3 Neutrosophic Sets and Systems, 18/2017

University of New Mexico

1HXWURVRSKLF6RIW0DWUL[DQGLWVDSSOLFDWLRQ
WR'HFLVLRQ0DNLQJ
Tuhin Bera1 , Nirmal Kumar Mahapatra 2
1 Department of Mathematics, Boror S. S. High School, Bagnan, Howrah-711312,WB, India. E-mail: tuhin78bera@gmail.com
2 Department of Mathematics, Panskura Banamali College, Panskura RS-721152,WB, India. E-mail: nirmal hridoy@yahoo.co.in (corresponding author)

Abstract: The motivation of this paper is to extend the concept of
Neutrosophic soft matrix (NSM) theory. Some basic definitions of
classical matrix theory in the parlance of neutrosophic soft set the-
ory have been presented with proper examples. Then, a theoretical
studies of some traditional operations of NSM have been developed.
Finally, a decision making theory has been proposed by developing
an appropriate solution algorithm, namely, score function algorithm
and it has been illustrated by suitable examples.

Keywords: Intuitionistic fuzzy soft matrix, Neutrosophic soft set, Neutrosophic soft matrix and different operators, Application in decision making.

1 Introduction Evaluation of non-membership values is also not always pos-

Researchers in economics, sociology, medical science, engineer-
ing, environment science and many other several fields deal daily
with the vague, imprecise and occasionally insufficient informa-
tion of modeling uncertain data. Such uncertainties are usually
handled with the help of the topics like probability, fuzzy sets
[1], intuitionistic fuzzy sets [2], interval mathematics, rough sets
etc. But, Molodtsov [3] has shown that each of the above topics
suffers from inherent difficulties possibly due to inadequacy of
their parametrization tool and there after, he initiated a novel con-
cept ‘soft set theory’ for modeling vagueness and uncertainties.
It is completely free from the parametrization inadequacy syn-
drome of different theories dealing with uncertainty. This makes
the theory very convenient, efficient and easily applicable in prac-
tice. Molodtsov [3] successfully applied several directions for the
applications of soft set theory, such as smoothness of functions,
game theory, operation research, Riemann integration, Perron in-
tegration and probability etc. In 2010, Cagman and Enginoglu [4]
introduced a new soft set based decision making method which
selects a set of optimum elements from the alternatives. Maji et
al. [5, 6] have done further research on soft set theory.
Presence of vagueness demanded ‘fuzzy soft set’ [7] to come
into picture. But satisfactory evaluation of membership values is
not always possible because of the insufficiency in the available
information situation. For that, Maji et al. [8, 9] have introduced
the notion ‘intuitionistic fuzzy soft set’ in 2001. Matrices play
an important role in the broad area of science and engineering.
Classical matrix theory sometimes fails to solve the problems in-
volving uncertainties. Hence, several authors proposed the ma-
trix representation of soft set, fuzzy soft set, intuitionistic fuzzy
soft set and applied these in certain decision making problems,
for instance Cagman and Enginoglu [10], Yong and Chenli [11],
Borah et al. [12], Neog and Sut [13], Broumi et al. [14], Mondal
and Roy [15], Chetia and Das [16], Basu et al. [17], Rajara-
jeswari and Dhanalakshmi [18].
sible for the same reason as in case of membership values and
so, there exist an indeterministic part upon which hesitation sur-
vives. As a result, Smarandache [19, 20] has introduced the con-
cept of Neutrosophic Set (NS) which is a generalisation of clas-
sical sets, fuzzy set, intuitionistic fuzzy set etc. Later, Maji [21]
has introduced a combined concept Neutrosophic soft set (NSS).
Using this concept, several mathematicians have produced their
research works in different mathematical structures, for instance
Deli [22, 24], Broumi and Smarandache [25]. Later, this con-
cept has been modified by Deli and Broumi [26]. Accordingly,
Bera and Mahapatra [23, 27-31] introduce some view on alge-
braic structure on neutrosophic soft set. The development of de-
cision making algorithms over neutrosophic soft set theory are
seen in the literatures [32-37].
The present study aims to extend the NSM theory by develop-
ing the basic definitions of classical matrix theory and by estab-
lishing some results in NSS theory context. The organisation of
the paper is as following :
Section 2 deals some preliminary necessary definitions which
will be used in rest of this paper. In Section 3, the concept of
NSM has been discussed broadly with suitable examples. Then,
some traditional operators of NSM are proposed along with some
properties in Section 4. In Section 5, a decision making algo-
rithm has been developed and applied in two different situations.
Firstly, it has been adopted in a class room to select the best stu-
dent in an academic year and then in national security system to
emphasize the security management in five mega cities. This al-
gorithm is much more brief and simple rather than others. More-
over, a decision can be made with respect to a lot of parameters
concerning the fact easily by that. That is why, this algorithm is
more generous, we think. Finally, the conclusion of the present
work has been stated in Section 6.

Tuhin Bera, Nirmal Kumar Mahapatra, Neutrosophic Soft Matrix and its application to Decision Making
 Neutrosophic Sets and Systems, 18/2017 4

2 Preliminaries 2.4 Definition [20]

In this section, we recall some necessary definitions related to Let X be a space of points (objects), with a generic element
fuzzy set, intuitionistic fuzzy soft matrix, neutrosophic set, neu- in X denoted by x. A neutrosophic set A in X is charac-
trosophic soft set, NSM for the sake of completeness. terized by a truth-membership function TA , an indeterminacy-
membership function IA and a falsity-membership function FA .
TA (x), IA (x) and FA (x) are real standard or non-standard sub-
2.1 Definition [28] sets of ]− 0, 1+ [. That is TA , IA , FA : X →]− 0, 1+ [. There
is no restriction on the sum of TA (x), IA (x), FA (x) and so,
−
A binary operation ∗ : [0, 1] × [0, 1] → [0, 1] is continuous t - 0 ≤ sup TA (x) + sup IA (x) + sup FA (x) ≤ 3+ .
norm if ∗ satisfies the following conditions:
(i) ∗ is commutative and associative.
(ii) ∗ is continuous. 2.5 Definition [3]
(iii) a ∗ 1 = 1 ∗ a = a, ∀a ∈ [0, 1].
(iv) a ∗ b ≤ c ∗ d if a ≤ c, b ≤ d with a, b, c, d ∈ [0, 1]. Let U be an initial universe set and E be a set of parameters. Let
A few examples of continuous t-norm are a ∗ b = ab, a ∗ b = P (U ) denote the power set of U . Then for A ⊆ E, a pair (F, A)
min{a, b}, a ∗ b = max{a + b − 1, 0}. is called a soft set over U , where F : A → P (U ) is a mapping.

2.2 Definition [28] 2.6 Definition [21]

A binary operation  : [0, 1] × [0, 1] → [0, 1] is continuous t -
conorm (s - norm) if  satisfies the following conditions :
(i)  is commutative and associative.
(ii)  is continuous.
(iii) a  0 = 0  a = a, ∀a ∈ [0, 1].
(iv) a  b ≤ c  d if a ≤ c, b ≤ d with a, b, c, d ∈ [0, 1].
A few examples of continuous s-norm are a  b = a + b −
ab, a  b = max{a, b}, a  b = min{a + b, 1}.
2.3 Definition [16]
Let U be an initial universe, E be the set of parameters and A ⊆
E. Let, (fA , E) be an intuitionistic fuzzy soft set over U . Then
a subset of U × E is uniquely defined by RA = {(u, e) : e ∈
A, u ∈ fA (e)} which is called a relation form of (fA , E). The
membership function and non-membership functions are written
by μRA : U × E → [0, 1] and νRA : U × E → [0, 1] where
μRA (u, e) ∈ [0, 1] and νRA (u, e) ∈ [0, 1] are the membership
value and non-membership value, respectively of u ∈ U for each
e ∈ E. If (μij , νij ) = (μRA (ui , ej ), νRA (ui , ej )), we can define
a matrix [(μij , νij )]m×n =
Let U be an initial universe set and E be a set of parameters. Let
N S(U ) denote the set of all NSs of U . Then for A ⊆ E, a pair
(F, A) is called an NSS over U , where F : A → N S(U ) is a
mapping.
This concept has been modified by Deli and Broumi [26] as
given below.
2.7 Definition [26]
Let U be an initial universe set and E be a set of parameters. Let
N S(U ) denote the set of all NSs of U . Then, a neutrosophic soft
set N over U is a set defined by a set valued function fN repre-
senting a mapping fN : E → N S(U ) where fN is called approx-
imate function of the neutrosophic soft set N . In other words, the
neutrosophic soft set is a parameterized family of some elements
of the set N S(U ) and therefore it can be written as a set of or-
dered pairs,
N = {(e, {< x, TfN (e) (x), IfN (e) (x), FfN (e) (x) >: x ∈ U }) :
e ∈ E}

⎛ ⎞ where TfN (e) (x), IfN (e) (x), FfN (e) (x) ∈ [0, 1] are respectively
(μ11 , ν11 ) (μ12 , ν12 ) ... (μ1n , ν1n )
⎜ (μ21 , ν21 ) (μ22 , ν22 ) ... (μ2n , ν2n ) ⎟ called truth-membership, indeterminacy-membership, falsity-
⎜ ⎟ membership function of fN (e). Since supremum of each T, I, F
⎜ .. .. .. .. ⎟
⎝ . . . . ⎠ is 1 so the inequality 0 ≤ TfN (e) (x)+IfN (e) (x)+FfN (e) (x) ≤ 3
(μm1 , νm1 ) (μm2 , νm2 ) ... (μmn , νmn ) is obvious.

which is called an m × n IFSM of the IFSS (fA , E) over U .

Therefore, we can say that a IFSS (fA , E) is uniquely charac- 2.7.1 Example
terised by the matrix [(μij , νij )]m×n and both concepts are inter-
changeable. The set of all m × n IFS matrices over U will be Let U = {h1 , h2 , h3 } be a set of houses and E =
denoted by IFSMm×n . {e1 (beautiful), e2 (good location), e3 , (green surrounding)} be a

Tuhin Bera, Nirmal Kumar Mahapatra, Neutrosophic Soft Matrix and its application to Decision Making
 5 Neutrosophic Sets and Systems, 18/2017

set of parameters describing the nature of houses. Let, 2.9 Definition [26]
fN (e1 ) = {< h1 , (0.5, 0.6, 0.3) >, < h2 , (0.4, 0.7, 0.6) >, 1. Let N be a neutrosophic soft set over N (U ). Then a subset
of N (U ) × E is uniquely defined by : RN = {(fN (x), x) : x ∈
< h3 , (0.6, 0.2, 0.3) >}
E, fN (x) ∈ N (U )} which is called a relation form of (N, E).
fN (e2 ) = {< h1 , (0.6, 0.3, 0.5) >, < h2 , (0.7, 0.4, 0.3) >, The characteristic function of RN is written as :
< h3 , (0.8, 0.1, 0.2) >}
fN (e3 ) = {< h1 , (0.7, 0.4, 0.3) >, < h2 , (0.6, 0.7, 0.2) >, ΘRN : N (U ) × E → [0, 1] × [0, 1] × [0, 1] by
ΘRN (u, x) = (TfN (x) (u), IfN (x) (u), FfN (x) (u))
< h3 , (0.7, 0.2, 0.5) >}
where TfN (x) (u), IfN (x) (u), FfN (x) (u) are truth-membership,
Then N = {[e1 , fN (e1 )], [e2 , fN (e2 )], [e3 , fN (e3 )]} is an NSS
indeterminacy-membership and falsity-membership of u ∈ U ,
over (U, E). The tabular representation of the NSS N is given in
respectively.
Table 1.
2. Let U = {u1 , u2 , · · · , um }, E = {x1 , x2 , · · · , xn } and N be
fN (e1 ) fN (e2 ) fN (e3 ) a neutrosophic soft set over N (U ). Then,
h1 (0.5,0.6,0.3) (0.6,0.3,0.5) (0.7,0.4,0.3) RN fN (x1 ) fN (x2 ) ··· fN (xn )
h2 (0.4,0.7,0.6) (0.7,0.4,0.3) (0.6,0.7,0.2) u1 ΘRN (u1 , x1 ) ΘRN (u1 , x2 ) · · · ΘRN (u1 , xn )
h3 (0.6,0.2,0.3) (0.8,0.1,0.2) (0.7,0.2,0.5) u2 ΘRN (u2 , x1 ) ΘRN (u2 , x2 ) · · · ΘRN (u2 , xn )
Table 1 : Tabular form of NSS N . .. .. .. .. ..
. . . . .
um ΘRN (um , x1 ) ΘRN (um , x2 ) ··· ΘRN (um , xn )
2.8 Definition [26] If aij = ΘRN (ui , xj ), we can define a matrix
1. The complement of a neutrosophic soft set N is denoted by ⎛ ⎞
a11 a12 ··· a1n
N o and is defined by : ⎜ a21 a22 ··· a2n ⎟
⎜ ⎟
[aij ] = ⎜ .. .. .. .. ⎟
N o = {(e, {< x, FfN (e) (x), 1 − IfN (e) (x), TfN (e) (x) >: x ∈ ⎝ . . . . ⎠
U }) : e ∈ E} am1 am2 ··· amn

2. Let N1 and N2 be two NSSs over the common universe (U, E). such that aij = (TfN (xj ) (ui ), IfN (xj ) (ui ), FfN (xj ) (ui )) =
a a a
Then N1 is said to be the neutrosophic soft subset of N2 if ∀e ∈ (Tij , Iij , Fij ), which is called an m × n neutrosophic soft ma-
E and x ∈ U trix (NS-matrix) of the neutrosophic soft set N over N (U ).
According to this definition, a neutrosophic soft set N is
TfN1 (e) (x) ≤ TfN2 (e) (x), IfN1 (e) (x) ≥ IfN2 (e) (x), uniquely characterised by a matrix [aij ]m×n . Therefore, we shall
FfN1 (e) (x) ≥ FfN2 (e) (x). identify any neutrosophic soft set with it’s soft NS-matrix and
use these two concepts as interchangeable. The set of all m × n
We write N1 ⊆ N2 and then N2 is the neutrosophic soft superset NS-matrix over N (U ) will be denoted by Ñm×n . From now on
of N1 . we shall delete the subscripts m × n of [aij ]m×n , we use [aij ]
3. Let N1 and N2 be two NSSs over the common universe (U, E). instead of [aij ]m×n , since [aij ] ∈ Ñm×n means that [aij ] is an
Then their union is denoted by N1 ∪ N2 = N3 and is defined by m × n NS-matrix for i = 1, 2, · · · , m and j = 1, 2, · · · , n.

N3 = {(e, {< x, TfN3 (e) (x), IfN3 (e) (x), FfN3 (e) (x) >: x ∈ 2.10 Definition [26]
U }) : e ∈ E}
Let [aij ], [bij ] ∈ Ñm×n . Then,
where TfN3 (e) (x) = TfN1 (e) (x)  TfN2 (e) (x), IfN3 (e) (x) = 1. [aij ] is a zero NS-matrix, denoted by [0̃], if aij =
IfN1 (e) (x) ∗ IfN2 (e) (x), FfN3 (e) (x) = FfN1 (e) (x) ∗ FfN2 (e) (x). (0, 1, 1), ∀i, j.
4. Let N1 and N2 be two NSSs over the common universe (U, E). 2. [aij ] is a universal NS-matrix, denoted by [1̃], if aij =
Then their intersection is denoted by N1 ∩ N2 = N4 and is de- (1, 0, 0), ∀i, j.
fined by : ˜ ij ], if
3. [aij ] is an NS-submatrix of [bij ], denoted by [aij ]⊆[b
a b a b a b
Tij ≤ Tij , Iij ≥ Iij , Fij ≥ Fij , ∀i, j.
N4 = {(e, {< x, TfN4 (e) (x), IfN4 (e) (x), FfN4 (e) (x) >: x ∈
U }) : e ∈ E} 4. [aij ] and [bij ] are equal NS- matrices, denoted by [aij ] = [bij ],
if aij = bij , ∀i, j.
where TfN4 (e) (x) = TfN1 (e) (x) ∗ TfN2 (e) (x), IfN4 (e) (x) = 5. Complement of [aij ] is denoted by [aij ]o and is defined as
IfN1 (e) (x)  IfN2 (e) (x), FfN4 (e) (x) = FfN1 (e) (x)  FfN2 (e) (x). another NS-matrix [cij ] such that cij = (Fija , 1 − Iij
a
, Tija ), ∀i, j.

Tuhin Bera, Nirmal Kumar Mahapatra, Neutrosophic Soft Matrix and its application to Decision Making
 Neutrosophic Sets and Systems, 18/2017 6

3 Neutrosophic soft matrix e1 e2 e3

h1 (0.5,0.6,0.3) (0.6,0.3,0.5) (0.7,0.4,0.3)
In this section, we have introduced some definitions and have h2 (0.4,0.7,0.6) (0.7,0.4,0.3) (0.6,0.7,0.2)
included some new operations related to NSM. h3 (0.6,0.2,0.3) (0.8,0.1,0.2) (0.7,0.2,0.5)

Hence, the NSM corresponding to this NSS N over (U, E) is :

3.1 Definition
⎛ ⎞
Let U = {u1 , u2 , · · · , um } and E = {e1 , e2 , · · · , en } be the (0.5, 0.6, 0.3) (0.6, 0.3, 0.5) (0.7, 0.4, 0.3)
universal set of objects and the parametric set, respectively. Sup- [aij ]3×3 = ⎝ (0.4, 0.7, 0.6) (0.7, 0.4, 0.3) (0.6, 0.7, 0.2) ⎠
pose, N be a neutrosophic soft set over (U, E) given by N = {< (0.6, 0.2, 0.3) (0.8, 0.1, 0.2) (0.7, 0.2, 0.5)
(e, fN (e)) >: e ∈ E} where
Next, let E1 = {e1 (cheap), e2 (moderate), e3 (high), e4 (very high)}
fN (e) = {< u, (TfN (e) (u), IfN (e) (u), FfN (e) (u)) >: u ∈ U }. be another set of parameters describing the cost of houses in U .
The relation form of an NSS M over (U, E1 ) is written as :
Thus, fN (e) corresponds a relation on {e} × U i.e., fN (e) =
{(e, ui ) : 1 ≤ i ≤ m} for each e ∈ E. It is obviously a sym- e1 e2 e3 e4
metric relation. Now, consider a relation RE on U × E given h1 (.4, .5, .5) (.5, .7, .6) (.2, .5, .8) (.5, .6, .4)
by RE = {(u, e) : e ∈ E, u ∈ fN (e)}. It is called a relation h2 (.6, .4, .7) (.6, .3, .4) (.7, .6, .5) (.8, .4, .3)
form of the NSS N over (U, E). The characteristic function of h3 (.7, .3, .4) (.5, .2, .5) (.8, .4, .4) (.1, .6, .6)
RE is χRE : U × E → [0, 1] × [0, 1] × [0, 1] and is defined
as : χRE (u, e) = (TfN (e) (u), IfN (e) (u), FfN (e) (u)). The tabular Here, the NSM corresponding to the NSS M over (U, E1 ) is
representation of RE is given in Table 2. [bij ]3×4 =
⎛ ⎞
e1 e2 ··· en (.4, .5, .5) (.5, .7, .6) (.2, .5, .8) (.5, .6, .4)
u1 χRE (u1 , e1 ) χRE (u1 , e2 ) ··· χRE (u1 , en ) ⎝ (.6, .4, .7) (.6, .3, .4) (.7, .6, .5) (.8, .4, .3) ⎠
u2 χRE (u2 , e1 ) χRE (u2 , e2 ) ··· χRE (u2 , en ) (.7, .3, .4) (.5, .2, .5) (.8, .4, .4) (.1, .6, .6)
.. .. .. .. ..
. . . . .
um χRE (um , e1 ) χRE (um , e2 ) ··· χRE (um , en ) 3.2 Definition
Table 2 : Tabular form of RE
Let A = [aij ] ∈ N SMm×n where aij = (Tija , Iij a
, Fija ). Then,
If aij = χRE (ui , ej ), then we can define a matrix 1. A is called a square NSM if m = n i.e., if the number of rows
⎛ ⎞ and the number of columns are equal. The NSS corresponding to
a11 a12 · · · a1n this NSM has the same number of objects and parameters.
⎜ a21 a22 · · · a2n ⎟
⎜ ⎟ 2. A square NSM A = [aij ]n×n is called upper triangular NSM
[aij ]m×n = ⎜ . .. .. .. ⎟
⎝ .. . . . ⎠ if aij = (0, 1, 1), ∀i > j and is called lower triangular NSM if
am1 am2 ··· amn aij = (0, 1, 1), ∀i < j.
A is called triangular NSM if it is either neutrosophic soft up-
where aij = (TfN (ej ) (ui ), IfN (ej ) (ui ), FfN (ej ) (ui )) = per triangular or neutrosophic soft lower triangular matrix.
(Tija , Iij
a
, Fija ). 3. The transpose of a square NSM A = [aij ]n×n is another
Thus, we shall identify any neutrosophic soft set with it’s NSM square NSM of same order obtained from [aij ] by interchanging
and use these two concepts as interchangeable. Since we con- it’s rows and columns. It is denoted by At . Thus At = [aij ]t =
sider the full parametric set E, so each NSS N over (U, E) cor- [(Tija , Iij
a
, Fija )]t = [(Tji
a a
, Iji a
, Fji )]. The NSS corresponding to
responds a unique NSM [aij ]m×n where cardinality of U and E t
A becomes a new NSS over the same universe and the same
are m and n, respectively. To get another NSM of the same order parametric set.
over (U, E), we need to define another NSS over (U, E). The set
4. A square NSM A = [aij ]n×n is said to be a symmet-
of all NSMs of order m × n is denoted by N SMm×n . Whenever
ric NSM if At = A i.e., if aij = aji or (Tija , Iij a
, Fija ) =
U and E are fixed, we get all NSMs of unique order i.e., to obtain a a a
(Tji , Iji , Fji ), ∀i, j.
an NSM of distinct order, at least any of U and E will have to be
changed.
3.3 Definition
3.1.1 Example
Let A = [aij ] ∈ N SMm×n , where aij = (Tija , Iij
a
, Fija ). Then,
Consider the Example [2.7.1]. The relation form of the NSS N the scalar multiple of NSM A by a scalar k is defined by kA =
over the said (U, E) is [kaij ]m×n where 0 ≤ k ≤ 1.

Tuhin Bera, Nirmal Kumar Mahapatra, Neutrosophic Soft Matrix and its application to Decision Making
 7 Neutrosophic Sets and Systems, 18/2017

3.3.1 Example 3.6 Definition

Let A = [aij ]2×3 = Let A = [aij ] ∈ N SMm×n , where m = n and aij =
 (Tija , Iij
a
, Fija ). Then, the trace of NSM A is denoted by tr(A)
(0.4, 0.5, 0.5) (0.5, 0.7, 0.6) (0.5, 0.6, 0.4)
m
(0.6, 0.4, 0.7) (0.7, 0.3, 0.4) (0.8, 0.4, 0.3) and is defined as tr(A) = i=1 [Tiia − (Iiia + Fiia )].

be an NSM. Then the scalar multiple of this matrix by k = 0.5

is kA = [kaij ]2×3 = 3.6.1 Example

(0.20, 0.25, 0.25) (0.25, 0.35, 0.30) (0.25, 0.30, 0.20) Let A = [aij ]3×3 =
(0.30, 0.20, 0.35) (0.35, 0.15, 0.20) (0.40, 0.20, 0.15)
⎛ ⎞
(0.5, 0.6, 0.3) (0.6, 0.3, 0.5) (0.7, 0.4, 0.3)
3.4 Proposition ⎝ (0.4, 0.7, 0.6) (0.7, 0.4, 0.3) (0.6, 0.7, 0.2) ⎠
(0.6, 0.2, 0.3) (0.8, 0.1, 0.2) (0.7, 0.2, 0.5)
Let A = [aij ], B = [bij ] ∈ N SMm×n where aij =
(Tija , Iij
a
, Fija ). For two scalars s, k ∈ [0, 1],
be an NSM. Then tr(A) = (0.5−0.6−0.3)+(0.7−0.4−0.3)+
(i) s(kA) = (sk)A. (ii) s ≤ k ⇒ sA ≤ kA. (iii) A ⊆ B ⇒ (0.7 − 0.2 − 0.5) = −0.4
sA ⊆ sB.

Proof.
3.7 Proposition
(i) s(kA) = s[kaij ] = s[(kTija , kIij a
, kFija )]
Let A = [aij ] ∈ N SMn×n , where aij = (Tija , Iija
, Fija ). If
= [(skTija , skIij
a
, skFija )] = sk[(Tija , Iij
a
, Fija )] k ∈ [0, 1] is a scalar, then tr(kA) = k tr(A).
= sk[aij ] = (sk)A.
n
n
Proof. tr(kA) = i=1 [kTiia − (kIiia + kFiia )] = k i=1 [Tiia −
(ii) Since Tija , Iij
a
, Fija ∈ [0, 1], ∀i, j so, sTija ≤ kTija , sIij a
≤ (Iiia + Fiia )] = k tr(A).
a a a
kIij , sFij ≤ kFij .
Now, sA = [(sTija , sIij a
, sFija )] ≤ [(kTija , kIij
a
, kFija )] = kA.
3.8 Max-Min Product of NSMs

(iii) A ⊆ B ⇒ [aij ] ⊆ [bij ] Two NSMs A and B are said to be conformable for the
product A ⊗ B if the number of columns of the NSM
⇒ Tija ≤ Tijb , Iij
a b
≥ Iij , Fija ≥ Fijb , ∀i, j A be equal to the number of rows of the NSM B and
⇒ sTija ≤ sTijb , sIij
a b
≥ sIij , sFija ≥ sFijb , ∀i, j this product becomes also an NSM. If A = [aij ]m×n
⇒ s[aij ] ⊆ s[bij ] and B = [bjk ]n×p , then A ⊗ B = [cik ]m×p where
aij = (Tija , Iij
a
, Fija ), bjk = (Tjk b b
, Ijk b
, Fjk ) and cik =
⇒ sA ⊆ sB
(maxj min(Tij , Tjk ), minj max(Iij , Ijk ), minj max(Fija , Fjk
a b a b b
)).
Clearly, B ⊗ A can not be defined here.
3.5 Theorem
Let A = [aij ]m×n be an NSM where aij = (Tija , Iij
a
, Fija ). Then,
3.8.1 Example
(i) (kA)t = k At for k ∈ [0, 1] being a scalar.
⎛ ⎞
(ii) (At )t = A. (0.5, 0.6, 0.3) (0.6, 0.3, 0.5)
(iii) If A = [aij ]n×n is an upper triangular (lower triangular) Let A = [aij ]3×2 = ⎝ (0.4, 0.7, 0.6) (0.7, 0.4, 0.3) ⎠
NSM, then At is lower triangular (upper triangular) NSM. (0.6, 0.2, 0.3) (0.8, 0.1, 0.2)
Proof.(i) Here (kA)t , k At ∈ N SMn×m . Now,
and B = [bjk ]2×3 =
(kA)t = [(kTija , kIij
a
, kFija )]t = [(kTji
a a
, kIji a
, kFji )]

= a
k[(Tji a
, Iji a
, Fji )] = k[(Tija , Iij
a
, Fija )]t = k At . (0.4, 0.5, 0.5) (0.5, 0.7, 0.6) (0.5, 0.6, 0.4)
(0.6, 0.4, 0.7) (0.7, 0.3, 0.4) (0.8, 0.4, 0.3)
(ii) Here At ∈ N SMn×m and so (At )t ∈ N SMm×n . Now,
be two NSMs. Then, A ⊗ B = [cik ]3×3 =
(At )t = ([(Tija , Iij
a
, Fija )]t )t = [(Tji
a a
, Iji a t
, Fji )]
a a a
⎛ ⎞
= [(Tij , Iij , Fij )] = A. (0.6, 0.4, 0.5) (0.6, 0.3, 0.5) (0.6, 0.4, 0.4)
⎝ (0.6, 0.4, 0.6) (0.7, 0.4, 0.4) (0.7, 0.4, 0.3) ⎠
(iii) Straight forward. (0.6, 0.4, 0.5) (0.7, 0.3, 0.4) (0.8, 0.4, 0.3)

Tuhin Bera, Nirmal Kumar Mahapatra, Neutrosophic Soft Matrix and its application to Decision Making
 Neutrosophic Sets and Systems, 18/2017 8


One calculation is provided herewith for convenience of A ⊗ B. Fijc = (w1 +w2 )
(Fija )w1 · (Fijb )w2 , ∀i, j and w1 , w2 > 0.
c a b a b a b
= max{min(T21 ), min(T22 )} 2Tij Tij
T21
j
, T11 , T21 (vii) Harmonic mean A  B = C where Tijc = c
a +T b , Iij
Tij
=
ij
a b a b
= max{min(0.4, 0.4), min(0.7, 0.6)} = 0.6 2Iij Iij
, Fijc =
2Fij Fij
, ∀i, j.
a +I b
Iij a +F b
Fij
c a b a b ij ij
I21 = min{max(I21 , I11 ), max(I22 , I21 )}
j (viii) Weighted harmonic mean A w B = C where Tijc =
w1 +w2 c w1 +w2 c w1 +w2
= min{max(0.7, 0.5), max(0.4, 0.4)} = 0.4 w1 w2 , Iij = w1 w2 , Fij = w1 w2 , ∀i, j and w1 , w2 > 0.
a + a + a +
T
ij Tb I
ij Ib F
ij Fb
c a b a b ij ij ij
F21 = min{max(F21 , F11 ), max(F22 , F21 )}
j
= min{max(0.6, 0.5), max(0.3, 0.7)} = 0.6 4.1 Proposition
Thus, c21 = (0.6, 0.4, 0.6) and so on. Let A = [aij ], B = [bij ] ∈ N SMm×n , where aij =
(Tija , Iij
a
, Fija ). Then,
(i) (A ∪ B)t = At ∪ B t , (A ∩ B)t = At ∩ B t .
3.9 Theorem (ii) (A  B)t = At  B t , (A w B)t = At w B t .
t t t w t t w t
Let A = [aij ]m×n , B = [bjk ]n×p be two NSMs where (iii) (A  B)t = At  Bt , (A w B)t = At w Bt .
aij =(Tija , Iij
a
, Fija ). Then, (A ⊗ B)t = B t ⊗ At (iv) (A  B) = A  B , (A  B) = A  B .
t t t t t
Proof. Let A ⊗ B = [cik ]m×p . Then (A ⊗ B)t = [cki ]p×m , At = Proof. (i) Here A ∪ B, (A ∪ B) , A , B , A ∪ B ∈ N SMm×n .
[aji ]n×m , B t = [bkj ]p×n and so the order of (B t ⊗At ) is (p×m). Now,
Now,
(A ∪ B)t = [(Tija  Tijb , Iij
a b
∗ Iij , Fija ∗ Fijb )]t
(A ⊗ B)t = a
[(Tji b
 Tji a
, Iji b
∗ Iji a
, Fji b
∗ Fji )]
c c c
= [(Tki , Iki , Fki )]p×m = a
[(Tji a
, Iji a
, Fji b
)] ∪ [(Tji b
, Iji b
, Fji )]
b a b a
= [(max min(Tkj , Tji ), min max(Ikj , Iji ), = [(Tija , Iij
a
, Fija )]t ∪ [(Tijb , Iij
b
, Fijb )]t
j j
b
min max(Fkj a
, Fji ))]p×m = At ∪ B t .
j

= b
[(Tkj b
, Ikj b
, Fkj a
)]p×n ⊗ [(Tji a
, Iji a
, Fji )]n×m = B t ⊗ At . Next A ∩ B, (A ∩ B)t , At ∩ B t ∈ N SMm×n . Now,

(A ∩ B)t = [(Tija ∗ Fijb , Iij

a b
 (1 − Iij ), Fija  Tijb )]t
a b a b a b
4 Operators of NSMs = [(Tji ∗ Fji , Iji  (1 − Iji ), Fji  Tji )]
a a a b b b
= [(Tji , Iji , Fji )] ∩ [(Tji , Iji , Fji )]
Let A = [(Tija , Iij
a
, Fija )], B = [(Tijb , Iij
b
, Fijb )] ∈ N SMm×n .
Then, = [(Tija , Iij
a
, Fija )]t ∩ [(Tijb , Iij
b
, Fijb )]t
(i) Union A ∪ B = C where Tijc = Tija  Tijb , Iij c a
= Iij ∗ = At ∩ B t .
b c a b
Iij , Fij = Fij ∗ Fij , ∀i, j.
Remaining others can be proved in the similar manner.
(ii) Intersection A ∩ B = C where Tijc = Tija ∗ Tijb , Iij
c
=
a b c a b
Iij  Iij , Fij = Fij  Fij , ∀i, j.
a b
4.2 Proposition
Tij +Tij
(iii)Arithmetic mean A  B = C where Tijc = , c
Iij =
a b
2 Let A = [aij ], B = [bij ] are upper triangular (lower triangular)
Iij +Iij F a +F b
2 Fijc = ij 2 ij , ∀i, j.
, NSMs of same order. Then (i) A ∪ B, A ∩ B (ii) A  B, A w B
(iv) Weighted arithmetic mean A w B = C where Tijc = (iii) A  B, A w B all are upper triangular (lower triangular)
a
w1 Tij b
+w2 Tij c
a
w1 Iij b
+w2 Iij a
w1 Fij b
+w2 Fij NSMs.
w1 +w2 , Iij = w1 +w2 , Fijc = w1 +w2 , ∀i, j and
w1 , w2 > 0. Proof. Straight forward.

(v) Geometric mean AB = C where Tijc c
= Tija · Tijb , Iij =
 
a · Ib , F c =
Iij Fija · Fijb , ∀i, j.
4.3 Theorem
ij ij

(vi) Weighted geometric mean A w B = C where Let A = [aij ], B = [bij ] be two symmetric NSMs of same order.

Then,
Tijc = (w1 +w2 ) (Tija )w1 · (Tijb )w2 ,
 (i) A ∪ At , A ∪ B, A ∩ B, A  B, A w B, A  B, A w B, A 
c a )w1 · (I b )w2 ,
Iij = (w1 +w2 ) (Iij ij B, A w B are so.

Tuhin Bera, Nirmal Kumar Mahapatra, Neutrosophic Soft Matrix and its application to Decision Making
 9 Neutrosophic Sets and Systems, 18/2017

(ii) A ⊗ B is symmetric iff A ⊗ B = B ⊗ A. = [(Fija , 1 − Iij

a
, Tija )]  [(Fijb , 1 − Iij
b
, Tijb )]
(iii) A ⊗ At , At ⊗ A both are symmetric. = [(Tija , Iij
a
, Fija )]o  [(Tijb , Iij
b
, Fijb )]o
Proof. Here At = A and B t = B as both are symmetric NSMs. = Ao  B o .
t w w
Clearly A ∪ A , A ∪ B, A ∩ B, A  B, A  B, A  B, A 
B, A  B, A w B, A ⊗ B, B ⊗ A, A ⊗ At , At ⊗ A all are well Next, for w1 , w2 > 0, we have,
defined as both the NSMs are same order and square. Now,
(A w B)o
(i) These are left to the reader.
(ii) (A ⊗ B)t = B t ⊗ At = B ⊗ A = A ⊗ B. w1 Tija + w2 Tijb w1 Iij a
+ w2 Iij b
w1 Fija + w2 Fijb o
= [( , , )]
(iii) (A ⊗ At )t = (At )t ⊗ At = A ⊗ At and (At ⊗ A)t = w1 + w2 w1 + w2 w1 + w2
At ⊗ (At )t = At ⊗ A. w1 Fija + w2 Fijb w1 Iija
+ w2 Iijb
w1 Tija + w2 Tijb
= [( ,1 − , )]
w1 + w2 w1 + w2 w1 + w2
4.4 Proposition w1 Fija + w2 Fijb w1 (1 − Iij a
) + w2 (1 − Iij b
)
= [( , ,
w1 + w2 w1 + w2
Let A = [(Tija , Iija
, Fija )], B = [(Tijb , Iij
b
, Fijb )] ∈ N SMm×n .
Then, w1 Tija + w2 Tijb
)]
w1 + w2
(i) (A ∪ B)o = Ao ∩ B o , (A ∩ B)o = Ao ∪ B o .
(ii) (A  B)o = Ao  B o , (A w B)o = Ao w B o . = [(Fija , 1 − Iij
a
, Tija )] w [(Fijb , 1 − Iij
b
, Tijb )]
= [(Tija , Iij
a
, Fija )]o w [(Tijb , Iij
b
, Fijb )]o = Ao w B o .
Proof. (i) Here (A ∪ B)o , Ao ∩ B o ∈ N SMm×n . Now,

4.5 Proposition (Commutative law)

(A ∪ B)o = [(Tija  Tijb , Iij
a b
∗ Iij , Fija ∗ Fijb )]o
Let A = [(Tija , Iij
a
, Fija )], B = [(Tijb , Iij
b
, Fijb )] ∈ N SMm×n .
= [(Fija ∗ Fijb , 1 − (Iij
a b
∗ Iij ), Tija  Tijb )] Then,
= [(Fija ∗ Fijb , (1 − Iij
a b
)  (1 − Iij ), Tija  Tijb )] (i) A ∪ B = B ∪ A, A ∩ B = B ∩ A (ii) A  B = B  A, A w
= [(Fija , 1 − Iij
a
, Tija )] ∩ [(Fijb , 1 − Iij
b
, Tijb )] B = B w A (iii) A  B = B  A, A w B = B w A (iv)
A  B = B  A, A w B = B w A.
= [(Tija , Iij
a
, Fija )]o ∩ [(Tijb , Iij
b
, Fijb )]o
= Ao ∩ B o . Proof. Obvious

Next, (A ∩ B)o , Ao ∪ B o ∈ N SMm×n . Now,

4.6 Proposition (Associative law)
o
(A ∩ B) = [(Tija ∗ Tijb , Iija
 Iijb
, Fija  Fijb )]o Let A = [(Tija , Iij a
, Fija )], B = [(Tijb , Iij
b
, Fijb )], C =
c c c
= [(Fija  Fijb , 1 − (Iij a b
 Iij ), Tija ∗ Tijb )] [(Tij , Iij , Fij )] ∈ N SMm×n . Then,
= [(Fija  Fijb , (1 − Iij a
) ∗ (1 − Iij b
), Tija ∗ Tijb )] (i) (A ∪ B) ∪ C = A ∪ (B ∪ C) (ii) (A ∩ B) ∩ C = A ∩ (B ∩ C)
(iii) (AB)C = A(B C) (iv) (AB)C = A(B C)
= [(Fija , 1 − Iija
, Tija )] ∪ [(Fijb , 1 − Iij b
, Tijb )] (v) (A  B)  C = A  (B  C).
= [(Tija , Iij
a
, Fija )]o ∪ [(Tijb , Iij b
, Fijb )]o
o o
Proof. (i) Clearly (A ∪ B) ∪ C, A ∪ (B ∪ C) ∈ N SMm×n . Now,
= A ∪B .
a b a b
(A ∪ B) ∪ C
Note : Here, (1 − Iij )  (1 − Iij ) = 1 − (Iij ∗ Iij ) and (1 −
a b a b = [(Tija  Tijb , Iij
a b
∗ Iij , Fija ∗ Fijb )] ∪ [(Tijc , Iij
c
, Fijc )]
Iij ) ∗ (1 − Iij ) = 1 − (Iij  Iij ) hold for dual pairs of non-
parameterized t-norms and s-norms e.g., a ∗ b = min{a, b} and = [((Tija  Tijb )  Tijc , (Iij
a b
∗ Iij c
) ∗ Iij , (Fija ∗ Fijb ) ∗ Fijc )]
a  b = max{a, b}, a ∗ b = max{a + b − 1, 0} and a  b = = [(Tija  (Tijb  Tijc ), Iij
a b
∗ (Iij c
∗ Iij ), Fija ∗ (Fijb ∗ Fijc ))]
min{a + b, 1} etc.
= A ∪ (B ∪ C)
(ii) Here (A  B)o , Ao  B o ∈ N SMm×n .
Similarly, the other results can be verified.
Tija + Tijb Iij
a
+ Iijb
Fija + Fijb o
(A  B)o = [( , , )]
2 2 2 4.7 Proposition (Distributive law)
Fija + Fijb a
Iij + Iijb
Tija + Tijb
= [( ,1 − , )] Let A = [(Tija , Iij a
, Fija )], B = [(Tijb , Iij
b
, Fijb )], C =
2 2 2 c c c
[(Tij , Iij , Fij )] ∈ N SMm×n . Then,
Fija + Fijb (1 − Iij a
) + (1 − Iij b
) Tija + Tijb
= [( , , )] (i) A ∩ (B  C) = (A ∩ B)  (A ∩ C), (A  B) ∩ C =
2 2 2

Tuhin Bera, Nirmal Kumar Mahapatra, Neutrosophic Soft Matrix and its application to Decision Making
 Neutrosophic Sets and Systems, 18/2017 10

(A ∩ C)  (B ∩ C). (A  C) ∪ (B  C).
(ii) A ∪ (B  C) = (A ∪ B)  (A ∪ C), (A  B) ∪ C = A  (B ∩ C) = (A  B) ∩ (A  C), (A ∩ B)  C =
(A ∪ C)  (B ∪ C). (A  C) ∩ (B  C).
(v) A  (B ∪ C) = (A  B) ∪ (A  C), (A ∪ B)  C =
Proof. (i) Here A ∩ (B  C), (A ∩ B)  (A ∩ C) ∈ N SMm×n .
(A  C) ∪ (B  C).
Now,
A  (B ∩ C) = (A  B) ∩ (A  C), (A ∩ B)  C =
A ∩ (B  C) (A  C) ∩ (B  C).
Tijb + Tijc Iij b
+ Iij c
Fijb + Fijc Proof. We shall here prove (i), (iv) and (v) only. The others can
= [(Tija , Iij
a
, Fija )] ∩ [( , , )]
2 2 2 be proved in the similar fashion.
Tijb + Tijc a Iij b
+ Iijc
Fijb + Fijc (i) Here A ∩ (B ∪ C), (A ∩ B) ∪ (A ∩ C) ∈ N SMm×n . Now,
= [(Tija ∗ , Iij  , Fija  )]
2 2 2
Tija ∗ Tijb + Tija ∗ Tijc Iij a
 Iijb
+ Iij a c
 Iij A ∩ (B ∪ C)
= [( , , = [(Tija , Iij
a
, Fija )] ∩ [(max{Tijb , Tijc }, min{Iij
b c
, Iij },
2 2
Fija  Fijb + Fija  Fijc min{Fijb , Fijc })]
)]
2 = [(min{Tija , max{Tijb , Tijc }}, max{Iij
a b
, min{Iij c
, Iij }},
= [(Tij ∗ Tij , Iij  Iij , Fija  Fijb )]
a b a b
max{Fija , min{Fijb , Fijc }})]
[(Tija ∗ Tijc , Iij
a
 Iijc
, Fija  Fijc )]
= [(max{min{Tija , Tijb }, min{Tija , Tijc }}, min{max{Iij
a b
, Iij },
= (A ∩ B)  (A ∩ C)
a c
max{Iij , Iij }}, min{max{Fija , Fijb }, max{Fija , Fijc }})]
Next (A  B) ∩ C, (A ∩ C)  (B ∩ C) ∈ N SMm×n . Now,
= [(min{Tija , Tijb }, max{Iij
a b
, Iij }, max{Fija , Fijb })]
(A  B) ∩ C ∪[(min{Tija , Tijc }, max{Iij
a c
, Iij }, max{Fija , Fijc })]
Tija + Tijb Iij a
+ Iij b
Fija + Fijb = (A ∩ B) ∪ (A ∩ C)
= [( , , )] ∩ [(Tijc , Iij
c
, Fijc )]
2 2 2
Next (A ∪ B) ∩ C, (A ∩ C) ∪ (B ∩ C) ∈ N SMm×n . Now,
Tija + Tijb a
Iij + Iij b
Fija + Fijb
= [( ∗ Tijc , c
 Iij ,  Fijc )]
2 2 2 (A ∪ B) ∩ C
Tija ∗ Tijc + Tijb ∗ Tijc Iij a c
 Iij + Iij b c
 Iij = [(max{Tija , Tijb }, min{Iij
a b
, Iij }, min{Fija , Fijb })]
= [( , ,
2 2 ∩[(Tijc , Iij
c
, Fijc )]
Fija  Fijc + Fijb  Fijc
)] = [(min{max{Tija , Tijb }, Tijc }, max{min{Iij
a b
, Iij c
}, Iij },
2
= [(Tija ∗ Tijc , Iij
a
 Iij c
, Fija  Fijc )] max{min{Fija , Fijb }, Fijc })]
[(Tijb ∗ Tijc , Iij
b c
 Iij , Fijb  Fijc )] = [(max{min{Tija , Tijc }, min{Tijb , Tijc }}, min{max{Iij
a c
, Iij },
b c
= (A ∩ C)  (B ∩ C) max{Iij , Iij }}, min{max{Fija , Fijc }, max{Fijb , Fijc }})]
= [(min{Tija , Tijc }, max{Iij
a c
, Iij }, max{Fija , Fijc })]
In a similar way, the remaining can be established.
∪[(min{Tijb , Tijc }, max{Iij
b c
, Iij }, max{Fijb , Fijc })]
= (A ∩ C) ∪ (B ∩ C)
4.8 Proposition (Distributive law)
Let A = [(Tija , Iij a
, Fija )], B = [(Tijb , Iij
b
, Fijb )], C = (iv) Here A  (B ∪ C), (A  B) ∪ (A  C) ∈ N SMm×n . Now,
c c c
[(Tij , Iij , Fij )] ∈ N SMm×n .
A  (B ∪ C)
If a ∗ b = min{a, b} and a  b = max{a, b}, then
= [(Tija , Iij
a
, Fija )]  [(max{Tijb , Tijc }, min{Iij
b c
, Iij },
(i) A ∩ (B ∪ C) = (A ∩ B) ∪ (A ∩ C), (A ∪ B) ∩ C =
(A ∩ C) ∪ (B ∩ C). min{Fijb , Fijc })]
 
(ii) A ∪ (B ∩ C) = (A ∪ B) ∩ (A ∪ C), (A ∩ B) ∪ C = = [( Tija · max{Tijb , Tijc }, Iij a · min{I b , I c },
ij ij
(A ∪ C) ∩ (B ∪ C). 
(iii)A  (B ∪ C) = (A  B) ∪ (A  C), (A ∪ B)  C = Fija · min{Fijb , Fijc })]
(A  C) ∪ (B  C).   
A  (B ∩ C) = (A  B) ∩ (A  C), (A ∩ B)  C = = [(max{ Tija · Tijb , Tija · Tijc }, min{ Iij a · Ib ,
ij
(A  C) ∩ (B  C).   
a · I c }, min{ F a · F b ,
Iij Fija · Fijc })]
(iv) A  (B ∪ C) = (A  B) ∪ (A  C), (A ∪ B)  C = ij ij ij

Tuhin Bera, Nirmal Kumar Mahapatra, Neutrosophic Soft Matrix and its application to Decision Making
 11 Neutrosophic Sets and Systems, 18/2017

  
= [( Tija · Tijb , Iij
a · Ib , Fija · Fijb )] 2 · max{Tija , Tijb } · Tijc 2 · min{Iij
a b
, Iij c
} · Iij
ij = [( , ,
   max{Tijb , Tijc } + Tijc b , Ic } + Ic
min{Iij ij ij
∪[( Tija · Tijc , Iija · Ic , Fija · Fijc )]
ij 2 · min{Fija , Fijb } · Fijc
= (A  B) ∪ (A  C) )]
min{Fijb , Fijc } + Fijc
Next (A ∪ B)  C, (A  C) ∪ (B  C) ∈ N SMm×n . Now, 2Tija Tijc 2Tijb Tijc a c
2Iij Iij
= [(max{ a c , b c
}, min{ a c ,
Tij + Tij Tij + Tij Iij + Iij
(A ∪ B)  C b c
2Iij Iij 2Fija Fijc 2Fijb Fijc
= [(max{Tija , Tijb }, min{Iij a b
, Iij }, min{Fija , Fijb })] }, min{ , })]
b + Ic
Iij ij Fija + Fijc Fijb + Fijc
[(Tijc , Iij
c
, Fijc )]
  2Tija Tijc a c
2Iij Iij 2Fija Fijc
= [( a , a + I c , F a + F c )]
= [( max{Tija , Tijb } · Tijc , min{Iij a , Ib } · Ic ,
ij ij Tij + Tijc Iij ij ij ij

min{Fija , Fijb } · Fijc , )] 2Tijb Tijc b c
2Iij Iij 2Fijb Fijc
∪[( , , )]
   Tijb + Tijc b
Iij + c
Iij Fijb + Fijc
= [(max{ Tija · Tijc , Tijb · Tijc }, min{ Iij a · Ic ,
ij = (A  C) ∪ (B  C)
  
b · I c }, min{ F a · F c ,
Iij Fijb · Fijc })]
ij ij ij
  
= [( Tija · Tijc , Iij a · Ic ,
ij Fija · Fijc )] 4.9 Proposition (Idempotent law)
  
∪[( Tijb · Tijc , Iij b · Ic ,
ij Fijb · Fijc )] Let A = [(Tija , Iij
a
, Fija )] ∈ N SMm×n . Then,
= (A  C) ∪ (B  C) (i) A w A = A (ii) A w A = A (iii) A w A = A.

(v) Here A  (B ∪ C), (A  B) ∪ (A  C) ∈ N SMm×n . Now, Proof. For all i, j and w1 , w2 > 0 we have,
a a
w1 Tij +w2 Tij w1 I a +w2 I a w1 F a +w2 F a
A  (B ∪ C) (i) A w A = [( w1 +w2 , wij1 +w2 ij , wij1 +w2 ij , )] =
= [(Tija , Iij
a
, Fija )]  [(Tija , Iij
a
, Fija )] = A.

[(max{Tijb , Tijc }, min{Iij
b c
, Iij }, min{Fijb , Fijc })] (ii) A w A = [( (w1 +w2 ) (Tija )w1 · (Tija )w2 ,
 
2 · Tija · max{Tijb , Tijc } 2 · Iij
a b
· min{Iij c
, Iij } (w1 +w2 ) a )w1 · (I a )w2 , (w1 +w2 ) (F a )w1 · (F a )w2 )]
(Iij
= [( a b c
, a b c
, 
ij

ij ij
Tij + max{Tij , Tij } Iij + min{Iij , Iij } a a
= [( (w1 +w2 ) (Tij )w1 +w2 , (w1 +w2 ) (Iij )w1 +w2 ,
2 · Fija · min{Fijb , Fijc } 
)] (w1 +w2 )
(Fija )w1 +w2 )] = [(Tija , Iij
a
, Fija )] = A.
Fija + min{Fijb , Fijc }
2Tija Tijb 2Tija Tijc a b
2Iij Iij (iii) A w A = [( w1 +w2
w1 w
+ T a2
, w1 +w2
w1 w
+ I a2
, w1 +w2
w1 w
+ F a2
)] =
= [(max{ , a c }, min{ , Ta Ia Fa
Tija Tijb Tij + Tij a b ij ij ij ij ij ij
+ Iij + Iij [(Tija , Iij
a
, Fija )] = A.
a c
2Iij Iij 2Fija Fijb 2Fija Fijc
a + Ic }, min{ , })]
Iij ij Fija + Fijb Fija + Fijc
2Tija Tijb a b
2Iij Iij 2Fija Fijb
= [( , , )] 5 Neutrosophic soft matrix theory in de-
Tija + Tijb Iij a + Ib F a + F b
ij ij ij
2Tija Tijc a c
2Iij Iij 2Fija Fijc
cision making (score function algo-
∪[( a c , a c , F a + F c )] rithm)
Tij + Tij Iij + Iij ij ij
= (A  B) ∪ (A  C)
5.1 Definition
Next (A ∪ B)  C, (A  C) ∪ (B  C) ∈ N SMm×n . Now,
1. Let A = [aij ]m×n be an NSM where aij = (Tija , Iij a
, Fija ).
(A ∪ B)  C Then the value of the matrix A is denoted by V (A) and is defined
= [(max{Tija , Tijb }, min{Iij
a b
, Iij }, min{Fija , Fijb })] a
as : V (A) = [vij ]m×n where vij a
= Tija − Iij
a
− Fija , ∀i, j.
[(Tijc , Iij
c
, Fijc )] 2. The score of two NSMs A and B is defined as S(A, B) =
a b
[sij ]m×n where sij = vij + vij . So, S(A, B) = V (A) + V (B).
3. The total score for each object in U is Σnj=1 sij .

Tuhin Bera, Nirmal Kumar Mahapatra, Neutrosophic Soft Matrix and its application to Decision Making
 Neutrosophic Sets and Systems, 18/2017 12

5.2 Properties of Score Function N = {fN (e1 ), fN (e2 ), fN (e3 ), fN (e4 ), fN (e5 )} where

Value matrices are classical real matrices which follow all prop- fN (e1 ) = {< s1 , (0.6, 0.4, 0.5) >, < s2 , (0.7, 0.4, 0.2) >,
erties of classical real matrices. The score function is basically < s3 , (0.9, 0.4, 0.2) >}
a real matrix in classical sense derived from two or more value
fN (e2 ) = {< s1 , (0.5, 0.5, 0.6) >, < s2 , (0.8, 0.5, 0.1) >,
matrices. So score functions obey all properties of real matrices.
< s3 , (0.6, 0.7, 0.5) >}
fN (e3 ) = {< s1 , (0.7, 0.3, 0.4) >, < s2 , (0.8, 0.5, 0.3) >,
5.3 Methodology < s3 , (0.5, 0.6, 0.7) >}
Suppose, N number of decision makers wish to select an ob- fN (e4 ) = {< s1 , (0.7, 0.5, 0.3) >, < s2 , (0.6, 0.7, 0.5) >,
ject jointly from m number of objects i.e., universal set U with < s3 , (0.5, 0.5, 0.5) >}
respect to n number of features i.e., parametric set E. Each deci- fN (e5 ) = {< s1 , (0.6, 0.4, 0.6) >, < s2 , (0.6, 0.3, 0.7) >,
sion maker forms an NSS over (U, E) and corresponding to each < s3 , (0.8, 0.3, 0.3) >}}
NSS, each get an NSM of order m × n. It needs to compute the
value matrix corresponding to each matrix. Then the score matrix P = {fP (e1 ), fP (e2 ), fP (e3 ), fP (e4 ), fP (e5 )} where
and finally, the total score of each object will be calculated.
fP (e1 ) = {< s1 , (0.8, 0.3, 0.3) >, < s2 , (0.8, 0.5, 0.3) >,
< s3 , (1.0, 0.4, 0.2) >}
5.3.1 Algorithm
fP (e2 ) = {< s1 , (0.6, 0.4, 0.5) >, < s2 , (0.7, 0.6, 0.2) >,
Step 1 : Construct the NSMs from the given NSSs. < s3 , (0.8, 0.5, 0.4) >}
Step 2 : Calculate the value matrices of corresponding NSMs. fP (e3 ) = {< s1 , (0.8, 0.4, 0.1) >, < s2 , (0.7, 0.5, 0.5) >,
Step 3 : Compute the score matrix from value matrices and the
< s3 , (0.6, 0.7, 0.3) >}
total score for each object in U .
Step 4 : Find the object of maximum score and it is the optimal f (e
P 4 ) = {< s1 , (0.6, 0.6, 0.2) >, < s2 , (0.8, 0.6, 0.4) >,
solution. < s3 , (0.7, 0.3, 0.6) >}
Step 5 : If score is maximum for more than one object, then find f (e ) = {< s , (0.8, 0.4, 0.2) >, < s , (0.6, 0.4, 0.3) >,
P 5 1 2
Σnj=1 (sij )k , k ≥ 2 successively. Choose the object of maximum
< s3 , (0.7, 0.5, 0.4) >}}
score and hereby the optimal solution.
The above three NSSs are represented by the NSMs A, B and C,
respectively, as following :
5.3.2 Case study 1 (application in class room)
⎛ ⎞
Three students {s1 , s2 , s3 } from class - x in a school have been (.7, .2, .6) (.4, .6, .7) (.5, .5, .3) (.6, .6, .5) (.8, .3, .4)
shortened to win the best student award in an academic session. ⎝ (.6, .3, .5) (.7, .6, .3) (.7, .4, .4) (.5, .8, .6) (.7, .2, .6) ⎠
A team of three teachers {T1 , T2 , T3 } has been formed by the (.8, .3, .5) (.5, .5, .4) (.6, .4, .6) (.4, .7, .4) (.9, .1, .2)
Head Master of that school for this purpose. Final selection is ⎛ ⎞
(.6, .4, .5) (.5, .5, .6) (.7, .3, .4) (.7, .5, .3) (.6, .4, .6)
based on the set of parameters {e1 , e2 , e3 , e4 , e5 } indicating the ⎝
(.7, .4, .2) (.8, .5, .1) (.8, .5, .3) (.6, .7, .5) (.6, .3, .7) ⎠
quality of student, participation in school cultural programme,
(.9, .4, .2) (.6, .7, .5) (.5, .6, .7) (.5, .5, .5) (.8, .3, .3)
class room interactions, maintenance of discipline in class room, ⎛ ⎞
daily attendance, respectively. Teachers have given their valuable (.8, .3, .3) (.6, .4, .5) (.8, .4, .1) (.6, .6, .2) (.8, .4, .2)
opinions by the following NSSs separately i.e., first NSS given by ⎝ (.8, .5, .4) (.7, .6, .2) (.7, .5, .5) (.8, .6, .4) (.6, .4, .3) ⎠
first teacher and so on. (1, .4, .2) (.8, .5, .4) (.6, .7, .3) (.7, .3, .6) (.7, .5, .4)
M = {fM (e1 ), fM (e2 ), fM (e3 ), fM (e4 ), fM (e5 )} where Then the corresponding value matrices are :
⎛ ⎞
fM (e1 ) = {< s1 , (0.7, 0.2, 0.6) >, < s2 , (0.6, 0.3, 0.5) >, −.1 −.9 −.3 −.5 0.1
V (A) = ⎝ −.2 −.2 −.1 −.9 −.1 ⎠
< s3 , (0.8, 0.3, 0.5) >}
0.0 −.4 −.4 −.7 0.6
fM (e2 ) = {< s1 , (0.4, 0.6, 0.7) >, < s2 , (0.7, 0.6, 0.3) >, ⎛ ⎞
−.3 −.6 0.0 −.1 −.4
< s3 , (0.5, 0.5, 0.4) >}
V (B) = ⎝ 0.1 0.2 0.0 −.6 −.4 ⎠
fM (e3 ) = {< s1 , (0.5, 0.5, 0.3) >, < s2 , (0.7, 0.4, 0.4) >, 0.3 −.6 −.8 −.5 0.2
< s3 , (0.6, 0.4, 0.6) >} ⎛ ⎞
0.2 −.3 0.3 −.2 0.2
fM (e4 ) = {< s1 , (0.6, 0.6, 0.5) >, < s2 , (0.5, 0.8, 0.6) >, V (C) = ⎝ −.1 −.1 −.3 −.2 −.1 ⎠
< s3 , (0.4, 0.7, 0.4) >} 0.4 −.1 −.4 −.2 −.2

Tuhin Bera, Nirmal Kumar Mahapatra, Neutrosophic Soft Matrix and its application to Decision Making
 13 Neutrosophic Sets and Systems, 18/2017

⎛ ⎞
The score matrix is : (0.9, 0.4, 0.5) (0.8, 0.5, 0.5) (0.7, 0.5, 0.4)
⎜ (0.8, 0.5, 0.4) (0.9, 0.3, 0.3) (0.9, 0.3, 0.2) ⎟
⎛ ⎞ ⎜ ⎟
−0.2 −1.8 00.0 −0.8 −0.1 A=⎜
⎜ (0.7, 0.6, 0.6) (0.7, 0.6, 0.5) (0.5, 0.6, 0.7) ⎟
⎟
S(A, B, C) = ⎝ −0.2 −0.1 −0.4 −1.7 −0.6 ⎠ ⎝ (0.6, 0.4, 0.7) (0.6, 0.7, 0.8) (0.7, 0.4, 0.6) ⎠
00.7 −1.1 −1.6 −1.4 00.6 (0.5, 0.3, 0.8) (0.6, 0.8, 0.5) (0.6, 0.3, 0.4)
⎛ ⎞
⎛ ⎞ (1.0, 0.5, 0.4) (0.9, 0.4, 0.5) (0.8, 0.3, 0.2)
−2.9 ⎜ (0.9, 0.4, 0.5) (0.9, 0.2, 0.3) (0.9, 0.2, 0.1) ⎟
⎜ ⎟
and the total score = ⎝ −3.0 ⎠ B=⎜ (0.7, 0.7, 0.5) (0.8, 0.5, 0.4) (0.4, 0.5, 0.6) ⎟
⎜ ⎟
−2.8 ⎝ (0.6, 0.5, 0.3) (0.7, 0.7, 0.6) (0.5, 0.6, 0.6) ⎠
Hence, the student s3 will be selected for the best student award (0.6, 0.7, 0.4) (0.6, 0.8, 0.7) (0.7, 0.4, 0.3)
from class-x in that academic session.
Then the corresponding value matrices are :
⎛ ⎞
0.0 −.2 −.2
5.3.3 Case study 2 (application in security management) ⎜ −.1 0.3 0.4 ⎟
⎜ ⎟
V (A) = ⎜
⎜ −.5 −.4 −.8 ⎟
⎟
An important discussion on internal security management has ⎝ −.5 −.9 −.3 ⎠
been arranged by the order of Home Minister. Two officers −.6 −.7 −.1
have mate in that discussion to analyse and arrange the secu- ⎛ ⎞
rity management in five mega-cities e.g., Delhi(D), Mumbai(M), 0.1 0.0 0.3
⎜ 0.0 0.4 0.6 ⎟
Kolkata(K), Chennai(C), Bengaluru(B). The priority of manage- ⎜ ⎟
ment is given to the cities based on the set of parameters {a, b, c} V (B) = ⎜
⎜ −.5 −.1 −.7 ⎟
⎟
⎝ −.2 −.6 −.7 ⎠
indicating their geographical position(e.g., having international
boarder line, having sea coast etc ), population density, past his- −.5 −.9 0.0
tory of terrorist attack, respectively. Following NSSs refer the The score matrix and the total score for selection are :
opinions of two officers individually regarding that matter. ⎛ ⎞
00.1 −0.2 00.1
N1 = {fN1 (a), fN1 (b), fN1 (c)} where ⎜ −0.1 00.7 01.0 ⎟
⎜ ⎟
S(A, B) = ⎜ ⎜ −1.0 −.5 −1.5 ⎟
⎟
fN1 (a) = {< D, (0.9, 0.4, 0.5) >, < M, (0.8, 0.5, 0.4) >, ⎝ −0.7 −1.5 −1.0 ⎠
< K, (0.7, 0.6, 0.6) >, < C, (0.6, 0.4, 0.7) >, −1.1 −1.6 −0.1
⎛ ⎞
< B, (0.5, 0.3, 0.8) >} 00.0
fN1 (b) = {< D, (0.8, 0.5, 0.5) >, < M, (0.9, 0.3, 0.3) >, ⎜ 01.6 ⎟
⎜ ⎟
< K, (0.7, 0.6, 0.5) >, < C, (0.6, 0.7, 0.8) >, Total score = ⎜ ⎜ −3.0 ⎟
⎟
⎝ −3.2 ⎠
< B, (0.6, 0.8, 0.5) >}
−2.8
fN1 (c) = {< D, (0.7, 0.5, 0.4) >, < M, (0.9, 0.3, 0.2) >,
Hence, the priority of security management should be given in
< K, (0.5, 0.6, 0.7) >, < C, (0.7, 0.4, 0.6) >,
descending order to Mumbai, Delhi, Bangaluru, Kolkata and
< B, (0.6, 0.3, 0.4) >} Chennai.
N2 = {fN2 (a), fN2 (b), fN2 (c)} where
6 Conclusion
fN2 (a) = {< D, (1.0, 0.5, 0.4) >, < M, (0.9, 0.4, 0.5) >,
< K, (0.7, 0.7, 0.5) >, < C, (0.6, 0.5, 0.3) >, In this paper, some definitions regarding neutrosophic soft ma-
< B, (0.6, 0.7, 0.4) >} trices have been brought and some new operators have been in-
cluded, illustrated by suitable examples. Moreover, application
fN2 (b) = {< D, (0.9, 0.4, 0.5) >, < M, (0.9, 0.2, 0.3) >,
of neutrosophic soft matrix theory in decision making problems
< K, (0.8, 0.5, 0.4) >, < C, (0.7, 0.7, 0.6) >, have been made. We expect, this paper will promote the future
< B, (0.6, 0.8, 0.7) >} study on different algorithms in several other decision making
fN2 (c) = {< D, (0.8, 0.3, 0.2) >, < M, (0.9, 0.2, 0.1) >, problems.
< K, (0.4, 0.5, 0.6) >, < C, (0.5, 0.6, 0.6) >,
< B, (0.7, 0.4, 0.3) >} References
These two NSSs are represented by the NSMs A and B, respec- [1] L. A. Zadeh, Fuzzy sets, Information and control, 8, (1965),
tively, as following : 338-353.

Tuhin Bera, Nirmal Kumar Mahapatra, Neutrosophic Soft Matrix and its application to Decision Making
 Neutrosophic Sets and Systems, 18/2017 144

[2] K. Atanassov, Intuitionistic fuzzy sets, Fuzzy sets and sys- [19] F. Smarandache, Neutrosophic set, A generalisation of the
tems, 20, (1986),87-96.
[3] D. Molodtsov, Soft set theory- first results, Computer and
Mathematics with Applications, 37, (1999), 19-31.
[4] N. Cagman and S. Enginoglu, Soft set theory and uni-int de-
cision making, European J. Oper. Res., 207(2), (2010), 848-
855.
[5] P. K. Maji, R. Biswas and A. R. Roy, An application of soft
sets in a decision making problem, Comput. Math. Appl., 44,
(2002), 1077-1083.
[6] P. K. Maji, R. Biswas and A. R. Roy, Soft set theory, Comput.
Math. Appl., 45, (2003), 555-562.
intuitionistic fuzzy sets, Inter.J.Pure Appl.Math., 24, (2005),
287-297.
[20] F. Smarandache, Neutrosophy, Neutrosophic Probability,
Set and Logic, Amer. Res. Press, Rehoboth, USA., (1998),
p. 105, http://fs.gallup.unm.edu/eBook-neutrosophics4.pdf
(fourth version).
[21] P. K. Maji, Neutrosophic soft set, Annals of Fuzzy Mathe-
matics and Informatics, 5(1), (2013), 157-168.
[22] I. Deli, Interval-valued neutrosophic soft sets and its deci-
sion making, International Journal of Machine Learning and
Cybernetics, DOI: 10.1007/s13042-015-0461-3.

[7] P. K. Maji, R. Biswas and A. R. Roy, Fuzzy soft sets, J. Fuzzy [23] T. Bera and N. K. Mahapatra, On neutrosophic soft func-
Math., 9(3), (2001), 589-602. tion, Annals of fuzzy Mathematics and Informatics, 12(1),
(2016), 101-119.
[8] P. K. Maji, R. Biswas and A. R. Roy, Intuitionistic fuzzy soft
sets, J. Fuzzy Math., 9(3), (2001), 677-692. [24] I. Deli and S. Broumi, Neutrosophic soft relations and some
properties, Annals of Fuzzy Mathematics and Informatics,
[9] P. K. Maji, R. Biswas and A. R. Roy, On intuitionistic fuzzy 9(1), (2015), 169-182.
soft sets, J. Fuzzy Math., 12(3), (2004), 669-683.
[25] S. Broumi and F. Smarandache, Intuitionistic neutrosophic
[10] N. Cagman and S. Enginoglu, Soft matrix theory and it’s soft set, Journal of Information and Computing Science, 8(2),
decision making, Comput. Math. Appl., 59, (2010), 3308- (2013), 130-140.
3314.
[26] I. Deli and S. Broumi, Neutrosophic Soft Matrices and
[11] Y. Yong and J. Chenli, Fuzzy soft matrices and their appli- NSM-decision Making, Journal of Intelligent and Fuzzy Sys-
cations, part 1, LNAI, 7002, (2011), 618-627. tems, 28(5), (2015), 2233-2241.
[12] M. J. Borah, T. J. Neog and D. K. Sut, Fuzzy soft matrix [27] T. Bera and N. K. Mahapatra, Introduction to neutrosophic
theory and it’s decision making, IJMER, 2, (2012), 121-127. soft groups, Neutrosophic Sets and Systems, 13, (2016), 118-
[13] T. J. Neog and D. K. Sut, An application of fuzzy soft sets in 127, doi.org/10.5281/zenodo.570845.
decision making problems using fuzzy soft matrices, IJMA, [28] T. Bera and N. K. Mahapatra, On neutrosophic normal
(2011), 2258-2263. soft groups, Int. J. Appl. Comput. Math., 2(4), (2016), DOI
[14] S. Broumi, F. Smarandache and M. Dhar, On fuzzy soft 10.1007/s40819-016-0284-2.
matrix based on reference function, Information engineering
[29] T. Bera and N. K. Mahapatra, (α, β, γ)-cut of neutrosophic
and electronic business, 2, (2013), 52-59.
soft set and it’s application to neutrosophic soft groups,
[15] J. I. Mondal and T. K. Roy, Intuitionistic fuzzy soft matrix Asian Journal of Math. and Compt. Research, 12(3), (2016),
theory, Mathematics and statistics, 1(2), (2013), 43-49, DOI: 160-178.
10.13189/ms.2013.010205.
[30] T. Bera and N. K. Mahapatra, On neutrosophic soft rings,
[16] B. Chetia and P. K. Das, Some results of intuitionistic fuzzy OPSEARCH, 1-25, (2016), DOI 10.1007/ s12597-016-0273-
soft matrix theory, Advanced in applied science research, 6.
3(1), (2012), 412-423.
[31] T. Bera and N. K. Mahapatra, Introduction to neutrosophic
[17] T. M. Basu, N. K. Mahapatra and S. K. Mondal, Intuition- soft topological space, OPSEARCH, (March, 2017), DOI
istic fuzzy soft matrix and it’s application in decision mak- 10.1007/s12597-017-0308-7.
ing problems, Annals of fuzzy mathematics and informatics,
7(1), (2014), 109-131. [32] S. Das, S. Kumar, S. Kar and T. Pal, Group deci-
sion making using neutrosophic soft matrix : An al-
[18] P. Rajarajeswari and P. Dhanalakshmi, Intuitionistic fuzzy gorithmic approach, Journal of King Saud University
soft matrix theory and it’s application in decision making, - Computer and Information Sciences, (2017), https
IJERT, 2(4), (2013), 1100-1111. ://doi.org/10.1016/j.jksuci.2017.05.001.

Tuhin Bera, Nirmal Kumar Mahapatra, Neutrosophic Soft Matrix and its application to Decision Making
 15
[33] S. Pramanik, P. P. Dey and B. C. Giri, TOPSIS for single
valued neutrosophic soft expert set based multi-attribute de-
cision making problems, Neutrosophic Sets and Systems, 10,
(2015), 88-95.
[34] P. P. Dey, S. Pramanik and B. C. Giri, Generalized neutro-
sophic soft multi-attribute group decision making based on
TOPSIS, Critical Review 11, (2015), 41-55.
[35] S. Pramanik and S. Dalapati, GRA based multi criteria de-
cision making in generalized neutrosophic soft set environ-
ment, Global Journal of Engineering Science and Research
Management, 3(5), (2016), 153-169.
Tuhin Bera, Nirmal Kumar Mahapatra, Neutrosophic Soft Matrix and its application to Decision Making
Neutrosophic Sets and Systems, 18/2017
[36] P. P. Dey, S. Pramanik and B. C. Giri, Neutrosophic soft
multi-attribute group decision making based on grey rela-
tional analysis method, Journal of New Results in Science,
10, (2016), 25-37.
[37] P. P. Dey, S. Pramanik and B. C. Giri, Neutrosophic soft
multi-attribute decision making based on grey relational pro-
jection method, Neutrosophic Sets and Systems 11, (2016),
98-106.
Received: October 2, 2017. Accepted: October 25, 2017.