Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

Int. J. Fuzzy Syst.

https://doi.org/10.1007/s40815-020-0803-5

Evaluation of the Performance of Search and Rescue Robots Using


T-spherical Fuzzy Hamacher Aggregation Operators
Kifayat Ullah1 · Tahir Mahmood1 · Harish Garg2

Received: 26 September 2019 / Revised: 3 January 2020 / Accepted: 7 January 2020


© Taiwan Fuzzy Systems Association 2020

Abstract Multi-attribute decision-making approach is a aggregation operators have two variable parameters, namely
widely used algorithm that needs some aggregation tools and q and γ which affects the decision-making process and their
several such aggregation operators have been developed in sensitivity towards decision-making results is analyzed. A
past decades to serve the purpose. Hamacher aggregation comparative analysis of the results obtained using proposed
operator is one such operator which is based on Hamacher Hamacher aggregation operators in view of the variable
t-norm and t-conorm. It is observed that the Hamacher parameters q and γ is established to discuss any advantages
aggregation operators of intuitionistic fuzzy set, Pythagorean or disadvantages.
fuzzy set and that of picture fuzzy set has some limitations
in their applicability. To serve the purpose, in this paper, Keywords Hamacher aggregation operators · Multi-
some Hamacher aggregation operators based on T-spherical attribute decision-making · Picture fuzzy sets · Spherical
fuzzy numbers are introduced. The concepts of T-spherical fuzzy sets · T-spherical fuzzy sets
fuzzy Hamacher-weighted averaging and T-spherical fuzzy
Hamacher-weighted geometric aggregation operators are
proposed which described four aspects of human opinion 1 Introduction
including yes, no, abstinence and refusal with no limitations.
Such type of aggregation operators efficiently describes the Zadeh’s [1] concept of fuzzy set (FS) provided an opening
cases that left unsolved by the existing aggregation oper- for the scientists to deal with problems of real-life involv-
ators. The validity of the proposed aggregation operators ing uncertainty or imprecision. Atanassov [2] extended the
is examined, and some basic properties are discussed. The concept of Zadeh’s FS to propose the framework of intu-
proposed new Hamacher aggregation operators are used to itionistic fuzzy set (IFS) which takes a membership degree
analyze the performance of search and rescue robots using (MD) and non-membership degree (NMD) to describe the
a multi-attribute decision-making approach as their perfor- uncertainty of an event with a restriction that the sum of both
mance in an emergency is eminent. The proposed Hamacher MD and NMD cannot exceed 1. This concept was strength-
ened by Yager [3] by introducing the idea of Pythagorean
B Harish Garg fuzzy set (PyFS) which increases the range for assigning
harishg58iitr@gmail.com values of MD and NMD. Yager also proposed the frame-
Kifayat Ullah work of generalized orthopair FS [4] commonly known as
kifayat.phdma72@iiu.edu.pk q-rung orthopair FS (q-ROFS) which associated the duplets
Tahir Mahmood of Atanassov’s IFS with a parameter q that increases their
tahirbakhat@iiu.edu.pk range to infinity. Somehow, the pairs of IFS, PyFS and q-
1 ROFS deal with the imprecision’s that exists in real-life for
Department of Mathematics and Statistics, International
Islamic University, Islamabad 44000, Pakistan instance multi-attribute decision-making (MADM), pattern
2 recognition that involve human opinion, but these duplets
School of Mathematics, Thapar Institute of Engineering &
Technology, Deemed University, Patiala, Punjab 147004, only discuss two aspects of human opinion i.e. favor and dis-
India favor, but a human opinion has some sort of abstinence and

123
International Journal of Fuzzy Systems

refusal degrees as well. This was suggested by Cuong [5] in


2013 that the pairs representing an IFS or its generalized form
described only the MD and NMD of human opinion where
the abstinence degree (AD) and refusal degree (RD) are kind
of ignored which leads to the loss of information. For this
purpose, Cuong [5] suggested the idea of picture fuzzy set
(PFS) in the form of triplets which uses a MD, AD, NMD
and RD with a restriction that their sum must not exceed 1.
Cuong’s framework of PFS was further improved by Mah-
mood et al. [6] to spherical fuzzy set (SFS) and T-spherical
fuzzy set (TSFS) which increased the range for assigning
MD, AD and NMD. The geometrical analysis of the spaces
of PFS, SFS and TSFS is shown in Fig. 1.
MADM is one of the most widely studied and efficient
algorithms that are used whenever a set of distinct alternatives
needs to be assessed based on some attributes. In recent years,
researchers widely implemented the MADM approaches Fig. 1 Comparison of spaces of TSFNs with PFNs and SFNs
under the different environment using aggregation opera-
tors. For example, under IFS and PyFS environment, some
averaging and geometric aggregation operators are defined in tors while Gao [29] developed prioritized Pythagorean fuzzy
[7–10]. Liu and Wang [11] proposed the averaging and geo- HA operators for MADM problems. Wei [30] also proposed
metric operators for generalized orthopair FS and studied Pythagorean power HA operators by merging the concept of
their applications in MADM problems. Garg [12] and Wang power aggregation and HA operators. Further, it is observed
et al. [13] presented these operators for PFSs. However, using that these operators [28–30] can model human opinion of
Einstein t-norm operations, such aggregation operators are yes and no type and yet it has no applicability whenever we
studied by the researchers [14–17] under the different envi- have information in the form of a duplet like (0.85, 0.65)
ronment. Kaur and Garg [18] proposed cubic intuitionistic where the sum of squares of both MG and NMG exceeds
aggregation operators for MADM problems. Under T-SFS 1 i.e. 0.852 + 0.652 = 1.145  1. Recently, Darko and
information, Liu et al. [19] developed some power Muirhead Liang [31] introduced the concept of HA operators for q-
mean operators for TSFSs and studied their practicality in rung orthopair fuzzy numbers which somehow improves the
MADM. However, some improved interactive aggregation HA operators discussed in [25–30] but again this can also
operators in T-spherical fuzzy environment are proposed by model only two aspects of human opinion i.e. favor and dis-
Garg et al. [20]. Apart from it, some applications of T-SFSs favor and a human opinion has abstinence and refusal factors
using different kinds of measures and operators are summa- as well as described by Cuong [5] and Mahmood et al. [6].
rized in [19,21–23]. Due to the non-applicability of HA operators proposed
In addition to algebraic and Einstein norm based opera- in [25–31], Wei [32] proposed HA operators for in-picture
tions, Hamacher t-norm and t-conorm [24] is an important fuzzy settings. The picture fuzzy HA operators have the
class of the t-norm, which is also a generalization of algebraic ability to describe the abstinence as well as refusal degree
and Einstein norm and hence it based Hamacher aggre- of human opinion and hence there is less chance of the
gation (HA) operators are much useful than the existing information loss. Jana and Pal [33] investigated the enter-
ones. In the literature, there are several HA operators intro- prise selection problem using picture fuzzy HA operators.
duced so far and all of these have their own constraints. For So far, the chances of information loss are decreased by
instance, Huang [25] introduced the HA operators in intu- Wei [32] and Jana and Pal [33] by introducing the picture
itionistic fuzzy settings while Liu [26]proposed HA operators fuzzy HA operators but their applicability becomes a prob-
for interval-valued IFSs. Garg [27] proposed entropy-based lem when we have information in the form of triplets like
HA operators for MADM purpose. These existing operators (0.85, 0.35, 0.65) where the sum of MD, AD and NMD
[25–27] are limited in access and can only be used in intu- exceed 1 i.e. 0.852 + 0.352 + 0.652 = 1.2675  1. This
itionistic fuzzy settings i.e. such type of aggregation tools can means the HA operators proposed in [32,33] have their lim-
only model yes and no type of human opinion. Further, such itations when it comes to applicability in many cases and
kind of HA operators gets failed whenever we have informa- certain range where these operators are applicable is depicted
tion in the form of duplets like (0.85, 0.65) where the sum of in Fig. 1.
both degrees exceeds 1 i.e. 0.85+0.65 = 1.5  1. To address So far, it is observed that the theory of HA operators devel-
it, Wu and Wei [28] proposed Pythagorean fuzzy HA opera- oped for IFSs, PyFSs and q-ROFSs [25–31] and for PFSs

123
K. Ullah et al.: Evaluation of Perfomance of search and Rescue...

developed in [32,33] have their applicability problems and This Remark 1 proved the diverse and generalized nature
in some cases the information loss is more. All this leads of TSFNs over existing fuzzy frameworks. Clearly SFN has
us to develop such HA operators which can not only solve the condition 0 ≤ m 2 + i 2 + n 2 ≤ 1 due to which not
the problems discussed in [25–33] but also can be applicable every triplet can be considered as SFN e.g. (0.71, 0.52, 0.91)
in cases where the existing HA operators gets failed. There- as 0 ≤ 0.712 + 0.522 + 0.912  1. Similarly, not every
fore, in this paper, we developed the concept of T-spherical triplet can also be regarded as PFN as well. On the other
fuzzy HA operators that have the ability of processing the hand, for every triplet of the form (m, i, n), we can find a
information which the existing operators fail to process. The q ∈ Z+ such that 0 ≤ m q + i q + n q ≤ 1. In current exam-
proposed new HA operators are utilized in evaluating the ple, the triplet (0.71, 0.52, 0.91) is a TSFN for q = 5 i.e.
performance of search and rescue robots using a MADM 0 ≤ 0.715 + 0.525 ≤ + 0.915 = 0.842475 ≤ 1. Further if
approach as such robots have great significance in state of we look at Atanassov’s IFNs or Yager’s PyFNs or q-ROFNs,
emergencies and the literature shows no significance work all these frameworks cannot describe the AD in human opin-
for evaluating the performance of search and rescue robots ion and are based on MD and NMD only which caused the
based on MADM. loss of information. The 3D diagram depicted in Fig. 1 shows
The rest of the paper is organized as follows. Section 2 the comparisons of the spaces of TSFNs with SFNs and
deals with some basic definitions. In Sect. 3, we define some PFNs.
new Hamacher operations and it’s based weighted averaging The aggregation theory has drawn enough attention of
and geometric aggregation operators for TSFSs. Section 4 scientists based on t-norms and t-conorms such as algebraic,
provides a MADM approach based on the proposed oper- Einstein and Hamacher. Our goal here is to demonstrate the
ators and their viability is discussed over the evaluation of existing HA operators and their drawbacks and to develop
search and rescue robots. Section 5 provides a comparative new HA operators in T-spherical fuzzy environment:
analysis of the new and existing HA operators in the decision-
making process showing the drawbacks of the previous study. Definition 2 [24] The Hamacher t-norm Thn and t-conorm
Section 6 is about discussion and advantages of the proposed Thcn are defined as:
study while Sect. 7 summarized the article along with some
a.b
future directions. Thn (a, b) = , γ > 0,
γ + (1 − γ ) (a + b − ab)
(a, b) ∈ [0, 1]2 (1)
2 Preliminaries a + b − ab − (1 − γ ) ab
Thcn (a, b) = , γ > 0,
1 − (1 − γ ) ab
In this section, some basic notions of TSFS and terms related (a, b) ∈ [0, 1]2 (2)
to proposed work are presented.
Definition 1 [6] A triplet T = (m, i, n) is referred as a T- Further, this t-norm Thn (a, b) is also regarded as Hamacher
spherical fuzzy number (TSFN) where m, i and n denote the product while the t-conorm Thcn (a, b) is known as Hamacher
MD, AD and NMD of the TSFN T provided that for some sum.
q ∈ Z+ , 0 ≤ m √
q + i q + n q ≤ 1. The RD in this case is
Remark 2 The Hamacher product and the Hamacher sum
defined as r = q 1 − (m q + i q + n q ). defined above reduces to algebraic product and sum for
Remark 1 A TSFN is a generalized form of existing fuzzy γ = 1 and to Einstein sum and product for γ = 2
frameworks and it reduces to: respectively.

Definition 3 [31] The averaging and geometric HA opera-


• Spherical fuzzy number (SFN); by taking q as 2 (Mah-
tors for some q-ROFNs are defined as
mood et al. [6]).
• Picture fuzzy number (PFN); by taking q as 1 (Cuong
q − ROFHWA (T1 ,T2 , , . . .Tn )
[5]). ⎛ l


• q- rung orthopair fuzzy number (q-ROFN); by taking i q w j l q wj
 j=1 1+(γ −1)m j − j=1 1−m j
⎜ ⎟
⎜ l 1+(γ −1)m q w j +(γ −1) l 1−m q w j ,
q

as zero (Yager [4]). ⎟


⎜ j=1 j=1 ⎟
• Pythagorean fuzzy number (PyFN); by taking i as zero
j j
⎜ ⎟
=⎜ √ l wj ⎟ (3)
and q as 2 (Yager [3]). ⎜ q γ
j=1 n j

⎜ ⎟
• Intuitionistic fuzzy number (IFN); by taking i as zero and ⎝ q l
q

wj l q
2w j ⎠
j=1 1+(γ −1) 1−n j +(γ −1) j=1 nj
q as 1 (Atanassov [2]).
• Fuzzy number; by taking i and n as zero and q as 1 (Zadeh
[1]).

123
International Journal of Fuzzy Systems

⎛ m Am B ⎞
q − ROFHWG (T1 ,T2 , , . . .Tn ) 
q
 q q q q
,
⎜ γ +(1−γ ) m A +m B −m A m B ⎟
⎛ √ l ⎞ ⎜ ⎟
q γ
wj
⎜ q q q q q q ⎟
j=1 m j
,⎟ ⎜ q i A +i B −i A i B −(1−γ )i A i B , ⎟
⎜ q l

w j 
2w j ⎜
(ii) A ⊗ B = ⎜ q q
1−(1−γ )i A i B ⎟
⎜ j=1 1+(γ −1) 1−m j
q q
+(γ −1) lj=1 m j ⎟ ⎟
⎜ ⎟ ⎜ ⎟
⎜  ⎟ ⎜ q q q q q q ⎟
=⎜  l

q w j l

q wj
⎟ ⎝ q n A +n B −n A n B −(1−γ )n A n B ⎠
⎜  1+(γ −1)n j − j=1 1−n j ⎟ q q
1−(1−γ )n A n B
⎜ 
q j=1



⎝  l q wj  q wj ⎠ ⎛ ⎞
j=1 1+(γ −1)n j +(γ −1) lj=1 1−n j
 q λ  q λ
1+(γ −1)m A − 1−m A
⎜ q λ , ⎟
q
 
q λ 
(4) ⎜ 1+(γ −1)m A +(γ −1) 1−m A ⎟
⎜ ⎟
⎜ √ ⎟
⎜ q γ iλ ⎟
λA = ⎜  
A
   , ⎟
Definition 4 [32] The averaging and geometric HA opera- (iii) ⎜ q q λ q 2λ ⎟
1+(γ −1) 1−i A +(γ −1) i A
⎜ ⎟
tors for a some PFNs are defined as: ⎜ √ ⎟
⎜ qγ n λA ⎟
⎝   q λ  q 2λ ⎠
PFHWA (T1 ,T2 , . . . ,Tn ) q
1+(γ −1) 1−n A +(γ −1) n A
⎛ l w j l wj ⎞ ⎛ √ ⎞
j=1 (1+(γ −1)m j ) − j=1 (1−m j ) γ m λA
q
⎜ j=1 (1+(γ −1)m j )w j +(γ −1) lj=1 (1−m j )w j ,
l 
⎟ ⎜
  q λ  q 2λ , ⎟
⎜ ⎟
q
⎜ 1+(γ −1) 1−m A +(γ −1) m A ⎟
⎜  wj ⎟ ⎜  ⎟
⎜ γ lj=1 i j ⎟ ⎜  ⎟
⎜ l ⎟ q λ  q λ
= ⎜ j=1 (1+(γ −1)(1−i j )) j +(γ −1) lj=1 (i j )w j ,
w ⎟ (5) ⎜
(iv) Aλ = ⎜ q 
1+(γ −1)i A − 1−i A ⎟

⎜ ⎟ q λ  q λ ,
⎜ ⎟ ⎜ 1+(γ −1)i A +(γ −1) 1−i A ⎟
⎜  wj
⎟ ⎜ ⎟
⎝  γ lj=1 n j
⎠ ⎜  ⎟
l wj ⎜  q λ  q λ ⎟
l wj
j=1 (1+(γ −1)(1−n j )) +(γ −1) j=1 (n j ) ⎜ q 1+(γ −1)n A − 1−n A ⎟
⎝  q λ  q λ ⎠
1+(γ −1)n A +(γ −1) 1−n A

PFHWG (T1 ,T2 , . . . ,Tn ) The Hamacher operations proposed in Definition 5 are more
⎛ l wj ⎞ generalized than the existing Hamacher operations of IFSs
γ j=1 m j
⎜ l (1+(γ −1)(1−m j ))w j +(γ −1) l (m j )w j , ⎟ [25], PyFSs [28], q-ROFSs [31] and PFSs [32]. Unlike
⎜ j=1 j=1 ⎟ Hamacher operations of IFSs, PyFSs and PFSs, the T-
⎜ l w j l

⎜ wj ⎟
⎜  j=1 (1+(γ −1)i j ) − j=1 (1−i j ) ⎟ spherical fuzzy Hamacher operations describe the MD, the
= ⎜ l (1+(γ −1)i j )w j +(γ −1) l (1−i j )w j , ⎟ (6)
⎜ j=1 j=1 ⎟ AD, the NMD and the RD with no limitations. The reason is
⎜ l w j l
⎟ that for every triplet (m, i, n), there is a q ∈ Z+ that make
⎜ wj ⎟
⎝ l j=1 (1+(γ −1)iwj ) − j=1 (1−i j )
l wj ⎠ the triplet a TSFN. Further, it can be easily obtaining some
j=1 (1+(γ −1)i j ) +(γ −1) j=1 (1−i j )
j
special cases of the proposed operations given in Definition 5
and are listed as:

3 T-spherical Fuzzy Hamacher Operations and (i) If we set q = 2, then proposed operations reduced to
Aggregation Operators operations for SFNs.
(ii) If we take q = 1, then proposed operations reduced to
In this section, we define some new Hamacher operations operations for picture fuzzy numbers [32].
and it’s based weighted averaging and geometric aggregation (iii) If we take i = 0 for each TSFN, then it reduces to
operators for TSFNs. q-rung orthopair fuzzy Hamacher operations [31].
(iv) If we take q = 2 and i = 0, then it becomes
Definition 5 For two TSFNs A = (m A , i A , n A ) and B = Pythagorean fuzzy Hamacher operations [28].
(m B , i B , n B ) and for λ, γ > 0. The T-spherical fuzzy (v) If we take q = 1 and i = 0, then given operations
Hamacher operations are: reduced to intuitionistic fuzzy Hamacher operations
⎛ q q q q ⎞
q m A +m B −m A m B −(1−γ )m A m B
q q [25].
⎜ q
1−(1−γ )m A m B
q , ⎟
⎜ ⎟
⎜  i Ai B
, ⎟ 3.1 T-spherical Fuzzy Hamacher-Weighted Averaging
(i) A ⊕ B = ⎜   ⎟
⎜ q q q q q
γ +(1−γ ) i A +i B −i A i B ⎟ (TSFHWA) Operators
⎝  n AnB ⎠
q
 q q q q
γ +(1−γ ) n A +n B −n A n B
This section is based on the averaging aggregation operators
based on Hamacher operations. In our onward work, w =

123
K. Ullah et al.: Evaluation of Perfomance of search and Rescue...

(w1 , w2 , . . .
wn )T will denote the weight vector such that The result in Eq. (8) holds true for l = 2.
wi > 0 and n1 w j = 1. The terms j and k will denote the Suppose the result is true for l = k and
indexing sets such that j, k = 1, 2, 3, . . . l.
TSFHWA (T1 ,T2 , . . .Tk )
Definition 6 Let Ti = (m, i, n) be a collection of TSFNs. ⎛ ⎞
Then, TSFHWA operator is a map T n → T such that  k

q w j k

q wj
 1+(γ −1)m j − j=1 1−m j
⎜ j=1

⎜ k 1+(γ −1)m q
w j +(γ −1) k 1−m q
w j , ⎟
q

l ⎜ j=1 j j=1 j ⎟
TSFHWA (T1 ,T2 , . . .Tn ) = ⊕ w jT j (7) ⎜ ⎟
⎜ √ k wj ⎟
j=1 ⎜ q γ i ⎟

= ⎜ q k
j=1 j

w j
2w j ⎟ ,
q  q ⎟
Now, using Definition 5, we propose the following result. ⎜ j=1 1+(γ −1) 1−i j +(γ −1) kj=1 i j ⎟
⎜ ⎟
⎜ √ k wj ⎟
Theorem 1 Let Ti = (m, i, n) be a collection of TSFNs. ⎜ q γ
j=1 jn ⎟
⎝ 

w j 
2w j

q k q k q
Then, TSFHWA operator is having the form j=1 1+(γ −1) 1−n j +(γ −1) j=1 nj

TSFHWA (T1 ,T2 , . . .Tn ) Now we prove for l = k + 1


⎛ l


q w j l q wj
 1+(γ −1)m − 1−m
⎜ j=1 j j=1 j
⎟ TSFHWA (T1 ,T2 , , . . .Tk ,Tk+1 )
⎜ l 1+(γ −1)m q
w j +(γ −1) l 1−m q
w j , ⎟
q

⎜ j=1 j j=1 j ⎟
⎜ ⎟ = TSFHWA (T1 ,T2 , , . . .Tk ) ⊕ Tk+1
⎜ √ l wj ⎟
⎜ q γ
j=1 i j ⎟ ⎛ k



= ⎜ q l

w j , ⎟  q w j k q wj
l q
2w j ⎟ (8) 1+(γ −1)m j − j=1 1−m j
⎜ −1)
q
+(γ −1) ⎟ ⎜ j=1

⎜ k 1+(γ −1)m q
w j +(γ −1) k 1−m q
w j , ⎟
1+(γ 1−i i q
⎜ ⎟
j=1 j j=1 j
⎜ √ l wj ⎟ ⎜ j=1 j j=1 j ⎟
⎜ q γ
j=1 n j ⎟ ⎜ ⎟
⎝ 

w j
2w j ⎠ ⎜ √ k wj ⎟
 ⎜ q γ
j=1 i j ⎟
q l q
1+(γ −1) 1−n j +(γ −1) l q
nj ⎜
= ⎜ q k

w j
2w j ⎟ ,
j=1 j=1
q  q ⎟
⎜ j=1 1+(γ −1) 1−i j +(γ −1) kj=1 i j ⎟
⎜ ⎟
⎜ √ k wj ⎟
Proof We use mathematical induction to prove the result. ⎜ q γ
j=1 jn ⎟
For l = 2 ⎝ 

w j 
2w j

q k q k q
j=1 1+(γ −1) 1−n j +(γ −1) j=1 nj

w1T1 ⊕ w2T2 ⎛  q w  q w

⎛  ⎞ q

1+(γ −1)m k+1 k+1 − 1−m k+1 k+1
   ,
q w  q w ⎜ 1+(γ −1)m k+1 k+1 −(γ −1) 1−m k+1 k+1 ⎟
q w q w
1+(γ −1)m 1 1 − 1−m 1 1

q
 q w  q w , ⎜ ⎟
⎜ 1+(γ −1)m 1 1 +(γ −1) 1−m 1 1 ⎟ ⎟ ⎜ √ wk+1 ⎟
⎜ ⎟ ⎜ 
q γi
, ⎟
⊕⎜ ⎟
k+1

q γ i w1    
⎜ ⎟ ⎜ q
1+(γ −1) 1−i k+1 k+1 +(γ −1) i k+1 k+1 ⎟
q w q 2w
=⎜
  q 2w1 , ⎟
1
⎜  q w1 ⎟ ⎜ ⎟

q
1+(γ −1) 1−i 1 +(γ −1) i 1
⎟ ⎜ √ w
q γ n k+1 ⎟
⎜ √
q γ n w1 ⎟ ⎝  k+1 ⎠
 q wk+1  q 2wk+1
⎝  
1 ⎠ q
1+(γ −1) 1−n k+1 +(γ −1) n k+1
q q w1  q 2w1
1+(γ −1) 1−n 1 +(γ −1) n 1
⎛ ⎞ TSFHWA (T1 ,T2 , . . .Tk+1 )
 q w 
1+(γ −1)m 2 2 − 1−m 2 2
q w ⎛ k+1


,⎟ q w j k+1 q wj

q
 q w2  q w  1+(γ −1)m − 1−m
⎜ 1+(γ −1)m 2 −(γ −1) 1−m 2 2 ⎟ ⎜ j=1 j j=1 j

⎜ k+1 1+(γ −1)m q
w j +(γ −1) k+1 1−m q
w j , ⎟
q
⎜ √
q γ i w2
⎟ ⎜ ⎟
⎜  ,⎟ ⎜
j=1 j j=1 j

⊕⎜ ⎟
2
 q w2  q 2w2 ⎜ ⎟
⎜ q
1+(γ −1) 1−i 2 +(γ −1) i 2 ⎟ ⎜
√ k+1 w j
γ j=1 i j ⎟
⎜ ⎟ q
,
⎜ √q γ n w2 ⎟ =⎜

⎜ q k+1 1+(γ −1) 1−i q w j +(γ −1) k+1 i q 2w j ⎟



⎝  
2 ⎠ ⎜ ⎟
q w2  q 2w2
q
1+(γ −1) 1−n 2 +(γ −1) n 2 ⎜ j=1 j j=1 j

⎜ √ k+1 w j ⎟
⎜ q γ
j=1 n j ⎟
w1T1 ⊕ w2T2 ⎝ 

w j 
2w j

⎛  2

⎞ q k+1
j=1
q
1+(γ −1) 1−n j +(γ −1) k+1
j=1
q
nj
q w j 2 q wj
 1+(γ −1)m − 1−m
⎜ j=1 j j=1 j

⎜ 2 1+(γ −1)m q
w j +(γ −1) 2 1−m q
w j , ⎟
q

⎜ j=1 j j=1 j ⎟ Hence the result holds for l = k + 1 and, therefore, holds
⎜ ⎟

⎜ √ 2 wj ⎟ true for all values of l.
⎜ q γ j=1 i j ⎟
=⎜


, ⎟
⎜ q 2 1+(γ −1) 1−i q w j +(γ −1) 2 i q 2w j ⎟ Now, we state some basic properties of the proposed
⎜ ⎟
⎜ ⎟
j=1 j j=1 j
TSFHWA operator in the following Theorem.
⎜ √ 2 wj ⎟
⎜ q γ
j=1 n j ⎟
⎝ 

w j 
2w j
⎠ Theorem 2 The HA operator of TSFNs satisfies the follow-
q 2 q 2 q
1+(γ −1) 1−n j +(γ −1) nj
j=1 j=1 ing properties:

123
International Journal of Fuzzy Systems

(1) (Idempotency) If T j = T = (m, i, n) ∀ j = The TSFHWA operator only weighs while the TSFHOWA
1, 2, 3, . . . l. Then operator weighs the ordered position of the T-spherical fuzzy
arguments. We need to develop such operators that discuss
TSFHWA (T1 , T2 , T3 . . .Tn ) = T the ordered position as well as the argument itself. Therefore,
we propose the concept of hybrid operator that takes into
 
(2) (Boundedness)
 If T− = min j m j , max  j i j , max j n j account both the argument and its ordered position.
and T+ = max j m j , min j i j , min j n j . Then
Definition 8 Let Ti = (m, i, n) be a collection of TSFNs.
Then, TSFHHA operator is a map T n → T such that
T− ≤ TSFHWA (T1 , T2 , T3 . . .Tn ) ≤ T+
l
(3) (Monotonicity) Let T j and P j be two TSFNs such that TSFHHA (T1 ,T2 , . . .Tn ) = ⊕ w jṪσ ( j) (11)
j=1
T j ≤ P j ∀ j. Then
where Ṫσ ( j) is the jth largest of the TSFN Ṫ j = lw jT j with
TSFHWA (T1 , T2 , T3 . . .Tn ) w j as the weight vector ofT-spherical fuzzy arguments T j
≤ TSFHWA (P1 , P2 , P3 . . . Pn ) such that w j ∈ [0, 1] and n1 w j = 1 and l is the balancing
coefficient.
These can be proved analogously.
Theorem 4 Let Ti = (m, i, n) be a collection of TSFNs.
The TSFHWA aggregation operator weighs the TSFN Then, TSFHHA operator is having the form
only. In MADM problem, there are circumstances when the
ordered position of the TSFN matters. For, those situations, TSFHHA (T1 ,T2 , . . .Tn )
the concept of ordered weighted averaging operators play a ⎛ l
w j 
w j ⎞
 q
1+(γ−1)ṁ σ( j)
q
− lj=1 1−ṁ σ( j)
⎜
,⎟
j=1
significant role and TSFHOWA operator is proposed as fol- q

⎜ lj=1 1+(γ−1)ṁ qσ( j) w j +(γ−1) lj=1 1−ṁ qσ( j) w j ⎟
lows. ⎜ ⎟
⎜ l ⎟
⎜ √
q γ
wj ⎟
⎜ j=1 i σ( j)
, ⎟
Definition 7 Let Ti = (m, i, n) be a collection of TSFNs. =⎜ 

w j l
2w j ⎟
⎜ q l 1+(γ−1) 1−i q q
+(γ−1) j=1 i σ( j) ⎟
Then, TSFHOWA operator is a map T n → T such that ⎜ j=1 σ( j) ⎟
⎜  ⎟
⎜ √
q γ l
w j ⎟
⎝ j=1 ṅ σ( j) ⎠
l
l

w j 
2w j
TSFHOWA (T1 ,T2 , . . . ,Tn ) = ⊕ w jTσ ( j) (9) q
j=1
q
1+(γ−1) 1−ṅ σ( j)
q
+(γ−1) lj=1 ṅ σ( j)
j=1
(12)
where σ ( j) is such that Tσ ( j−1) ≥ Tσ ( j) ∀ j.
Remark 4 The Eq. (12) reduced to TSFHWA operator if we
Theorem 3 Let Ti = (m, i, n) be a collection of TSFNs.  T
take w j = 1l , 1l , 1l . . . 1l while it reduced to TSFHOWA
Then, TSFHOWA operator is having the form  T
operator if we take w j = 1l , 1l , 1l . . . 1l .
TSFHOWA (T1 ,T2 , . . . ,Tn ) Remark 5 The TSFHHA operator given in Eq. (12) satisfies
l the properties of, monotonicity and boundedness stated in
= ⊕ w jTσ ( j)
j=1 Theorem 2.
⎛ l
w j 
w j ⎞
 q
1+(γ −1)m σ ( j)
q
− lj=1 1−m σ ( j)
⎜
,⎟ 3.2 T-Spherical fuzzy Hamacher-weighted geometric
j=1
q

⎜ lj=1 1+(γ −1)m qσ ( j) w j +(γ −1) lj=1 1−m qσ ( j) w j ⎟ (TDFHWG) operator
⎜ ⎟
⎜ √ l wj ⎟
⎜ q γ
j=1 i σ ( j) ⎟

=⎜ 
, ⎟

w j l q
2w j ⎟ This section is based on the geometric aggregation operators
⎜ q l
1+(γ −1) 1−i
q
+(γ −1) i ⎟
⎜ j=1 σ ( j)
l
j=1 σ ( j)
⎟ based on Hamacher operations.
⎜ √q γ
wj

⎝ j=1 n σ ( j) ⎠
q l
q

w j 
q

2w j Definition 9 Let Ti = (m, i, n) be a collection of TSFNs.
1+(γ −1) 1−n σ ( j) +(γ −1) lj=1 n σ ( j)
j=1
Then, TSFHWG operator is a map T n → T such that
(10)
l
TSFHWG (T1 ,T2 , . . .Tn ) = ⊕ w jT j (13)
Remark 3 The TSFHOWA operator given in Eq. (10) satis- j=1
fies the properties of idempotency, monotonicity and bound-
edness stated in Theorem 2. Now, using Definition 5, we propose the following result.

123
K. Ullah et al.: Evaluation of Perfomance of search and Rescue...

lw
Theorem 5 Let Ti = (m, i, n) be a collection of TSFNs. where Ṫσ ( j) is the jth largest of the TSFN Ṫ j = T j j with
Then, TSFHWG operator is having the form w j as the weight vector ofT-spherical fuzzy arguments T j
such that w j ∈ [0, 1] and n1 w j = 1 and l is the balancing
TSFHWG (T1 ,T2 , , . . . ,Tn ) coefficient.
⎛ √ l wj ⎞
q γ j=1 m j

w j
2w j ,⎟ Theorem 7 Let Ti = (m, i, n) be a collection of TSFNs.
⎜ q l q  q ⎟
⎜ j=1 1+(γ −1) 1−m j +(γ −1) lj=1 m j ⎟ Then, TSFHHG operator is having the form
⎜  ⎟
⎜  l

q w j l

q wj ⎟
⎜  1+(γ −1)i − 1−i ⎟ TSFHHG (T1 ,T2 , . . .Tn )
⎜ 
q j=1

j j=1

j

, ⎟
=⎜ l q wj l q wj ⎟ ⎛ ⎞
⎜ j=1 1+(γ −1)i j +(γ −1) j=1 1−i j ⎟  l
w j 
w j
⎜  ⎟  1+(γ −1) ṁ
q
− l
1−ṁ
q
⎜ ⎟ ⎜ j=1 σ ( j) j=1 σ ( j)

 l ⎜ l 1+(γ −1)ṁ q
w j +(γ −1) l 1−ṁ q
w j , ⎟

w j 
w j q
⎜  q l q ⎟
⎜  j=1 1+(γ −1)n j − j=1 1−n j ⎟ ⎜ j=1 σ ( j) j=1 σ ( j) ⎟
⎝ q
l
w l
w ⎠ ⎜ ⎟
q j
1+(γ −1)n j +(γ −1)
q j
1−n j ⎜ √ l wj ⎟
j=1 j=1
⎜ q γ i
j=1 σ ( j) ⎟
=⎜


, ⎟
⎜ q l 1+(γ −1) 1−i q wj l q 2w j ⎟
(14) ⎜ σ ( j) +(γ −1) j=1 i σ ( j) ⎟
⎜ ⎟
j=1
⎜ √ l wj ⎟
Proof The proof of this result is similar to Theorem 1, and ⎜ q γ
j=1 ṅ σ ( j) ⎟
⎝ 

w j 
2w j

hence we omit their proof here.
 q l
j=1
q
1+(γ −1) 1−ṅ σ ( j) +(γ −1) l
j=1
q
ṅ σ ( j)

In addition, it is noted that the proposed TSFHWG oper- (18)


ator satisfies the properties as described in Theorem 2 for a
collection of TSFNs. Remark 7 The Eq. (18) reduced to TSFHWG operator if we
 T
Definition 10 Let Ti = (m, i, n) be a collection of TSFNs. take w j = 1l , 1l , 1l · · · 1l while it reduced to TSFHOWG
 T
Then, TSFHOWG operator is a map T n → T such that operator if we take w j = 1l , 1l , 1l · · · 1l .

l Apart from it, we can conclude that the aggregation operators


TSFHOWG (T1 ,T2 , . . .Tn ) = ⊕ w jTσ ( j) (15) defined in the existing environments such as IFSs, PyFSs, q-
j=1
ROFSs, PFSs and SFSs, can be reduced from the proposed
where σ ( j) is such that Tσ ( j−1) ≥ Tσ ( j) ∀ j. operators. In other words, the existing HA operators become
special cases of the proposed TSFHWA and TSFHWG oper-
Theorem 6 Let Ti = (m, i, n) be a collection of TSFNs.
ators. For instance, consider the TSFHWA and TSFHWG
Then, TSFHOWG operator is having the form
aggregation operators proposed in Eqs. (8) and (14), respec-
l tively, and then we can analyze the following special cases.
TSFHOWG (T1 ,T2 , . . . ,Tn ) = ⊕ w jTσ ( j)
j=1
⎛ ⎞ (1) If we take q = 2, in the equations of TSFHWA and
 l
w j 
w j
TSFHWG, then it reduces to results of SFNs.
 q
1+(γ −1)m σ ( j)
q
− lj=1 1−m σ ( j)
⎜ j=1

⎜ l 1+(γ −1)m q
w j +(γ −1) l 1−m q
w j , ⎟ (2) If we take q = 1, in the equations of TSFHWA and
q

⎜ j=1 σ ( j) j=1 σ ( j) ⎟
⎜ ⎟ TSFHWG, then it reduces to results of PFNs [32].
⎜ √ l wj ⎟ (3) If we take i = 0, then these equations become operators
⎜ q γ
j=1 i σ ( j) ⎟

= ⎜ q l

w j , ⎟
q l q
2w j ⎟ for q-ROFNs [31].
⎜ j=1 1+(γ −1) 1−i σ ( j) +(γ −1) j=1 i σ ( j) ⎟ (4) If we take i = 0 and q = 2, then proposed operators
⎜ ⎟
⎜ √ l wj ⎟
⎜ q γ
j=1 n σ ( j) ⎟ reduces to operators of PyFNs [28].
⎝ 

w j
2w j ⎠
q l
1+(γ −1) 1−n σ ( j)
q 
+(γ −1) l q
n σ ( j)
(5) If we take i = 0 and q = 1, then it reduces to aggrega-
j=1 j=1
tion operators of IFNs [25].
(16)

Remark 6 The TSFHOWG operator given in Eq. (16) satis- 4 Applications of the Proposed Study in
fies the properties of idempotency, monotonicity and bound- Multi-attribute Decision-Making
edness.
Definition 11 Let Ti = (m, i, n) be a collection of TSFNs. This section aims to utilize the proposed Hamacher aggre-
Then, TSFHHG operator is a map T n → T such that gation operators of TSFNs in MADM problem. MADM
is a process of selecting a best alternative among a list of
l finite alternatives using aggregation operators and similarity
TSFHHG (T1 ,T2 , . . .Tn ) = ⊕ w jṪσ ( j) (17)
j=1 or distance measures etc. Here, the interesting fact is that

123
International Journal of Fuzzy Systems

Fig. 2 Flow chart of the


proposed MADM algorithm

information about the alternative is in the form of TSFNs Step 3: This step is based on normalization of decision matrix
which not only discuss the membership and non-membership provided in Step 1 to ensure that every attribute is of benefit
grade but also the abstinence and refusal degree in uncertain type. If an attribute is of cost type, then we use the definition
environment. Let us assume the collection of alternatives be of complement of TSFN proposed by Mahmood et al. [6] to
denoted by Ak (k is finite) and the attributes based on which make all values of benefit type.
the alternatives are assessed be denoted by G j ( j is finite) Step 4: Aggregate the normalized data obtained in Step 3
under the weight vector w j . Let Dk× j = (T)k× j = (m, i, n) using either TSFHWA or TSFHWG operator given in Eqs.
be the decision matrix containing the information about the (8) and (14), respectively.
alternatives in the form of TSFNs. The MADM algorithm Step 5: Compute the score values of the aggregated TSFNs
containing proposed HA operators is demonstrated as fol- by Eq. (19)
lows.
S (T) = m q − i q − n q (19)

4.1 Algorithm Step 6: Rank the numbers based on score values and choose
the most desired one(s).
The steps of MADM algorithm using HA operators are
described below followed by a comprehensive flowchart in 4.2 Application in the Performance of Search and
Fig. 2. Rescue Robots
Step 1: This step involves information gathering form
decision-makers about the given alternatives. The decision- In this section, we utilize the proposed MADM algorithm to
makers provide their evaluation about alternatives in the form evaluate the performance of search and rescue robots.
of TSFNs keeping the view the attributes. The performance of search and rescue robots [34] is emi-
Step 2: The aim here is to compute the value of q which nent in situations of emergencies. Rescue robots took place of
makes every value in the decision matrix a TSFN. We take the responders in case of emergency as they can take photos
the least value of q for which every triplet of the decision of the scenes and could record some live videos that can help
matrix is a TSFN. a lot in understanding the intensity of the situation. These

123
K. Ullah et al.: Evaluation of Perfomance of search and Rescue...

Table 1 Decision matrix


G1 G2 G3 G4
containing opinion of experts
about robots A1 (0.64, 0.34, 0.71) (0.72, 0.45, 0.63) (0.51, 0.59, 0.68) (0.56, 0.88, 0.71)
A2 (0.74, 0.69, 0.56) (0.65, 0.59, 0.68) (0.91, 0.50, 0.63) (0.87, 0.61, 0.79)
A3 (0.54, 0.33, 0.39) (0.75, 0.54, 0.67) (0.84, 0.66, 0.70) (0.72, 0.88, 0.35)
A4 (0.77, 0.63, 0.45) (0.88, 0.60, 0.44) (0.67, 0.77, 0.55) (0.70, 0.40, 0.50)

Table 2 Aggregated values by


TSFHWA operator TSFHWG operator
TSFHWA and TSFHWG
operators A1 (0.63136459, 0.527244, 0.694886) (0.61476783, 0.70588141, 0.89609987)
A2 (0.81580694, 0.613081, 0.665947) (0.80208765, 0.62469032, 0.89583012)
A3 (0.72527687, 0.556551, 0.479825) (0.69448866, 0.71784277, 0.88174917)
A4 (0.77847638, 0.577191, 0.478821) (0.772873, 0.63365776, 0.87523554)

search and rescue robots can be helpful in caves, tunnels and


wilderness for finding victims or a potential hazard. The moti- 0.5

vation to use these search and rescue robots is their aped and 0
completeness of the task without increasing risks to victims -0.5
A1 A2 A3 A4

or rescuer.
-1
Using TSFHWA operator Using TSFHWG operator
Example 1 In this example, we consider the problem of
evaluating the performance of search and rescue robots. Fig. 3 Comparison of the ranking of alternatives using TSFHWA and
This example is adapted from [35] where the performance TSFHWG operators
of search and rescue robots was examined using intu-
itionistic fuzzy information. However, in this environment,
only two aspects of human opinion were studied which Step 4: With w = (0.32, 0.23, 0.18, 0.27 )T , the aggregated
leads to information loss as the abstinence and refusal values of each alternative are obtained using TSFHWA and
degree of human opinion are neglected. Based on the TSFHWG operator. The results of them are listed in Table 2.
existing literature [35], the attributes that have an essen- Step 5: By Eq. (19), the score value of A k s are computed
tial role in the evaluation of search and rescue robots as −0.07829324, 0.24920943, 0.17835228 and 0.26568982
include G1 ; viability, G2 ; athletic ability, G3 ; working ability corresponding to the values obtained through TSFHWA oper-
and G4 ; communication and control capability. Let the num- ator while −0.87437565, −0.44116969, −0.67101258 and
ber of search and rescue robots that needs to be assessed −0.41245345 for the values obtained through TSFHWG
be 4 denoted by A1 , A2 , A3 , A4 . The weight vector for the operator.
four attributes be w = (0.32, 0.23, 0.18, 0.27 )T . The eval- Step 6: The ranking order from these score values are
uation here involves the opinion of the experts which they obtained as A4 > A2 > A3 > A1 . Thus the search and
provide in the form of a decision matrix keeping in view the rescue robot number A4 is the most reliable one for deal-
four attributes. The opinion of the experts is expressed in the ing with emergency situations. The comparison between the
form of TSFNs describing the MD, AD, NMD and RD of different operators results are given in Fig. 3.
alternatives based on attributes. The experts from the respec-
tive field gave their opinion and a decision matrix is provided 4.3 Effect of “γ ” on ranking of alternatives
in Table 1.
In this subsection, we aim to analyze the effect of variation
We solved the problem using the proposed algorithm and the in γ on the ranking results. For this purpose, we solved the
stepwise computations are provided as: considered problem for various values of γ and the rank-
Step 1: The collective information of the expert is given in ing results are given in Table 3. From this table, it clearly
Table 1 in the form of TSFNs. indicates that varying γ have no effect on ranking in case
Step 2: From their rating, we can compute that the smallest of TSFHWA operator. On the other hand, it is observed that
q which satisfy m q + i q + n q ≤ 1 is q = 5. the ranking results get changed in case of TSFHWG operator
Step 3: All the attributes are of same nature, so there is no for γ = 105. This shows the significance of variation in γ in
need of normalizing the data. decision analysis. At this stage, as for γ ≥ 105, the results

123
International Journal of Fuzzy Systems

Table 3 Impact of γ on to the ranking results


0.3
γ TSFHWA operator TSFHWG operator
0.2
5 A4 > A2 > A3 > A1 A4 > A2 > A3 > A1
10 A4 > A2 > A3 > A1 A4 > A2 > A3 > A1 0.1

50 A4 > A2 > A3 > A1 A4 > A2 > A3 > A1 0


A1 A2 A3 A4
100 A4 > A2 > A3 > A1 A4 > A2 > A3 > A1
-0.1
105 A4 > A2 > A3 > A1 A2 > A4 > A3 > A1
Below q=6 Above q=6
500 A4 > A2 > A3 > A1 A2 > A4 > A3 > A1
Fig. 4 Ranking pattern above and below stability values in case of
TSFHWA operator

Table 4 Ranking results for various values of γ = 2 and various values


0
of “q”
-0.1 A1 A2 A3 A4

q TSFHWA operator TSFHWG operator -0.2


-0.3
5 A4 > A2 > A3 > A1 A4 > A2 > A3 > A1 -0.4

6 A2 > A4 > A3 > A1 A4 > A2 > A3 > A1 -0.5


-0.6
11 A2 > A4 > A3 > A1 A2 > A4 > A3 > A1
-0.7
50 A2 > A4 > A3 > A1 A2 > A4 > A3 > A1
-0.8
100 A2 > A4 > A3 > A1 A2 > A4 > A3 > A1 -0.9

200 A2 > A4 > A3 > A1 A2 > A4 > A3 > A1 -1


Below q=11 Above q=11

Fig. 5 Ranking pattern above and below stability values in case of


TSFHWG operator
are stable so the it must be taken as γ = 105 for consistent
results.
Table 5 Input matrix for Example 2

4.4 Effect of variation in “q” on ranking results G1 G2 G3 G4

A1 (0.64, 0.71) (0.72, 0.63) (0.51, 0.68) (0.56, 0.71)


In TSFHWA and TSFHWG operators, γ is not the only vari- A2 (0.74, 0.56) (0.65, 0.68) (0.91, 0.63) (0.87, 0.79)
able constant but the value of “q” also plays an important A3 (0.54, 0.39) (0.75, 0.67) (0.84, 0.70) (0.72, 0.35)
role. Here we, analyze the impact of “q” on ranking results.
A4 (0.77, 0.45) (0.88, 0.44) (0.67, 0.55) (0.70, 0.50)
In the considered problem, if we vary the value of “q” from 5
onwards, then the final ranking order of the given alternative
is shown in Table 4. From it, we see that the ranking pattern by dropping the abstinence degree from each triplet of the
gets changed at q = 6 in case of TSFHWA operator while the decision matrix.
ranking pattern gets changed at q = 11 in case of TSFHWG
operators. However, the ranking pattern after q = 11 remains Example 2 Revisit the Example 1 and drop the abstinence
the same in both cases no matter how high we raise the value grade from the decision matrix provided in Table 1. The
of “q”. This shows a stability in ranking results at q = 11. updated decision matrix is summarized in Table 5.
The ranking pattern for below and above the stability points
Clearly, seen that all the duplets are q-ROFNs for q = 4.
in case of TSFHWA and TSFHWG operator is geometrically
Now, by taking q-ROFHWA and q-ROFHWG operators [31],
shown in Figs. 4 and 5, respectively.
we aggregate the information and results are summarized in
Table 6.
Based on their computed score values, we can obtain the
5 A Comparative of the Results Obtained Using final ranking order of the alternatives as A4 > A2 > A3 >
Proposed and Existing Methods A1 and hence get the most optimal one is A4 . From this
computed results, we obtained the best alternative is differ-
The goal of this section is to establish a comparison of the ent from the earlier obtained using TSFHWA and TSFHWG
proposed work with that of existing literature to demonstrate operators. This shows that by dropping the abstinence grade,
the advantages of the new proposed work over existing the- some information loss accrued which causes the change
ory. For it, we considered the numerical example as q-ROFS in ranking pattern. Therefore, the use of TSFHWA and

123
K. Ullah et al.: Evaluation of Perfomance of search and Rescue...

Table 6 Aggregated information using q-ROFHWA and q-ROFHWG cussed in Example 2 where the AD of the T-spherical fuzzy
operators information is dropped, and we observed the consequences
q-ROFHWA operator q-ROFHWG operator in the results. Here, it is worth mentioning that the framework
of PFS also allows you to describe the four aspects of human
A1 (0.62759479, 0.701401) (0.60888807, 0.9862327)
opinion but in a limited range. Therefore, the more conve-
A2 (0.81273797, 0.672338) (0.77880727, 0.98623275)
nient option is to use T-spherical fuzzy framework. Some key
A3 (0.71992669, 0.482812) (0.68152261, 0.9862327) advantages of the proposed work are stated below:
A4 (0.77591318, 0.480725) (0.75470923, 0.9862327)

Table 7 Ranking of alternatives using existing and proposed methods (a) T-spherical fuzzy HA operators generalizes existing HA
Aggregation operator Ref. Ranking operators describing the abstinence and refusal degree of
human opinion with unlimited range while the existing
TSFHWA This Paper A 4 > A2 > A3 > A1 HA operators either leads to the loss of information or
TSFHWG This Paper A 4 > A2 > A3 > A1 have limited range.
TSFWG Mahmood et al. [6] A4 > A2 > A3 > A1 (b) The existing HA operators becomes special cases of the
TSFWA Ullah et al. [36] A2 > A3 > A3 > A1 proposed T-spherical fuzzy HA operators.
TSFWG Ullah et al. [36] A4 > A2 > A3 > A1 (c) T-spherical fuzzy HA operators can solve the problems
TSFWGIA Garg et al. [20] A4 > A3 > A1 > A2 studied in existing literature while the existing HA opera-
TSFPMM Liu et al. [19] A4 > A3 > A1 > A2 tors of IFSs, PyFSs, q-ROFSs and PFSs cannot take over
the problems discussed in T-spherical fuzzy environment.
(d) The results obtained using T-spherical fuzzy HA oper-
TSFHWG operators is eminent for modeling of human opin- ators are stable than those obtained using the existing
ion without any loss of information. literature. This is because the proposed HA operators
have two variable parameters which are responsible for
Example 3 In this example, the information of the decision- changing the ranking of alternatives but provides some
makers about the alternatives in Example 1 is processed by stable results at a certain stage as discussed in Example 1.
several existing aggregation tools to assess the validity of the
proposed results. Here, it is important to note that the existing
HA operators of IFSs [25–27], PyFSs [28–30], q-ROFSs [31]
and PFSs [32] cannot solve the data and information which 7 Conclusion
are present in Table 1. Now, we use the aggregation operators
proposed by Mahmood et al. [6], Ullah et al. [36] and Garg In this paper, we proposed the concept of Hamacher opera-
et al. [20], Liu et al. [19] on the data provided in Table 1 to tions for TSFNs based on Hamacher t-norm and Hamacher
compare the results. The ranking results obtained using the t-conorm. Based on the newly developed operations, the
stated existing aggregation operators are listed in Table 7. It concepts of Hamacher averaging and Hamacher geometric
is evident that the results obtained using proposed TSFHWA aggregation operators of TSFNs are also proposed and their
and TSFHWG operators are consistent as these are much basic and essential properties are examined, and their fitness
similar to the ranking results obtained by Garg et al. [20] and validity are investigated. Keeping in view the importance
and Liu et al. [19]. This must be noted that the aggregation of the ordered position and argument itself, the concepts of
operators proposed by Garg et al. [20] is an improved version Hamacher ordered weighted averaging and geometric and
of the aggregation operators proposed by Mahmood et al. [6] Hamacher hybrid aggregation operators are also developed.
and Ullah et al. [36]. We briefly investigated the consequences of the proposed
aggregation operators and described some certain conditions
under which the proposed operations reduced to different
existing fuzzy environments i.e. the existing HA operators
6 Discussion and Advantages of the Proposed become special cases of proposed HA operators. To demon-
Work strate the significance of proposed work, a MADM algorithm
is developed in view of HA operators of TSFNs and a numer-
Work in T-spherical fuzzy environment allows you to discuss ical example is solved to observe the viability of new HA
four aspects of human opinion, namely MD, AD, NMD and operators and decision-making algorithm. Further, we ana-
RD unlike intuitionistic, Pythagorean and q-rung orthopair lyze the effects of the variable parameters q and γ in the
fuzzy environments. Hence, the chance of losing information decision making process and described the stability stage of
in T-spherical fuzzy environment will always be much lesser ranking results. A comparative analysis, of the ranking of
than other fuzzy environments. One such example is dis- alternatives obtained using proposed and existing aggrega-

123
International Journal of Fuzzy Systems

tion operators is established to signify the importance of the 16. Garg, H.: Generalized Pythagorean fuzzy geometric aggregation
proposed work. operators using Einstein t-norm and t-conorm for multicriteria
decision-making process. Int. J. Intell. Syst. 32(6), 597–630 (2017)
In the near future, we aim to extend the idea to the inter- 17. Garg, H.: A new generalized Pythagorean fuzzy information aggre-
val valued numbers and apply to other diverse fields such as gation using Einstein operations and its application to decision
investment banking, pattern recognition, supplier selection making. Int. J. Intell. Syst. 31(4), 886–920 (2016)
problems, etc using diverse fuzzy environment [37–39]. One 18. Kaur, G., Garg, H.: Generalized cubic intuitionistic fuzzy aggre-
gation operators using t-norm operations and their applications to
can also extend the approach to solve some healthcare prob- group decision-making process. Arab. J. Sci. Eng. 44(3), 2775–
lems [40] etc. Apart from this, we will develop a software 2794 (2019)
prototype to automate the decision making process so that 19. Liu, P., Khan, Q., Mahmood, T., Hassan, N.: T-spherical fuzzy
the proposed approach can be performed by managers easily power Muirhead mean operator based on novel operational laws
and their application in multi-attribute group decision making.
in emergency situations. IEEE Access 7, 22613–22632 (2019)
20. Garg, H., Munir, M., Ullah, K., Mahmood, T., Jan, N.: Algorithm
Compliance with Ethical Standards for T-spherical fuzzy multi-attribute decision making based on
improved interactive aggregation operators. Symmetry 10(12), 670
Conflicts of Interest The authors declare that they have no conflict of (2018). https://doi.org/10.3390/sym10120670
interests. 21. Ullah, K., Mahmood, T., Jan, N.: Similarity measures for T-
spherical fuzzy sets with applications in pattern recognition. Sym-
Ethical Approval This article does not contain any studies with human
metry 10(6), 193 (2018). https://doi.org/10.3390/sym10060193
participants or animals performed by any of the authors.
22. Quek, S.G., Selvachandran, G., Munir, M., Mahmood, T., Ullah,
K., Son, L.H., Thong, P.H., Kumar, R., Priyadarshini, I.: Multi-
attribute multi-perception decision-making based on generalized
T-spherical fuzzy weighted aggregation operators on neutro-
References sophic sets. Mathematics 7(4), 780 (2019). https://doi.org/10.3390/
math7090780
1. Zadeh, L.A.: Fuzzy sets. Inf. Control 8(3), 338–353 (1965) 23. Ullah, K., Garg, H., Mahmood, T., Jan, N., Ali, Z.: Correlation
2. Atanassov, K.T.: Intuitionistic fuzzy sets. Fuzzy Sets Syst. 20(1), coefficients for T-spherical fuzzy sets and their applications in clus-
87–96 (1986) tering and multi-attribute decision making. Soft Comput 24(3),
3. Yager, R.R.: Pythagorean fuzzy subsets. In: IFSA world congress 1647–1659 (2020). https://doi.org/10.1007/s00500-019-03993-6
and NAFIPS annual meeting (IFSA/NAFIPS), 2013 Joint. IEEE. 24. Oussalah, M.: On the use of Hamacher’s t-norms family for infor-
(2013) https://doi.org/10.1109/IFSA-NAFIPS.2013.6608375 mation aggregation. Inf. Sci. 153, 107–154 (2003)
4. Yager, R.R.: Generalized orthopair fuzzy sets. IEEE Trans. Fuzzy 25. Huang, J.Y.: Intuitionistic fuzzy Hamacher aggregation operators
Syst. 25(5), 1222–1230 (2017) and their application to multiple attribute decision making. J. Intell.
5. Cu’ ò’ ng, B.C.: Picture fuzzy sets. J. Comput. Sci. Cybern. 30(4), Fuzzy Syst. 27(1), 505–513 (2014)
409–420 (2014) 26. Liu, P.: Some Hamacher aggregation operators based on the
6. Mahmood, T., Ullah, K., Khan, Q., Jan, N.: An approach towards interval-valued intuitionistic fuzzy numbers and their application
decision making and medical diagnosis problems using the con- to group decision making. IEEE Trans. Fuzzy Syst. 22(1), 83–97
cept of spherical fuzzy sets. Neural Comput. Appl. 31, 7041–7053 (2013)
(2019) 27. Garg, H.: Intuitionistic fuzzy hamacher aggregation operators with
7. Xu, Z.: Intuitionistic fuzzy aggregation operators. IEEE Trans. entropy weight and their applications to multi-criteria decision-
Fuzzy Syst. 15(6), 1179–1187 (2007) making problems. Iran. J. Sci. Technol. Trans. Electr. Eng. 43(3),
8. Xu, Z., Yager, R.R.: Some geometric aggregation operators based 597–613 (2019)
on intuitionistic fuzzy sets. Int. J. Gen. Syst. 35(4), 417–433 (2006) 28. Wu, S.J., Wei, G.W.: Pythagorean fuzzy Hamacher aggregation
9. Garg, H.: Novel intuitionistic fuzzy decision making method based operators and their application to multiple attribute decision mak-
on an improved operation laws and its application. Eng. Appl. Artif. ing. Int. J. Knowl. Based Intell. Eng. Syst. 21(3), 189–201 (2017)
Intell. 60, 164–174 (2017) 29. Gao, H.: Pythagorean fuzzy hamacher prioritized aggregation oper-
10. Peng, X., Yang, Y.: Some results for Pythagorean fuzzy sets. Int. J. ators in multiple attribute decision making. J. Intell. Fuzzy Syst.
Intell. Syst. 30(5), 1133–1160 (2015) 35(2), 2229–2245 (2018)
11. Liu, P., Wang, P.: Some q-rung orthopair fuzzy aggregation oper- 30. Wei, G.W.: Pythagorean fuzzy Hamacher power aggregation oper-
ators and their applications to multiple-attribute decision making. ators in multiple attribute decision making. Fundam. Inf. 166(1),
Int. J. Intell. Syst. 33(2), 259–280 (2018) 57–85 (2019)
12. Garg, H.: Some picture fuzzy aggregation operators and their appli- 31. Darko, A.P., Liang, D.: Some q-rung orthopair fuzzy Hamacher
cations to multicriteria decision-making. Arab. J. Sci. Eng. 42(12), aggregation operators and their application to multiple attribute
5275–5290 (2017) group decision making with modified EDAS method. Eng. Appl.
13. Wang, C., Zhou, X., Tu, H., Tao, S.: Some geometric aggrega- Artif. Intell. 87, 103259 (2020)
tion operators based on picture fuzzy sets and their application in 32. Wei, G.: Picture fuzzy Hamacher aggregation operators and their
multiple attribute decision making. Italian J. Pure Appl. Math. 37, application to multiple attribute decision making. Fundam. Inf.
477–492 (2017) 157(3), 271–320 (2018)
14. Wang, W., Liu, X.: Intuitionistic fuzzy geometric aggregation 33. Jana, C., Pal, M.: Assessment of enterprise performance based on
operators based on Einstein operations. Int. J. Intell. Syst. 26(5), picture fuzzy hamacher aggregation operators. Symmetry 11(1),
1049–1075 (2011) 75 (2019)
15. Zhao, X., Wei, G.: Some intuitionistic fuzzy Einstein hybrid aggre- 34. Murphy, R.R., Tadokoro, S., Nardi, D., Jacoff, A., Fiorini, P.,
gation operators and their application to multiple attribute decision Choset, H., Erkmen, A.M.: Search and rescue robotics. Springer
making. Knowl. Based Syst. 37, 472–479 (2013)

123
K. Ullah et al.: Evaluation of Perfomance of search and Rescue...

handbook of Robotics pp. 1151–1173 (2008). https://doi.org/10. Tahir Mahmood was born in
1007/978-3-540-30301-5_51 1975 in Attock, Pakistan. He did
35. Zhou, J., Baležentis, T., Streimikiene, D.: Normalized weighted his Ph.D. in Mathematics from
Bonferroni Harmonic mean-based intuitionistic fuzzy operators Quaid-i-Azam University Islam-
and their application to the sustainable selection of search and res- abad, Pakistan in 2012. Nowa-
cue robots. Symmetry 11(2), 218 (2019). https://doi.org/10.3390/ days, he is working in Interna-
sym11020218 tional Islamic University Islam-
36. Ullah, K., Hassan, N., Mahmood, T., Jan, N., Hassan, M.: Eval- abad Pakistan in the Department
uation of investment policy based on multi-attribute decision- of Mathematics and Statistics as
making using interval valued T-spherical fuzzy aggregation Assistant Professor. His area of
operators. Symmetry 11(3), 357 (2019). https://doi.org/10.3390/ research is soft sets, Algebraic
sym11030357 and Fuzzy Algebraic Structures.
37. Garg, H., Kumar, K.: A novel possibility measure to interval-valued So far, he has produced 41 mas-
intuitionistic fuzzy set using connection number of set pair analysis ter students, 01 PhD student and
and their applications. Neural Comput. Appl. (2019). https://doi. has published more than 100
org/10.1007/s00521-019-04291-w research articles in well-reputed journals.
38. Wang, L., Garg, H., Li, N.: Interval-valued q-rung orthopair 2-tuple
linguistic aggregation operators and their applications to decision Harish Garg is working as
making process. IEEE Access 7(1), 131962–131977 (2019) an Assistant Professor (Senior
39. Garg, H.: Nancy, linguistic single-valued neutrosophic power Grade) with School of Math-
aggregation operators and their applications to group decision- ematics at Thapar Institute of
making problems. IEEE CAA J. Autom. Sin. (2019). https://doi. Engineering & Technology
org/10.1109/JAS.2019.1911522 (Deemed University) Patiala,
40. Garg, H., Kaur, G.: Quantifying gesture information in brain hem- India. He has authored over 215+
orrhage patients using probabilistic dual hesitant fuzzy sets with papers published in refereed
unknown probability information. Comput. Ind. Eng. 140, 106211 International Journals includ-
(2020). https://doi.org/10.1016/j.cie.2019.106211 ing Information Sciences, IEEE
TRANSACTIONS ON FUZZY
SYSTEMS, International Jour-
nal of Intelligent Systems,
Kifayat Ullah received his bach- Cognitive Computation, Artifi-
elor and master’s degrees in cial Intelligence Review, Applied
mathematics from International Soft Computing, Experts Systems with Applications, IEEE ACCESS,
Islamic University, Islamabad, International Journal of Fuzzy Systems, Expert Systems, Journal of
Pakistan. Currently, Kifayat Ullah Manufacturing Systems, Applied Mathematics & Computations, ISA
is a research fellow at Ghent Transactions, IEEE/CAA JOURNAL OF AUTOMATIC SINICA, IEEE
University, Belgium and a PhD TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL
scholar at the Department of INTELLIGENCE, Applied Intelligence, Computer and Industrial Engi-
Mathematics and Statistics, Inter- neering, Soft Computing, Computer and Operations Research, Journal
national Islamic University, Islam- of Experimental & Theoretical Artificial Intelligence, International
abad. His research interests Journal of Uncertainty, Fuzziness and Knowledge-based Systems, Jour-
include generalizations of fuzzy nal of Industrial and Management Optimization, International Journal of
sets and intuitionistic fuzzy sets, Uncertainty Quantification, and many more. His research interests are in
multi-attribute decision making, the fields of computational intelligence, multicriteria decision-making,
pattern recognition, cluster anal- fuzzy decision-making, Pythagorean fuzzy sets, computing with words
ysis and medical diagnosis. He also did some work on fuzzy graph and soft computing. Dr. Garg is the Associate Editor for JIFS, MPE,
theory and its applications. Kifayat Ullah has published more than 20 IJCIS, TEDE, JIMO, CAIE, Complexity, CAAI Transactions on Intel-
research articles in well-reputed journals. ligence Technology and so on. His Google citations are over 6640+. For
more details, visit http://sites.google.com/site/harishg58iitr/.

123

You might also like