy
(boundary) and falsity (negative) functions respectively, ̃ represents the

preference relation of ith criterion over jth criterion. The value of ̃ ̃

.
 Check consistency of decision matrix :
To ensure decision quality, we have to consider the consistency of decision
matrix during the evaluation process. If the decision matrix has a transitive
relation i.e. for all , and , then the decision matrix is consistent.
(i.e. , , = , , . , , ).
To determine the consistency degree perform the following:
Calculate consistency index (CI) and consistency ratio (CR) as follows:
 All values in the first column of the decision matrix should be multiplied by
the priority of the first item; this process continues for all columns of the
comparison matrix. Sum the values across rows to obtain a vector of values labeled
"weighted sum".
 The elements of the weighted sum vector should be divided by the
corresponding priority for each criterion.
 Compute the mean of the values found in the previous step; this mean is
denoted  .
Since  is still neutrosophic number, then we need to de-neutrosophy it by
( ) as follows:

  
( )= 3    (14)

 Compute the consistency index (CI) as follows:


CI (15)

where is the number of criteria being compared.

 Compute the consistency ratio, which is defined as:
CR= (16)

where RI is the consistency index of a randomly generated pair-wise comparison

matrix, as in Table 4.
0 2 3 4 5 6 7 8 9 10
0.0 0.0 0.58 0.90 1.12 1.24 1.32 1.4 1.45 1.49
Table 4. Saaty table for random consistency index (RI) per different number of criteria.

 Use the evaluation function to transform the neutrosophic number of decision

matrix to crisp number as follows:
 ̃ ̃ ̃
( ̃ )= 3 ̃ ̃ ̃ (17)

 Calculate the weight of criteria and alternatives with respect to each criterion by
taking the totality row averages, and performing a normalization process.
4.3 Neutrosophic AHP-QFD methodology
The steps of the proposed methodology are presented in Figure 2.

Start

Collect information about the problem and

analyze it

Identify quality requirements (WHATs)

Identify quality characteristics (HOWs)

Determine correlation matrix of HOWs

Calculate weights of (WHATs) using

neutrosophic AHP

Construct the relationship matrix between

(WHATs) and (HOWs)

Calculate weights of HOWs

According to weights of HOWs, determine

how to achieve target
 Determine competitor's matrix of suppliers

Calculate scores of suppliers using

neutrosophic AHP

Rank suppliers

End

Figure 2. The framework of proposed methodology.

In the proposed methodology, the mean for collecting information by experts resided
in a set of questionnaires. The first questionnaire was planned to identify the
company's stakeholder's requirements (WHATs). The second questionnaire was
intended to determine the evaluation criteria for selecting suppliers. The third
questionnaire was expected to show how stakeholders achieve their requirements (i.e.
determination process of HOWs). The fourth questionnaire requested the experts to
ascertain the decision matrix for calculating weights of requirements; and the final
questionnaire was needed to establish a list of competitors.
5. The case study: results, analysis and comparison analysis
The details about our case study and steps of the proposed methodology are
discussed extensively in this section, and a comparison with other existing MCDM
approaches is compassed.
5.1 Implementation of case study
Our case study has been implemented on a pharmaceutical manufacturing company in
Egypt. The company manufactures and develops several drugs and seeks to apply
different laws concerning drugs manufacturing, drugs experimentation and drugs
marketing. For the manufacturing process, the company needs suitable components
(i.e. chemicals and raw materials) and correct formulas. The company can obtain
chemicals and raw materials from different suppliers, and in order to select the best
one, we use the integrated neutrosophic AHP-QFD methodology, which will help
company in responding to all stakeholders requirements.
5.1.1 Collect and analyze information
 To determine requirements of company's stakeholders (WHATs) a questionnaire is
performed to collect their answers.
5.1.2 Identify quality requirements (WHATs)
From the stakeholder's answers at the first questionnaire, we found out that the
requirements for chemicals and raw materials purchasing process from outside
suppliers by the pharmaceutical company are as follows:
 Quality,
 Cost,
 Location ,
 Price,
 Delivery.
5.1.3 Identify quality characteristics (HOWs)
Considering the answers to the second questionnaire, the analysis process determined
seven criteria to supplier's evaluation. The criteria are as follows:
 Years of experience in this field,
 Certification system of quality,
 Geographical location,
 Constancy of financial status,
 Raw materials safety,
 Organizational behavior,
 Discounts.
5.1.4 Determine the correlation matrix between each criterion (HOWs)
The correlation matrix of the QFD contains these symbols: ++, which means a strong
positive correlation between HOWs, +, designing a positive correlation, –, denoting a
negative correlation, and – –, which means a strong negative correlation. The case
study correlation matrix between HOWs is presented in Figure 3.
5.1.5 Calculate weights of WHATs using neutrosophic AHP

〈 1, 1, 1 ; 0.50 ,0.50, 0.50〉 〈 1, 1, 1 ; 0.50 ,0.50, 0.50〉 〈 4, 5, 6 ; 0.80 ,0.15, 0.20〉 〈 6, 7, 8 ; 0.90 ,0.10, 0.10〉 〈 4, 5, 6 ; 0.80 ,0.15, 0.20〉

〈 1, 1, 1 ; 0.50 ,0.50, 0.50〉 〈 1, 1, 1 ; 0.50 ,0.50, 0.50〉 〈 4, 5, 6 ; 0.80 ,0.15, 0.20〉 〈 4, 5, 6 ; 0.90 ,0.10, 0.10〉 〈 4, 5, 6 ; 0.90 ,0.10, 0.10〉
 1 1 1 1 1 1 1 1 1
〈 , , ; 0.80 ,0.15, 0.20〉 〈 , , ; 0.80 ,0.15, 0.20〉 〈 1, 1, 1 ; 0.50 ,0.50, 0.50〉 〈 , , ; 0.30 ,0.75, 0.70〉 〈 2, 3, 4 ; 0.30 ,0.75, 0.70〉
6 5 4 6 5 4 4 3 2
1 1 1 1 1 1 1 1 1
〈 , , ; 0.90 ,0.10, 0.10〉 〈 , , ; 0.90 ,0.10, 0.10〉 〈 2, 3, 4 ; 0.30 ,0.75, 0.70〉 〈 1, 1, 1 ; 0.50 ,0.50, 0.50〉 〈 , , ; 0.80 ,0.15, 0.20〉
8 7 6 6 5 4 6 5 4
1 1 1 1 1 1 1 1 1
〈 , , ; 0.80 ,0.15, 0.20〉 〈 , , ; 0.90 ,0.10, 0.10〉 〈 , , ; 0.30 ,0.75, 0.70〉 〈 4, 5, 6 ; 0.80 ,0.15, 0.20〉 〈 1, 1, 1 ; 0.50 ,0.50, 0.50〉
6 5 4 6 5 4 4 3 2
Using Table 3, a team of experts asked to compare WHATs according to its
importance. The comparison is synthetized in Table 5.

Table 5. Comparison matrix for stakeholder's requirements.

By calculating consistency ratio (CR), we note that it is within the agreed limit (equal
to 0.1).
For calculating weights of criteria, do the following:
 Take the totality row averages using the following equation:
∑ 〈 , , ; , , 〉
(18)

 Since is still a triangular neutrosophic number, we need to de-neutrosophy

it using the following equation:

( = 3 (19)

 After de-neutrosophication of , we obtain ,which is a crisp value, which

we need to normalize using the following equation:
∑
(20)

The normalized weight values of the previous matrix will be as follows:

0.5, 0.4, 0.03, 0.04, 0.03.
5.1.6 Construct the relationship matrix between WHATs and HOWs
The relationship matrix takes neutrosophic values 9, 3 and 1, which means strong
relationship (S), moderate relationship (M) and weak relationship (W) respectively.
Check Table 3 for the triangular representations of these values. The influence of each
HOW on each WHAT is included in Table 6.

HOWs
 Financial status

Organizational
raw materials
Geographical
Certification
experience
Weights of

Discounts
system of

behavior
Years of
WHATs

location
quality

safety
WHATs

Quality 0.5 M S S S
Cost 0.4 W W M M M
Location 0.03 W
Price 0.04 W W M
Delivery 0.03 W M W M
Table 6. Relationship matrix between WHATs and HOWs.

Considering that the strong relationship (S) = 9 and the triangular neutrosophic
number of this value presented in Table 3 , we can obtain the crisp value by using
equation (17), and the same for moderate (3) and weak (1) relationships. The weights
of HOWs and crisp values of these relationship values are presented in Table 7 below:

HOWs
Financial status

Organizational
raw materials
Geographical
Certification
experience
Weights of

Discounts
system of

behavior
Years of
WHATs

location
quality

safety

Total

WHATs

Quality 0.5 1.85 10 10 10 32.35

Cost 0.4 0.5 0.5 1.85 1.85 1.85 6.95
Location 0.03 0.5 0.53
Price 0.04 0.5 0.5 1.85 2.89
Delivery 0.03 0.5 1.85 0.5 1.85 4.73
12.3 47.45
Total 1 3.85 10 2.85 1.85 11.85 3.7
5
Weights 8% 21% 6% 4% 30% 25% 8%
 of
HOWs
Table 7. The weights of HOWs and crisp values of these relationship values.

From the previous table, it is obvious that raw materials safety, organizational
behavior and certification system of quality are the most important criteria according
to company stakeholders.
5.1.7 According to weights of HOWs, determine how to achieve target
Since raw materials safety, organizational behavior and quality are the most important
criteria for company's stakeholders, we can now achieve the target by dealing with
certified suppliers that support these criteria.
5.1.8 Determine the list of competitors
We should determine our suppliers to select the best one. In the case study, the
company has five suppliers.
5.1.9 Calculate the scores for available suppliers using neutrosophic
AHP
Make decision matrices and compare suppliers according to each criterion. The
weights of suppliers using neutrosophic AHP are presented in Table 8.

Suppliers
Supplier 1 Supplier 2 Supplier 3 Supplier 4 Supplier 5
Criteria
Years of
0.42 0.21 0.26 0.08 0.33
experience
Certification
system of 0.11 0.12 0.12 0.13 0.18
quality
Geographical
0.16 0.06 0.11 0.30 0.09
location
Constancy of
financial 0.05 0.05 0.1 0.14 0.08
status
Raw materials
0.11 0.33 0.06 0.30 0.18
safety
 Organizational
0.05 0.11 0.06 0.10 0.10
behavior
Discounts 0.1 0.12 0.29 0.04 0.04
Table 8. Supplier's weights according to each criterion.

5.1.10 Make ranking process of suppliers

To make ranking of suppliers, multiply weights of suppliers (Table 8) with weights of
criteria as follows:

0.08
0.42 0.11 0.16 0.05 0.11 0.05 0.1
0.21 0.12
0.21 0.12 0.06 0.05 0.33 0.11 0.12 0.06 0.18
0.26 0.12 0.11 0.1 0.06 0.06 0.29 ∗ 0.04 0.11
0.08 0.13 0.30 0.14 0.30 0.10 0.04 0.3 0.17
0.25 0.15
0.33 0.18 0.09 0.08 0.18 0.10 0.04
0.08

From the previous results, we can rank suppliers as follows: supplier 2, supplier 4,
supplier 5, supplier 1 and supplier 3. Therefore, the best supplier for the
pharmaceutical manufacturing company is supplier 2. The obtained weights of
suppliers are accurate and very close due to the competitive power of the suppliers.
We can aggregate all previous results in Figure 3.
 Figure 3. The house
h of qualitty and pie chaart of pharmacceutical manuffacturing com
mpany case stu
udy.

5.1..11. Compparative analysis

a
In tthis sectionn, we comp
pare the neeutrosophic AHP-QFD
D methodoloogy with other
o
exissting MCDM
M approach
hes.
 G. Rajeesh and P. Malliga prooposed an AHP-QFD methodoloogy for suppplier
selectioon problem
m [73]; Leee et al. [7
74] propou
unded a deecision maaking
methoddology by in
ntegrating Q
QFD with analytic
a nettwork proceess (ANP); and
Kumaraaswamy et al.
a [75] sugggested a fraamework inttegrating thhe QFD with
h the
techniqque for ordeer preferencce by simillarity to ideeal solutionn (TOPSIS)) for
supplierr selection problem.
p Alll these reseearches have used crispp values in their
t
 calculations. Since the criteria for supplier selection process are always
conflicting, vague and inconsistent in nature, the crisp values does not
represent them efficiently.
 Alinezad et al. [76] integrated fuzzy AHP with QFD for evaluating and
selecting suppliers of a pharmaceutical company; Karsak and Dursun
proposed a fuzzy multi-criteria group decision making approach by combining
QFD with data envelopment analysis (DEA) for supplier selection problem
[77]; also, Dursun and Şener proposed a fuzzy multi-criteria group decision
making methodology depending on QFD and decision making trial and
evaluation laboratory (DEMATEL) for supplier selection problem [78].
All previous fuzzy multi-criteria decision making techniques failed to deal effectively
and efficiently with vague and inconsistent information for the following reasons:
 Fuzzy set only depends on the truth-membership function.
 Fuzzy set fails to consider the indeterminacy degree.
 In real life problems, the output of any decision making process can take the
following answers: agree, disagree and not sure. In this case, the fuzzy set only
take truthiness (agree), but fails to consider disagree (falsity) and not sure
(indeterminacy), so fuzzy set has a weak representation of vague and
inconsistent information.
We designed this model in order to handle all previous drawbacks and consider all
aspects of decision making process. Our model depends on neutrosophic set, which is
a generalization of fuzzy and intuitionistic fuzzy sets that can deal effectively and
efficiently with vague and inconsistent information by considering the truth,
indeterminacy and falsity membership functions. Therefore, the proposed model is
more adequate than previous approaches to deal with complex and vague issues in
real problems.
6. Conclusions and future works
Neutrosophic set represents the three-way decisions theory effectively, efficiently and
more flexible than fuzzy and intuitionistic fuzzy sets. By studying the three-way
decisions, we proposed two rules of decision making. An evaluation function of three-
way decisions is presented to calculate weights of alternatives. In addition, an
integrated neutrosophic AHP-QFD approach has been presented to select the best
supplier from different alternatives. To illustrate how this approach works, a case
study was presented. To determine the stakeholders requirements, the neutrosophic
QFD was employed, and to calculate the weights of stakeholder's requirements
(WHATs) and the weights of available suppliers, a neutrosophic AHP was used. The
main benefits of this methodology are handling of vague and inconsistent
information; achieving stakeholder's requirements efficiently and effectively; accurate
measure of weights for stakeholder's requirements and scores of suppliers through the
neutrosophic AHP. In traditional AHP, the scale values are crisp numbers, but in
neutrosophic AHP approach, they are triangular numbers, and because the scale
values can change from person to another, the neutrosophic numbers will make the
decisions less hazardous. In the future, we plan to use other ranking methods, such as
neutrosophic TOPSIS, to compute weights of the stakeholders requirements and
compare the results with different methods. We also intend to apply three-way
decisions theory based on neutrosophic set to various real life problems.
Competing interests
The authors announce that there is no conflict of interests concerning the publication
of this research.

References
[1] R. Goudey, "Do statistical inferences allowing three alternative decisions give
better feedback for environmentally precautionary decision-making?," Journal
of Environmental Management, vol. 85, pp. 338-344, 2007.
[2] P. W. Woodward and J. C. Naylor, "An application to Bayesian methods in
SPC," The Statistician, pp. 461-469, 1993.
[3] J. D. Lurie and H. C. Sox, "Principles of medical decision making," Spine, vol.
24, pp. 493-498, 1999.
[4] S. G. Pauker and J. P. Kassirer, "The threshold approach to clinical decision
making," New England Journal of Medicine, vol. 302, pp. 1109-1117, 1980.
[5] C. B. Schechter, "Sequential analysis in a Bayesian model of diastolic blood
pressure measurement," Medical Decision Making, vol. 8, pp. 191-196, 1988.
[6] M. Sherif and C. I. Hovland, "Social judgment: Assimilation and contrast
effects in communication and attitude change," 1961.
[7] A. Wald, "Sequential tests of statistical hypotheses," The Annals of
Mathematical Statistics, vol. 16, pp. 117-186, 1945.
[8] N. Azam and J. Yao, "Multiple criteria decision analysis with game-theoretic
 rough sets," Rough Sets and Knowledge Technology, pp. 399-408, 2012.
[9] J. W. Grzymala-Busse, "Generalized Parameterized Approximations," in
RSKT, 2011, pp. 136-145.
[10] J. W. Grzymala Busse and Y. Yao, "Probabilistic rule induction with the
LERS data mining system," International Journal of Intelligent Systems, vol.
26, pp. 518-539, 2011.
[11] J. P. Herbert and J. Yao, "Learning Optimal Parameters in Decision-Theoretic
Rough Sets," in RSKT, 2009, pp. 610-617.
[12] J. P. Herbert and J. Yao, "Game-theoretic rough sets," Fundamenta
Informaticae, vol. 108, pp. 267-286, 2011.
[13] X. Jia, W. Li, L. Shang, and J. Chen, "An Optimization Viewpoint of
Decision-Theoretic Rough Set Model," in RSKT, 2011, pp. 457-465.
[14] X. Jia, K. Zheng, W. Li, T. Liu, and L. Shang, "Three-way decisions solution
to filter spam email: an empirical study," in Rough Sets and Current Trends in
Computing, 2012, pp. 287-296.
[15] H. Li and X. Zhou, "Risk decision making based on decision-theoretic rough
set: a three-way view decision model," International Journal of
Computational Intelligence Systems, vol. 4, pp. 1-11, 2011.
[16] H. Li, X. Zhou, T. Li, G. Wang, D. Miao, and Y. Yao, "Decision-theoretic
rough sets theory and recent research," ed. Science Press, Beijing, 2011.
[17] H. Li, X. Zhou, J. Zhao, and B. Huang, "Cost-Sensitive Classification Based
on Decision-Theoretic Rough Set Model," in RSKT, 2012, pp. 379-388.
[18] H. Li, X. Zhou, J. Zhao, and D. Liu, "Attribute reduction in decision-theoretic
rough set model: a further investigation," in International Conference on
Rough Sets and Knowledge Technology, 2011, pp. 466-475.
[19] W. Li, D. Miao, W. Wang, and N. Zhang, "Hierarchical rough decision
theoretic framework for text classification," in Cognitive Informatics (ICCI),
2010 9th IEEE International Conference on, 2010, pp. 484-489.
[20] L. Dun, L. Huaxiong, and Z. Xianzhong, "Two decades' research on decision-
theoretic rough sets," in Cognitive Informatics (ICCI), 2010 9th IEEE
International Conference on, 2010, pp. 968-973.
[21] D. Liu, T. Li, and H. Li, "A multiple-category classification approach with
decision-theoretic rough sets," Fundamenta Informaticae, vol. 115, pp. 173-
 188, 2012.
[22] D. Liu, T. Li, and D. Liang, "A new discriminant analysis approach under
decision-theoretic rough sets," in International Conference on Rough Sets and
Knowledge Technology, 2011, pp. 476-485.
[23] D. Liu, T.-r. Li, and D. Liang, "Decision-Theoretic Rough Sets with
Probabilistic Distribution," in RSKT, 2012, pp. 389-398.
[24] D. Liu, T. Li, and D. Ruan, "Probabilistic model criteria with decision-
theoretic rough sets," Information Sciences, vol. 181, pp. 3709-3722, 2011.
[25] Y. Yao and X. Deng, "Sequential three-way decisions with probabilistic rough
sets," in Cognitive Informatics & Cognitive Computing (ICCI* CC), 2011 10th
IEEE International Conference on, 2011, pp. 120-125.
[26] J. Liu, F. Min, S. Liao, and W. Zhu, "Minimal Test Cost Feature Selection
with Positive Region Constraint," in RSCTC, 2012, pp. 259-266.
[27] G. Wang and H. Yu, "Multiple-category attribute reduct using decision-
theoretic rough set model," in International Conference on Rough Sets and
Current Trends in Computing, 2012, pp. 267-276.
[28] X. Yang, H. Song, and T.-J. Li, "Decision making in incomplete information
system based on decision-theoretic rough sets," in International Conference
on Rough Sets and Knowledge Technology, 2011, pp. 495-503.
[29] X. Yang and J. Yao, "Modeling multi-agent three-way decisions with
decision-theoretic rough sets," Fundamenta Informaticae, vol. 115, pp. 157-
171, 2012.
[30] Y. Yao and X. Deng, "Sequential three-way decisions with probabilistic rough
sets," in Cognitive Informatics & Cognitive Computing (ICCI* CC), 2011 10th
IEEE International Conference on, 2011, pp. 120-125.
[31] H. Yu, S. Chu, and D. Yang, "Autonomous knowledge-oriented clustering
using decision-theoretic rough set theory," Fundamenta Informaticae, vol.
115, pp. 141-156, 2012.
[32] H. Yu, Z. Liu, and G. Wang, "Automatically determining the number of
clusters using decision-theoretic rough set," Rough Sets and Knowledge
Technology, pp. 504-513, 2011.
[33] H. Yu and Y. Wang, "Three-Way Decisions Method for Overlapping
Clustering," in RSCTC, 2012, pp. 277-286.
[34] B. Zhou, "A New Formulation of Multi-category Decision-Theoretic Rough
 Sets," in RSKT, 2011, pp. 514-522.
[35] B. Zhou, Y. Yao, and J. Luo, "A Three-Way Decision Approach to Email
Spam Filtering," in Canadian Conference on AI, 2010, pp. 28-39.
[36] X. Zhou and H. Li, "A Multi-View Decision Model Based on Decision-
Theoretic Rough Set," in RSKT, 2009, pp. 650-657.
[37] I. PAWLAK, "Pawlak," Rough Sets, Theoretical Aspects of Reasoning about
Data. K L UWE R ACADE M IC PUB L IS4x E R S, 1991.
[38] U. o. R. D. o. C. Science and W. Ziarko, A probabilistic model of approximate
classification and decision rules with uncertainties in inductive learning:
Regina: Department of Computer Science, University of Regina, 1985.
[39] S. M. Wong and W. Ziarko, "Comparison of the probabilistic approximate
classification and the fuzzy set model," Fuzzy Sets and Systems, vol. 21, pp.
357-362, 1987.
[40] Y. Yao and S. K. M. Wong, "A decision theoretic framework for
approximating concepts," International Journal of Man-machine Studies, vol.
37, pp. 793-809, 1992.
[41] Y. Yao, "Decision-theoretic rough set models," Rough sets and knowledge
technology, pp. 1-12, 2007.
[42] Y. Yao, "Probabilistic approaches to rough sets," Expert systems, vol. 20, pp.
287-297, 2003.
[43] Y. Yao, "Three-Way Decision: An Interpretation of Rules in Rough Set
Theory," in RSKT, 2009, pp. 642-649.
[44] Y. Yao, "Three-way decisions with probabilistic rough sets," Information
Sciences, vol. 180, pp. 341-353, 2010.
[45] Y. Yao, "The superiority of three-way decisions in probabilistic rough set
models," Information Sciences, vol. 181, pp. 1080-1096, 2011.
[46] L. A. Zadeh, "Information and control," Fuzzy sets, vol. 8, pp. 338-353, 1965.
[47] K. T. Atanassov, "Intuitionistic fuzzy sets," Fuzzy sets and Systems, vol. 20,
pp. 87-96, 1986.
[48] F. Smarandache, A Unifying Field in Logics: Neutrosophic Logic. Neutrosophy,
Neutrosophic Set, Neutrosophic Probability: Neutrsophic Logic. Neutrosophy,
Neutrosophic Set, Neutrosophic Probability: Infinite Study, 2005.
[49] P. Liu and Y. Wang, "Multiple attribute decision-making method based on
single-valued neutrosophic normalized weighted Bonferroni mean," Neural
Computing and Applications, vol. 25, pp. 2001-2010, 2014.
[50] L. M. Ellram, "The supplier selection decision in strategic partnerships," Journal
of Supply Chain Management, vol. 26, pp. 8-14, 1990.
[51] C. A. Weber, J. R. Current, and W. Benton, "Vendor selection criteria and
methods," European journal of operational research, vol. 50, pp. 2-18, 1991.
[52] L. De Boer, E. Labro, and P. Morlacchi, "A review of methods supporting
supplier selection," European journal of purchasing & supply management,
vol. 7, pp. 75-89, 2001.
[53] E.-K. Lee, S. Ha, and S.-K. Kim, "Supplier selection and management system
considering relationships in supply chain management," IEEE transactions on
Engineering Management, vol. 48, pp. 307-318, 2001.
[54] W. Ho, X. Xu, and P. K. Dey, "Multi-criteria decision making approaches for
supplier evaluation and selection: A literature review," European Journal of
operational research, vol. 202, pp. 16-24, 2010.
[55] M. Bevilacqua, F. Ciarapica, and G. Giacchetta, "A fuzzy-QFD approach to
supplier selection," Journal of Purchasing and Supply Management, vol. 12,
pp. 14-27, 2006.
[56] S. H. Ha and R. Krishnan, "A hybrid approach to supplier selection for the
maintenance of a competitive supply chain," Expert systems with applications,
vol. 34, pp. 1303-1311, 2008.
[57] F. Faez, S. Ghodsypour, and C. O’Brien, "Vendor selection and order allocation
using an integrated fuzzy case-based reasoning and mathematical
programming model," International Journal of Production Economics, vol.
121, pp. 395-408, 2009.
[58] A. Bhattacharya, J. Geraghty, and P. Young, "Supplier selection paradigm: An
integrated hierarchical QFD methodology under multiple-criteria
environment," Applied Soft Computing, vol. 10, pp. 1013-1027, 2010.
[59] M. Punniyamoorthy, P. Mathiyalagan, and P. Parthiban, "A strategic model using
structural equation modeling and fuzzy logic in supplier selection," Expert
Systems with Applications, vol. 38, pp. 458-474, 2011.
[60] M. Zeydan, C. Çolpan, and C. Çobanoğlu, "A combined methodology for
supplier selection and performance evaluation," Expert Systems with
 Applications, vol. 38, pp. 2741-2751, 2011.
[61] M. Ni, X. Xu, and S. Deng, "Extended QFD and data-mining-based methods
for supplier selection in mass customization," International Journal of
Computer Integrated Manufacturing, vol. 20, pp. 280-291, 2007.
[62] M. Bevilacqua, F. E. Ciarapica, and B. Marchetti, "Development and test of a
new fuzzy-QFD approach for characterizing customers rating of extra virgin
olive oil," Food Quality and Preference, vol. 24, pp. 75-84, 2012.
[63] C. Gencer and D. Gürpinar, "Analytic network process in supplier selection: A
case study in an electronic firm," Applied mathematical modelling, vol. 31, pp.
2475-2486, 2007.
[64] L. Gao and A. Hailu, "Ranking management strategies with complex
outcomes: An AHP-fuzzy evaluation of recreational fishing using an
integrated agent-based model of a coral reef ecosystem," Environmental
Modelling& Software, vol. 31, pp. 3-18, 2012.
[65] O. Kilincci and S. A. Onal, "Fuzzy AHP approach for supplier selection in a
washing machine company," Expert systems with Applications, vol. 38, pp.
9656-9664, 2011.
[66] A. Zouggari and L. Benyoucef, "Simulation based fuzzy TOPSIS approach for
group multi-criteria supplier selection problem," Engineering Applications of
Artificial Intelligence, vol. 25, pp. 507-519, 2012.
[67] B. Jovanović and B. Delibašić, "Application of integrated QFD and fuzzy
AHP approach in selection of suppliers," Management: Journal of Sustainable
Business and Management Solutions in Emerging Economies, vol. 19, pp. 25-
35, 2014.
[68] Y. Yao, "An outline of a theory of three-way decisions," in Rough Sets and
Current Trends in Computing, 2012, pp. 1-17.
[69] M. Abdel-Baset, I. M. Hezam, and F. Smarandache, "Neutrosophic goal
programming," Neutrosophic Sets Syst, vol. 11, pp. 112-118, 2016.
[70] M. Kryszkiewicz, "Rough set approach to incomplete information systems,"
Information sciences, vol. 112, pp. 39-49, 1998.
[71] Y. Akao and G. H. Mazur, "The leading edge in QFD: past, present and
future," International Journal of Quality & Reliability Management, vol. 20,
pp. 20-35, 2003.
[72] T. L. Saaty, "What is the analytic hierarchy process?," in Mathematical
 models for decision supported: Springer, 1988, pp. 109-121.
[73] G. Rajesh and P. Malliga, "Supplier selection based on AHP QFD methodology,"
Procedia Engineering, vol. 64, pp. 1283-1292, 2013.
[74] Y.-T. Lee, W.-W. Wu, and G.-H. Tzeng, "An effective decision-making
method using a combined QFD and ANP approach," WSEAS Transactions on
Business and Economics, vol. 12, pp. 541-551, 2008.
[75] A. H. Kumaraswamy, A. Bhattacharya, V. Kumar, and M. Brady, "An integrated
QFD-TOPSIS methodology for supplier selection in SMEs," in Computational
Intelligence, Modelling and Simulation (CIMSiM), 2011 Third International
Conference on, 2011, pp. 271-276.
[76] A. Alinezad, A. Seif, and N. Esfandiari, "Supplier evaluation and selection
with QFD and FAHP in a pharmaceutical company," The International Journal
of Advanced Manufacturing Technology, vol. 68, pp. 355-364, 2013.
[77] E. E. Karsak and M. Dursun, "An integrated supplier selection methodology
incorporating QFD and DEA with imprecise data," Expert Systems with
Applications, vol. 41, pp. 6995-7004, 2014.
[78] M. Dursun and Z. Şener, "An Integrated DEMATEL-QFD Model for Medical
Supplier Selection," World Academy of Science, Engineering and
Technology, International Journal of Mechanical, Aerospace, Industrial,
Mechatronic and Manufacturing Engineering, vol. 8, pp. 592-596, 2014.
Mohamed Abdel-Basset Received his B.Sc., M.Sc and the Ph.D in operations research from
Faculty of Computers and Informatics, Zagazig University, Egypt. He is a head of department of
operations research and decision support , Faculty of Computers and Informatics, Zagazig
University. His current research interests are Optimization, Operations Research, Data Mining,
Computational Intelligence, Applied Statistics , Decision support systems , Robust Optimization,
Engineering Optimization, Multi-objective Optimization, Swarm Intelligence, Evolutionary
Algorithms, and Artificial Neural Networks. He is working on the application of multi-objective
and robust meta-heuristic optimization techniques. He is also an/a Editor/reviewer in different
international journals and conferences. He has published more than 100 articles in international
journals and conference proceedings.

M. Gunasekaran is currently working as a big data scientist in University of California, Unites

States. He received his PhD from the Vellore Institute of Technology University, India. He
received his Bachelor of Engineering and Master of Technology from Anna University and
Vellore Institute of Technology University respectively. He has worked as a Research Assistant
for a project on spatial data mining funded by Indian Council of Medical Research, Government
of India. His current research interests include data mining, big data analytics and soft
computing. He is the author/co-author of papers in conferences, book chapters and journals. He
got an award for young investigator from India and Southeast Asia by Bill and Melinda Gates
Foundation. He is a member of International Society for Infectious Diseases and Machine
Intelligence Research labs.

Mai Mohamed Received her BS degree and master degree from Zagazig University ,faculty of
computers and informatics,Egypt . she is currently research interest is computation intelligence,
neural networks, and Neutrosophic logic .

Naveen Chilamkurti is currently working as a Senior Lecturer at Department of Computer

Science and Computer Engineering, La Trobe University, Australia. He received his Ph.D. from
La Trobe University. He is also the Inaugural Editor-in-Chief for International Journal of
Wireless Networks and Broadband Technologies launched in July 2011. He has published about
125 journal and conference papers. His current research areas include intelligent transport
systems (ITS), wireless multimedia, wireless sensor networks, vehicle to infrastructure, vehicle
to vehicle communications, health informatics, mobile communications, WiMAX, mobile
security, mobile handover, and RFID. He currently serves on editorial boards of several
international journals. He is a senior member of IEEE. He is also an Associate Editor for Wiley
IJCS, SCN, Interscience JETWI, and IJIPT.
Mohamed Abdel-Basset

M. Gunasekaran

Mai Mohamed
Naveen Chilamkurti
Highlights

 Neutrosophic set is an excellent tool for dealing with vague and inconsistent information.
 Two rules of three-way decisions based on three membership degrees of neutrosophic set
are proposed.
 A new evaluation function is presented to calculate weights of alternatives.
 A Framework for supplier selection problem based on neutrosophic sets and AHP-QFD is
presented for smart environment .
 A case study is presented to illustrate the effectiveness of the proposed model.