SS symmetry

Article
Pythagorean Fuzzy Dombi Aggregation Operators
and Their Application in Decision Support System
Arshad Ahmad Khan 1 , Shahzaib Ashraf 2 , Saleem Abdullah 2 , Muhammad Qiyas 2 ,
Jianchao Luo 1, * and Sufyan Ullah Khan 1
1 College of Economics and Management, Northwest A& F University, Yangling, Shaanxi 712100, China;
arshadkhan@nwafu.edu.cn (A.A.K.), sufyan@nwafu.edu.cn (S.U.K.)
2 Department of Mathematics, Abdul Wali Khan University, Mardan 23200, Pakistan;
shahzaibashraf@awkum.edu.pk (S.A.); saleemabdullah@awkum.edu.pk (S.A.);
muhammadqiyas@awkum.edu.pk (M.Q.)
* Correspondence: jianchaoluo2000@yahoo.com




Received: 7 January 2019; Accepted: 8 March 2019; Published: 15 March 2019


Abstract: Keeping in mind the importance and well growing Pythagorean fuzzy sets, in this paper,
some novel operators for Pythagorean fuzzy sets and their properties are demonstrated. In this paper,
we develop a comprehensive model to tackle decision-making problems where strong points of view
are in the favour and against the some projects, entities or plans. Therefore, a new approach, based
on Pythagorean fuzzy set models by means of Pythagorean fuzzy Dombi aggregation operators is
proposed. An approach to deal with decision-making problems using Pythagorean Dombi averaging
and Dombi geometric aggregation operators is established. This model has a stronger capability than
existing averaging, geometric, Einstein, logarithmic averaging and logarithmic geometric aggregation
operators for Pythagorean fuzzy information. Finally, the proposed method is demonstrated through
an example of how the proposed method helps us and is effective in decision-making problems.

Keywords: Pythagorean fuzzy sets; Dombi operations; Pythagorean fuzzy Dombi aggregation
operators; decision support system

1. Introduction
This universe is loaded with qualms, imprecision and unclearness. In reality, the greater part of the
ideas we encounter in daily life are more unclear than exact. Managing with qualm or uncertainty is a
noteworthy issue in numerous territories—for example, economics, engineering, natural science,
medicinal science and sociology. Recently, many authors have become keen on demonstrating
unclearness. Traditional speculations like fuzzy sets [1], intuitionistic fuzzy sets [2], and Pythagorean
fuzzy sets [3] are notable and assume vital jobs in demonstrating uncertainty.
Notion of fuzzy sets introduced by Zadeh [1] revolutionized not only mathematics and logic but
also science and technology. It is a very nice tool to handle uncertainty. Here, some membership grade
is assigned to an object of a fuzzy sets. In many situations in the real world, apart from the grade of
membership, the grade of non-membership is also required. To handle such conditions, Atanassov
in [2] initiates the notion of intuitionistic fuzzy sets (IFSs), which are a significant improvement on fuzzy
sets. In IFSs, the sum of membership grade and non-membership grade of an object is always from the
unit interval. However, the fascinating scenario emerges when the membership and non-membership
of an object is given from the unit interval, but their sum exceeds. Ordinary IFSs fail to handle such
situations. Therefore, a more comprehensive model is required for such situations.
Yager enquired about this scenario in [3,4] and improved the concept of IFSs to Pythagorean fuzzy
sets (PFSs), which could be considered as a generalization of IFSs. The main difference between IFSs

Symmetry 2019, 11, 383; doi:10.3390/sym11030383 www.mdpi.com/journal/symmetry

Symmetry 2019, 11, 383 2 of 19

and PFSs is that, in IFSs, the sum of membership and non-membership is always from unit closed
interval, but, in PFSs, the sum of squares of membership grade and non-membership grade are real
numbers between 0 and 1.
After their pioneering work, Ren et al. [5] launched the TODIM approach to find the best
alternative in decision-making problems using PF information. Wei [6] developed the novel
aggregation operators namely, interaction weighted averaging and weighted geometric aggregation
operators. Wei and Lu [7] introduced the power aggregation operators to deal with PF information
in decision-making problems. Xu et al. [8] proposed the induced generalized OWA operators for
Pythagorean fuzzy information. Xue et al. [9] launched the LINMAP method to track the best
investment company in railway projects using Pythagorean fuzzy information. Yager [3] proposed
the weighted averaging, geometric, ordered averaging and ordered geometric aggregation operators
for Pythagorean fuzzy information. Rahman et al. [10] proposed the Pythagorean fuzzy aggregation
operators using Einstein t-norms and t-conorms. Zhang and Xu [11] presented the TOPSIS approach
for ranking the Pythagorean fuzzy numbers to deal with decision-making problems.
In 1982, Dombi [12] defined Dombi triangular-norm and Dombi triangular-conorm operations,
which have it made the preference of variability with the operation of parameters. For this advantage,
Liu et al. [13] used Dombi operations to intuitionistic fuzzy sets and developed multiple attribute
group decision-making problem using a Dombi Bonferroni mean operator under the intuitionistic
fuzzy information. Chen and Ye [14] proposed a multiple attribute decision-making problem utilizing
Dombi aggregations operators in the single-valued neutrosophic information. Shi et al. [15] extend
Dombi operations to neutrosophic cubic sets and use it for travel decision-making problems. Lu and
Ye [16] firstly defined a Dombi aggregation operator for linguistic cubic variables, and a multiple
attribute decision-making (MADM) method is developed in linguistic cubic setting. He [17] introduced
Typhoon disaster assessment based on Dombi hesitant fuzzy information aggregation operators.
Jana et al. [18] exhibit some aggregation operators under picture fuzzy data for assessing the distinct
priorities of the choices amid the decision-making process. Jana et al. [19] defined some bipolar fuzzy
Dombi aggregation operators on the basis of traditional arithmetic, geometric operations and Dombi
operations. Wei and Wei [20] presented some combination of operations of prioritized aggregation
operators and Dombi operations of SVNNs that consider the prioritized relationship between the
SVNNs and proposed some single-valued neutrosophic Dombi prioritized weighted aggregating
operators for the aggregation of SVNNs and also investigate their properties.
Motivated by the above discussion, we propose the novel aggregation operators for Pythagorean
fuzzy sets using Dombi t-norm and Dombi t-conorm. In the decision-making process, the aggregation
operators play the vital role in aggregating the fuzzy information. Thus, in this manner, we propose a
series of novel aggregation operators, namely, Dombi weighted average, Dombi weighted geometric,
Dombi ordered weighted average, Dombi ordered weighted geometic, Dombi hybrid weighted average
and Dombi hybrid weighted geometric aggregation operators for Pythagorean fuzzy information.
After that, launch the algorithm to deal with the decision-making problems based on the proposed
Dombi aggregation operators. A numerical example demonstrates how our proposed technique helps
and is effective in decision-making problems using Pythagorean fuzzy information.
The rest of this study is designed as follows. Section 2 briefly introduces the basic knowledge
of the extension of fuzzy sets. The novel Dombi aggregations operators are presented in Section 3.
Section 4 presents some discussions on the application of the proposed method. Section 5 discussed
the advantages of the proposed work and essential conclusions.

2. Preliminaries
This section consists of a brief review of norms, fuzzy sets and their generalization such as
intuitionistic fuzzy sets, Pythagorean fuzzy sets and their basic properties and results.
Symmetry 2019, 11, 383 3 of 19

Definition 1 ([21,22]). A mapping T b : Θ × Θ → Θ is said to be triangular-norm if, for each element, T

b
satisfies that
(1) T
b is commutative, monotonic and associative,
b (v∗ , 1) = v∗ , each v∗ ∈ T,
(2) T b
where Θ = [0, 1] is the unit interval.

Definition 2 ([21,22]). A mapping Sb : Θ × Θ → Θ is said to be triangular-conorm if, for each element, Sb

satisfies that
(1) Sb is commutative, monotonic and associative,
(2) Sb (v∗ , 0) = v∗ , each v∗ ∈ S,
b
where Θ = [0, 1] is the unit interval.

Now, the following is a list of different types of t-norm and t-conorm with generators.

Name t-norm Additive Generators

 
Algebraic T
b d, e ce = de ec t (ν) = − log ν
 A
de
t (ν) = log 2−
e ce = ec ν
Einstein T
bE d,
1+(1−de)(1−ce) ν
bH (d, ce) = de
ec γ+(1−γ)ν
Hamacher T
γ+(1−γ)(de+ce−de ec) , γ > 0 t (ν) = log ν ,γ >0
  !
γ −1 γ d −1
de e
γ = 1, t (ν) = − log ν
Frank Tε (d, ce) = log 1+
γ 6= 1, t (ν) = − log γγν−−11
b
γ γ −1

Name t-conorm Additive Generators

 
Algebraic SbA d,e ce = de+ ce − de ec s (ν) = − log (1 − ν)
 
d + c
Einstein SbE d, e ce = s (ν) = log 11+ ν
e e
1+de ec −ν
  e+ce−de
ec−(1−γ)de
d ec γ+(1−γ)(1−ν)
Hamacher S H d, ce =
b e ,γ > 0 s (ν) = log (1− ν )
,γ >0
1−(1−γ ) de
ec
 !
  γ 1− d − 1 γ 1− d − 1
e e
γ = 1, s (ν) = − log (1 − ν)
e ce = 1 − log
Frank Sbε d, γ 1+ γ −1 γ 6= 1, s (ν) = − log γ1γ−−ν −
1
1

Definition 3 ([2]). Let us consider a universal set E. An intuitionistic fuzzy set R on a set E consists of two
mappings, which are defined as:
R = {h Pθ (ε) , Nθ (ε)i |ε ∈ E} ,
such that the mappings Pθ : E → Θ and Nθ : E → Θ represent the positive and negative grades ε ∈ E to the set
R, and Θ = [0, 1] is the unit interval. Having the condition that 0 ≤ Pθ (ε) + Nθ (ε) ≤ 1, for all ε ∈ E, then
R is said to be an intuitionistic fuzzy set in E.

Definition 4 ([3]). Let us consider a universal set E. A Pythagorean fuzzy set C on a set E consists of two
mappings which are defined as:
C = {h Pθ (ε) , Nθ (ε)i |ε ∈ E} ,
such that the mappings Pθ : E → Θ and Nθ : E → Θ represent the positive and negative grades ε ∈ E to the set
C, and Θ = [0, 1] is the unit interval. Having the condition that 0 ≤ Pθ2 (ε) + Nθ2 (ε) ≤ 1, for all ε ∈ E, then
C is said to be the
q Pythagorean fuzzy set in E.
2 2


χθ (ε) = 1 − Pθ (ε) + Nθ (ε) is known to be a hesitancy degree of ε ∈ E to the set C.

Yager [3] proposed the basic operations of the Pythagorean fuzzy set as follows:

Definition 5 ([3]). For any two PFNs, C1 = Pθ1 (ε) , Nθ1 (ε) and C2 = Pθ2 (ε) , Nθ2 (ε) in E. The union,
intersection and compliment proposed as:
Symmetry 2019, 11, 383 4 of 19

(1) C1 ⊆ C2 i f f ∀ε ∈ E, Pθ1 (ε) ≤ Pθ2 (ε) and Nθ1 (ε) ≥ Nθ2 (ε);
(2) C1 = C2 i f f C1 ⊆ C2 and C2 ⊆ C1 ;



(3) C1 ∪ C2 = max Pθ1 , Pθ2 , min Nθ1 , Nθ2 ;



(4) C1 ∩ C2 = min Pθ1 , Pθ2 , max Nθ1 , Nθ2 ;
(5) C1c = Nθ1 , Pθ1 .

Definition 6 ([3]). For any two PFNs, C1 = Pθ1 (ε) , Nθ1 (ε) and C2 = Pθ2 (ε) , Nθ2 (ε) in E and
β ≥ 0.Then, thenq
operations of PFNs are proposed as o
(1) C1 ⊕ C2 = Pθ2 + Pθ2 − Pθ2 · Pθ2 , Nθ1 · Nθ2 ;
1 2 1 2
nq o
(2) β · C1 = 2
1 − (1 − Pθ ) , ( Nθ1 ) β ;
β
n q1 o
(3) C1 ⊗ C2 = Pθ1 · Pθ2 , Nθ2 + Nθ2 − Nθ2 · Nθ2 ;
n q 1
o2 1 2
β 2
(4) C1 = ( Pθ1 ) , 1 − (1 − Nθ ) .
β β
1

Yager [3] introduced some properties of the operational laws of Pythagorean fuzzy sets are
as follows:

Theorem 1. For any three SFNs C1 = Pθ1 (ε) , Nθ1 (ε) , C2 = Pθ2 (ε) , Nθ2 (ε) and C3 =
Pθ3 (ε) , Nθ3 (ε) in E and β 1 , β 2 ≥ 0. Then,
(1) C1 ⊕ C2 = C2 ⊕ C1 ;
(2) C1 ⊗ C2 = C2 ⊗ C1 ;
(3) β 1 (C1 ⊕ C2 ) = β 1 C1 ⊕ β 1 C2 , β 1 > 0;
β β
(4) (C1 ⊗ C2 ) β1 = C1 1 ⊗ C2 1 , β 1 > 0;
(5) β 1 C1 ⊕ β 2 C1 = ( β 1 + β 2 )C1 , β 1 > 0, β 2 > 0;
β β (β +β )
(6) C1 1 ⊗ C1 2 = C1 1 2 , β 1 > 0, β 2 > 0;
(7) (C1 ⊕ C2 ) ⊕ C3 = C1 ⊕ (C2 ⊕ C3 );
(8) (C1 ⊗ C2 ) ⊗ C3 = C1 ⊗ (C2 ⊗ C3 ).
D E
Definition 7 ([3]). For any PFN C p = Pθ p (ε) , Nθ p (ε) in E. Then, score and accuracy values are defined as
(1) Se(C p ) = P2 − N 2 ∈ [0, 1]
θp θp
e(C p ) = P2 + N 2 ∈ [0, 1]
(2) A θp θp

D E
Definition 8. For any PFNs C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2) in E. Then, comparison technique
proposed as,
(1) If Se(C1 ) < Se(C2 ) then C1 < C2 ,
(2) If Se(C1 ) > Se(C2 ) then C1 > C2 ,
(3) If Se(C1 ) = Se(C2 ) then
(a) Ae(C1 ) < A e(C2 ) then C1 < C2 ,
(b) Ae(C1 ) > A e(C2 ) then C1 > C2 ,
(c) A(C1 ) = A(C2 ) then C1 ≈ C2 .
e e

Garg [23,27] proposed that Pythagorean fuzzy aggregation operators are as follows:
D E
Definition 9 ([23]). For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. The structure
of Pythagorean fuzzy weighted averaging (PFWA) operator is
n
PFWA (C1 , C2 , ..., Cn ) = ∑ β pCp,
p =1
Symmetry 2019, 11, 383 5 of 19

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0 and ∑np=1 β p = 1.

D E
Definition 10 ([23]). For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E.
The structure of Pythagorean fuzzy order weighted averaging (PFOWA) operator is
n
PFOWA (C1 , C2 , ..., Cn ) = ∑ β p Cθ ω ( p ) ,
p =1

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0, ∑np=1 β p = 1 and pth biggest weighted value is
Cθω( p) consequently by total order Cθω(1) ≥ Cθω(2) ≥ ... ≥ Cθω(n) .
D E
Definition 11 ([23]). For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E.
The structure of Pythagorean fuzzy hybrid weighted averaging (PFHWA) operator is
n
PFHWA (C1 , C2 , ..., Cn ) = ∑ β p C∗θω( p) ,
p =1

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0, ∑np=1 β p = 1 and pth biggest weighted value is
 
C∗θ C∗θ = nβ p Cθω( p) , P ∈ N consequently by total order C∗θ ≥ C∗θ ≥ ... ≥ C∗θ . In addition,
ω ( p) ω ( p) ω (1) ω (2) ω (n)
associated weights are ω = (ω1 , ω2 , ..., ωn ) with ω p ≥ 0, Σnp=1 ω p = 1.
D E
Definition 12 ([23]). For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E.
The structure of Pythagorean fuzzy weighted geometric (PFWG) operator is
n
∏

β p
PFWG (C1 , C2 , ..., Cn ) = Cp ,
p =1

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0 and ∑np=1 β p = 1.

D E
Definition 13 ([23]). For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E.
The structure of Pythagorean fuzzy order weighted geometric (PFOWG) operator is
n  β p
PFOWG (C1 , C2 , ..., Cn ) = ∏ Cθ ω ( p ) ,
p =1

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0, ∑np=1 β p = 1 and pth biggest weighted value is
Cθω( p) consequently by total order Cθω(1) ≥ Cθω(2) ≥ ... ≥ Cθω(n) .
D E
Definition 14 ([23]). For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E.
The structure of Pythagorean fuzzy hybrid weighted geometric (PFHWG) operator is
n  β p
PFHWG (C1 , C2 , ..., Cn ) = ∏ C∗θω( p) ,
p =1

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0, ∑np=1 β p = 1 and pth biggest weighted value is
 
C∗θ C∗θ = nβ p Cθω( p) , P ∈ N consequently by total order C∗θ ≥ C∗θ ≥ ... ≥ C∗θ . In addition,
ω ( p) ω ( p) ω (1) ω (2) ω (n)
associated weights are ω = (ω1 , ω2 , ..., ωn ) with ω p ≥ 0, Σnp=1 ω p = 1.
Symmetry 2019, 11, 383 6 of 19

Definition 15 ([12]). Suppose that ( g, d) ∈ (0, 1) × (0, 1) are any real numbers with β ≥ 1. Then, Dombi
norms are defined as
b ( g, d) = 1
T , (1)
 β   β  β1
1− g 1− d
1+ g + d

1
Sb ( g, d) = 1 − . (2)
 β   β  β1
g d
1+ 1− g + 1− d

Example 1. Suppose that we take g = 0.7, d = 0.3 and β = 3. Then,

b ( g, d) = 1
T = 0.2995,
 3  3  13
1−0.7 1−0.3
1+ 0.7 + 0.3

1
Sb ( g, d) = 1 − o 1 = 0.7005.
0.7 3 0.3 3
n

3
1+ 1−0.7 + 1−0.3

3. Pythagorean Fuzzy Dombi Operators

Now, we propose novel Pythagorean fuzzy Dombi basic operations based on Definition 6.

Definition 16. For any two PFNs, C1 = Pθ1 (ε) , Nθ1 (ε) and C2 = Pθ2 (ε) , Nθ2 (ε) in E and
β ≥ 0. Then, the
 operations of PFNs based on Dombi operation are introduced as 
 
v 
1 1
u 
(1) C1 ⊕ C2 = 
u 1− !2β  β1
,v ;
!2β  β1
 !2β u 
u   u  !2β 
t
1+ 1 −1 + 1 −1
u  
 1 −1 1 −1
t1+ +

P2 P2
u
 θ1 θ2   N2 N2 
θ1 θ2
 
 
v 
1 1
u 
(2) β · C1 = 
u 1− !2β  β1
,v ;
!2β  β1
 u 
u   u  
t
1+ β 1 −1
u  
 1 −1
t1+ β

P2
u
 θ1   N2 
θ1
 
 
 v 
1 1
 
(3) C1 ⊗ C2 =  , u1 − ;
u
vu  !2β  1
!2β β u
 !2β !2β  β1 
u   
u   t
1+ 1 −1 + 1 −1
1 −1 1 −1
t1+ +

N2 N2
u
 P2 P2   θ1 θ2 
θ1 θ2
 
 
 v 
β 1 1
 
(4) C1 =  , u1 − .
u
vu  !2β  β1 u
 !2β  β1 
u   
u   t
1+ β 1 −1
1 −1
t1+ β

N2
u
 P2   θ1 
θ1

3.1. Pythagorean Fuzzy Dombi Weighted Averaging Operators

Based on the defined Dombi operators for PFNs, we defined the following weighted averaging
aggregation operators.
Symmetry 2019, 11, 383 7 of 19

D E
Definition 17. For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. The structure of
Pythagorean fuzzy Dombi weighted averaging (PFDWA) operator is
n
PFDWA (C1 , C2 , ..., Cn ) = ∑ β pCp,
p =1

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0 and ∑np=1 β p = 1.

D E
Theorem 2. For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. Then, the structure
of Pythagorean fuzzy Dombi weighted averaging (PFDWA) operator is defined using Dombi operations with
σ > 0;  v 
1
u1 −  1 ,
u
 u 
! 2σ  σ 
1+ ∑np=1 β p 1 −1
 t 
 P 2 
  θp  
PFDWA (C1 , C2 , ..., Cn ) = 

v 1
,
 (3)
 u 
u 
!2σ σ 1 
 u   
n 1
t1+ ∑ p=1 β p −1
 u 
N2
 θp 

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0 and ∑np=1 β p = 1.

Proof. Using mathematical induction to prove (3), we therefore proceed as:

(a) For n = 2, since  v 
1
u1 −  1 ,
u
 u 
! 2σ  σ 
 t 1+ β 1 1 −1 
  P 2  
 θ1 
β 1 C1 = 
 v 1


 u 
u 
!2σ σ 1 
 u   
 1 −1
t1+ β 1
u 
 N2 
θ1

and  v 
1
u1 − ,
u
 u
 !2σ  σ1 
 
 t 1+ β 2 1 −1

  P2  
 θ2 
β 2 C2 = 
 v 1


!2σ  σ1
 u  
 u 
u  
 1 −1
t1+ β 2
u 
 N2 
θ2
Symmetry 2019, 11, 383 8 of 19

then,

PFDWA (C1 , C2 )
= β 1 C1 + β 2 C2
 v 
1
u1 −  ,
u
 u !2σ !2σ  σ1 
 
 t 1+ β 1 1 −1 + β2 1 −1

  P2 P2  
 θ1 θ2 
= 
 v 1


!2σ  σ1
 u  
 u !2σ 
u  
 1 −1 1 −1
t1+ β 1 + β2
u 
 N2 N2 
θ1 θ2
 v 
1
u1 − ,
u
 u
 !2σ  σ1 
 
1+ ∑2p=1 β p 1 −1
 t 
 P2 
  θp  
= 
 v 1
.

!2σ  σ1
 u  
 u 
u  
t1+ ∑2p=1 1 −1
 u 
βp
N2
 θp 

(b) Now, (3) is true for n = k,

 v 
1
u1 − ,
u
 u
 !2σ  σ1 
 
1+ ∑kp=1 β p 1 −1
 t 
 P2 

  θp  
PFDWA C1 , C2 , ..., Ĕũk = 
 v 1
,

!2σ  σ1
 u  
 u 
u  
t1+ ∑kp=1 1 −1
 u 
βp
N2
 θp 

(c) Now, we prove that (3) for n = k + 1, that is, PFDWA C1 , C2 , ..., Ck , Ck+1 = ∑kp=1 β k Ck +

β k +1 Ck +1


PFDWA C1 , C2 , ..., Ck+1
 v   v 
1 1
u1 −  1 , u1 − ,
u u
 u ! 2σ  σ   u  !2σ  σ1 
  
1+ ∑kp=1 β p 1 −1 1 −1
 t   t
1 + β k +1

 P 2   P2 
  θp     θ k +1  
= 
 v 1
+
  v 1


 u 
u 
!2σ σ 1   u 
u !2σ  σ1

 u     u   
k 1 1
t1+ ∑ p=1 β p −1 −1
  
t1+ β k+1
u u 
N2 N2
 θp   θ k +1 
 v 
1
u1 − ,
u
 u
 !2σ  σ1 
 
1+ ∑kp+ 1 1 −1
 t βp

  =1 P2 
 θp  
= 
 v 1
.

!2σ  σ1
 u  
 u 
u  
t1+ ∑kp+ 1 1 −1
 u 
βp
 =1 N2
θp 
Symmetry 2019, 11, 383 9 of 19

Thus, (3) is true for n = z + 1. Hence, it satisfies for all n. Therefore,

 v 
1
u1 − ,
u
 u
 !2σ  σ1 
 
1+ ∑np=1 β p 1 −1
 t 
 P2 
  θp  
PFDWA (C1 , C2 , ..., Cn ) = 
 v 1
,

!2σ  σ1
 u  
 u 
u  
t1+ ∑np=1 1 −1
 u 
βp
N2
 θp 

which completed the proof.

Properties: PFDWA operator that satisfies some properties are enlisted below:
D E
(1) Idempotency: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. Then, if the
D E
collection of PFNs C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) are identical, which is,

PFDWA (C1 , C2 , ..., Cn ) = C.

D E
(2) Boundedness: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. C−
p =
D E D E
+
min p Pθ p , max p Nθ p and C p = max p Pθ p , min p Nθ p ( p = 1, 2, ..., n) in E; therefore,

C− +
p ⊆ PFDWA (C1 , C2 , ..., Cn ) ⊆ C p .
D E
(3) Monotonicity: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. If C p ⊆ C∗p
for ( p = 1, 2, ..., n) , then

PFDWA (C1 , C2 , ..., Cn ) ⊆ PFDWA (C1∗ , C2∗ , ..., C∗n ) .

D E
Definition 18. For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. The structure of
Pythagorean fuzzy Dombi order weighted averaging (PFDOWA) operator is
n
PFDOWA (C1 , C2 , ..., Cn ) = ∑ β p Cθ ω ( p ) ,
p =1

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0 and ∑np=1 β p = 1 and pth biggest weighted value is
Cθω( p) consequently by total order Cθω(1) ≥ Cθω(2) ≥ ... ≥ Cθω(n) .
D E
Theorem 3. For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. Then, the structure
of Pythagorean fuzzy Dombi order weighted averaging (PFDOWA) operator is defined using Dombi operations
with σ > 0;  v 
1
u1 −  ,
u
  1
2σ σ 
 u  
 u   
 t 1+ ∑np=1 β p  2 1 −1 
  P  
 θ 
 ω ( P) 
PFDOWA (C1 , C2 , ..., Cn ) = 

1
,
 (4)
v
 u 
u  
2σ σ1 
 u   
 u  n 1
 
u1+ ∑ −1
p =1 β p
 
2
 
t  N 
 θ 
ω ( P)

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0 and ∑np=1 β p = 1 and pth biggest weighted value is
Cθω( p) consequently by total order Cθω(1) ≥ Cθω(2) ≥ ... ≥ Cθω(n) .
Symmetry 2019, 11, 383 10 of 19

Proof. The procedure is similar to Theorem 2.

Properties: PFDOWA operator satisfies some properties are enlisted below;

D E
(1) Idempotency: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. Then, if the
D E
collection of PFNs C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) are identical, which is

PFDOWA (C1 , C2 , ..., Cn ) = C.

D E
(2) Boundedness: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. C−
p =
D E D E
+
min p Pθ p , max p Nθ p and C p = max p Pθ p , min p Nθ p ( p = 1, 2, ..., n) in E, therefore

C− +
p ⊆ PFDOWA (C1 , C2 , ..., Cn ) ⊆ C p .
D E
(3) Monotonicity: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. If C p ⊆ C∗p
for ( p = 1, 2, ..., n) , then

PFDOWA (C1 , C2 , ..., Cn ) ⊆ PFDOWA (C1∗ , C2∗ , ..., C∗n ) .

D E
Definition 19. For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. The structure of
Pythagorean fuzzy Dombi hybrid weighted averaging (PFDHWA) operator is
n
PFDHWA (C1 , C2 , ..., Cn ) = ∑ β p C∗θω( p) ,
p =1

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0, ∑np=1 β p = 1 and pth biggest weighted value is
 
C∗θ C∗θ = nβ p Cθω( p) , P ∈ N consequently by total order C∗θ ≥ C∗θ ≥ ... ≥ C∗θ . In addition,
ω ( p) ω ( p) ω (1) ω (2) ω (n)
associated weights are ω = (ω1 , ω2 , ..., ωn ) with ω p ≥ 0, Σnp=1 ω p = 1.
D E
Theorem 4. For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. Then, the structure
of Pythagorean fuzzy Dombi hybrid weighted averaging (PFDHWA) operator is defined using Dombi operations
with σ > 0;
 v 
u1 −  1
1 ,
u
 u   2σ σ
 
 u 
 
 
  
 u 
n

1
 
1 + ∑ − 1
 
 u p = 1 β p  ! 2  
 t
P∗

   
 
   
 θ 
ω ( P)
PFDHWA (C1 , C2 , ..., Cn ) =  ,
 
 v 1 
 u   2σ  σ 1 
 u 
u   

 u 
u      
n 1
u1 +  ∑ p =1 β p 
 
!2 −1 
u  
 
N∗
t     
 
 θ 
ω ( P)

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0, ∑np=1 β p = 1 and pth biggest weighted value is
 
C∗θ C∗θ = nβ p Cθω( p) , P ∈ N consequently by total order C∗θ ≥ C∗θ ≥ ... ≥ C∗θ . In addition,
ω ( p) ω ( p) ω (1) ω (2) ω (n)
associated weights are ω = (ω1 , ω2 , ..., ωn ) with ω p ≥ 0, Σnp=1 ω p = 1.

Proof. The procedure is similar to Theorem 2.

Properties: PFDHWA operator satisfies some properties that are enlisted below;
Symmetry 2019, 11, 383 11 of 19

D E
(1) Idempotency: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. Then, if
D E
collection of PFNs C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) are identical, which is

PFDHWA (C1 , C2 , ..., Cn ) = C.

D E
(2) Boundedness: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. C−
p =
D E D E
min p Pθ p , max p Nθ p and C+
p = max p Pθ p , min p Nθ p ( p = 1, 2, ..., n ) in E, therefore

C− +
p ⊆ PFDHWA (C1 , C2 , ..., Cn ) ⊆ C p .
D E
(3) Monotonicity: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. If C p ⊆ C∗p
for ( p = 1, 2, ..., n) , then

PFDHWA Ĕũ1 , C2 , ..., Cn ⊆ PFDHWA (C1∗ , C2∗ , ..., C∗n ) .

3.2. Pythagorean Fuzzy Dombi Weighted Geometric Operators

Based on the defined Dombi operators for PFNs, we defined the following weighted geometric
aggregation operators.
D E
Definition 20. For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. The structure of
Pythagorean fuzzy Dombi weighted geometric (PFDWG) operator is
n
∏

β p
PFDWG (C1 , C2 , ..., Cn ) = Cp ,
p =1

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0 and ∑np=1 β p = 1.

D E
Theorem 5. For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. Then, the structure
of Pythagorean fuzzy Dombi weighted geometric (PFDWG) operator is defined using Dombi operations with
σ > 0;  1 
v ,
u 
u 
!2σ σ1
 u   
n 1 −1
∑ p =1 p
 
+
u
t1 β 2
P
 
  θp  
PFDWG (C1 , C2 , ..., Cn ) =  v , (5)
 u1 − 
 u 1 
 1
!2σ σ 

 u  
n 1
1 + ∑ p =1 β p
 t 
−1
N2
 θp 

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0 and ∑np=1 β p = 1.

Proof. Using mathematical induction to prove (3), we therefore proceed as:

(a) For n = 2, since  1 
v ,
u 
u 
!2σ σ1
 u   
 u
+ β 1 −1 
t1 1 2
P
  

 θ1 
β 1 C1 =  v 
 u1 − 
 u 1 
 1
!2σ σ 

 u  
1
 t 
1+ β 1 −1
 N2 
θ1
Symmetry 2019, 11, 383 12 of 19

and  1 
v ,
!2σ  σ1
u 
 u 
u  
 1 −1

t1+ β 1
u
P2
  

 θ2 
β 2 C2 =  v ,
1
u1 −
 u 
!2σ  σ1
  
 u   
1
 
−1
t
1+ β 1
 N2 
θ2

then,

PFDWG (C1 , C2 )
= (C1 ) β1 + (C2 ) β2
 
1
v ,
 u 
u  !2σ !2σ  σ1 

1 1
u
−1 −1
 t1+ β 1 
+ β2
P2 P2
 
  θ1 θ2  
=  v
1

u1 −
 u 

 u  !2σ !2σ  σ1 

 
1 1
1+ β 1 −1 + β2 −1
 t 
N2 N2
 θ1 θ2 
 
1
v ,
 u 
u  !2σ  σ1 

t1+ ∑2p=1 β p 1
u
−1
 
P2
 
  θp  
=  v
1
.
u1 −
 u 

 u  !2σ  σ1 

 
1+ ∑2p=1 β p 1
−1
 t 
N2
 θp 

(b) Now, (3) is true for n = k,

 1 
v ,
!2σ  σ1
u 
 u 
u  
t1+ ∑kp=1 β p 1 −1
 u 
P2
 
  θp  
PFDWG (C1 , C2 , ..., Ck ) =  v .
1
u1 −
 u 
!2σ  σ1
  
 u   
1+ ∑kp=1 1
 
−1
t
βp
N2
 θp 

(c) Now, we prove that (3) is true for n = k + 1, which is PFDWG C1 , C2 , ..., Ck , Ck+1 =
k
∏ (Ck ) k + (Ck+1 ) k+1
β β
p =1

PFDWG C1 , C2 , ..., Ck+1
   
1 1
v , v ,
 u 
u  !2σ  σ1   u 
u  !2σ  σ1 
 
t1+ ∑kp=1 β p 1 1
u u
−1 −1
   t 1 + β k +1 
P2 P2
   
  θp     θ k +1  
=  v
1
+ v
1

u1 − u1 −
 u   u 

 u  !2σ  σ1  
  u  !2σ  σ1 

   
1+ ∑kp=1 β p 1
−1 1 + β k +1 1
−1
 t   t 
N2 N2
 θp   θ k +1 
 
1
v ,
 u 
u  !2σ  σ1 

t 1 + ∑ k +1 β p 1
u
−1
 
 p =1 P2
 
 θp  
=  v
1
.
u1 −
 u 

 u  !2σ  σ1 

 
1+ ∑kp+ 1 1
−1
 t 
βp
 =1 N2
θp 
Symmetry 2019, 11, 383 13 of 19

Thus, (5) is true for n = z + 1. Hence, it satisfies all n. Therefore,

 1 
v ,
!2σ  σ1
u 
 u 
u  
t1+ ∑np=1 β p 1 −1
 u 
P2
 
  θp  
PFDWG (C1 , C2 , ..., Cn ) =  v ,
1
u1 −
 u 
!2σ  σ1
  
 u   
1+ ∑np=1 1
 
−1
t
βp
N2
 θp 

which completed the proof.

Properties: PFDWG operator satisfies some properties are enlisted below;

D E
(1) Idempotency: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. Then, if
D E
collection of PFNs C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) are identical, which is

PFDWG (C1 , C2 , ..., Cn ) = C.

D E
(2) Boundedness: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. C−
p =
D E D E
+
min p Pθ p , max p Nθ p and C p = max p Pθ p , min p Nθ p ( p = 1, 2, ..., n) in E, therefore

C− +
p ⊆ PFDWG (C1 , C2 , ..., Cn ) ⊆ C p .
D E
(3) Monotonicity: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. If C p ⊆ C∗p
for ( p = 1, 2, ..., n) , then

PFDWG (C1 , C2 , ..., Cn ) ⊆ PFDWG (C1∗ , C2∗ , ..., C∗n ) .

D E
Definition 21. For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. The structure of
Pythagorean fuzzy Dombi order weighted geometric (PFDOWG) operator is
n  β p
PFDOWG (C1 , C2 , ..., Cn ) = ∏ Cθ ω ( p ) ,
p =1

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0 and ∑np=1 β p = 1 and pth biggest weighted value is
Cθω( p) consequently by total order Cθω(1) ≥ Cθω(2) ≥ ... ≥ Cθω(n) .
D E
Theorem 6. For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. Then, the structure
of Pythagorean fuzzy Dombi order weighted geometric (PFDOWG) operator is defined using Dombi operations
with σ > 0;
1
 
v ,
u 
u  
2σ σ 1
 u    
  
u1+ ∑ n β p  1 −1
u
 t  p =1

 P2  
 θ 
 ω ( P) 
PFDOWG (C1 , C2 , ..., Cn ) =  v
 ,
 u1 − 
u 1 
2σ  σ1 

 u 
 u 
 
 
 t 1+ ∑np=1 β p  1 −1 
 N2 
 θ 
ω ( P)

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0 and ∑np=1 β p = 1 and pth biggest weighted value is
Cθω( p) consequently by total order Cθω(1) ≥ Cθω(2) ≥ ... ≥ Cθω(n) .
Symmetry 2019, 11, 383 14 of 19

Proof. The procedure is similar to Theorem 5.

Properties: The PFDOWG operator satisfies some properties that are listed below;
D E
(1) Idempotency: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. Then, if
D E
collection of PFNs C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) are identical, which is

PFDOWG (C1 , C2 , ..., Cn ) = C.

D E
(2) Boundedness: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. C−
p =
D E D E
+
min p Pθ p , max p Nθ p and C p = max p Pθ p , min p Nθ p ( p = 1, 2, ..., n) in E, therefore

C− +
p ⊆ PFDOWG (C1 , C2 , ..., Cn ) ⊆ C p .
D E
(3) Monotonicity: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. If C p ⊆ C∗p
for ( p = 1, 2, ..., n) , then

PFDOWG (C1 , C2 , ..., Cn ) ⊆ PFDOWG (C1∗ , C2∗ , ..., C∗n ) .

D E
Definition 22. For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. The structure of
Pythagorean fuzzy Dombi hybrid weighted geometric (PFDHWG) operator is
n  β p
PFDHWG (C1 , C2 , ..., Cn ) = ∏ C∗θω( p) ,
p =1

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0, ∑np=1 β p = 1 and pth biggest weighted value is
 
C∗θ C∗θ = nβ p Cθω( p) , P ∈ N consequently by total order C∗θ ≥ C∗θ ≥ ... ≥ C∗θ . In addition,
ω ( p) ω ( p) ω (1) ω (2) ω (n)
associated weights are ω = (ω1 , ω2 , ..., ωn ) with ω p ≥ 0, Σnp=1 ω p = 1.
D E
Theorem 7. For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. Then, the structure
of Pythagorean fuzzy Dombi hybrid weighted geometric (PFDHWG) operator is defined using Dombi operations
with σ > 0;
 1 
v ,
u 
u  2σ  σ1
 u   
 u   
u     
 u1+ ∑ n β  1 − 1
 
 u  p =1 p 
 ∗
!2 
 

 t  P

 
  θ  
ω ( P)
PFDHWG (C1 , C2 , ..., Cn ) =  , (6)
 v 
 u1 −
u 1 
 u   2σ  σ1 
 u 
 
 
 u 
   
 
1+ ∑np=1 1
!2 −1 
   
 u
t βp 
N∗

   

 
 θ 
ω ( P)

where β p ( p = 1, 2, ..., n) are weight vectors with β p ≥ 0, ∑np=1 β p = 1 and pth biggest weighted value is
 
C∗θ C∗θ = nβ p Cθω( p) , P ∈ N consequently by total order C∗θ ≥ C∗θ ≥ ... ≥ C∗θ . In addition,
ω ( p) ω ( p) ω (1) ω (2) ω (n)
associated weights are ω = (ω1 , ω2 , ..., ωn ) with ω p ≥ 0, Σnp=1 ω p = 1.

Proof. The procedure is similar to Theorem 5.

Properties: The PFDHWG operator satisfies some properties that are listed below;
Symmetry 2019, 11, 383 15 of 19

D E
(1) Idempotency: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. Then, the
D E
collection of PFNs C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) is identical, which is

PFDHWG (C1 , C2 , ..., Cn ) = C.

D E
(2) Boundedness: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. C−
p =
D E D E
min p Pθ p , max p Nθ p and C+
p = max p Pθ p , min p Nθ p ( p = 1, 2, ..., n ) in E, therefore

C− +
p ⊆ PFDHWG (C1 , C2 , ..., Cn ) ⊆ C p .
D E
(3) Monotonicity: For any collection of PFNs, C p = Pθ p (ε) , Nθ p (ε) ( p = 1, 2, ..., n) in E. If C p ⊆ C∗p
for ( p = 1, 2, ..., n) , then

PFDHWG (C1 , C2 , ..., Cn ) ⊆ PFDHWG (C1∗ , C2∗ , ..., C∗n ) .

4. Algorithm for Multi-Attribute Decision-Making Using Pythagorean Fuzzy Information

In this section, a novel approach for decision problems using Pythagorean fuzzy information
is proposed. In this approach, the decision makers give the information in the form of Pythagorean
fuzzy sets.
Let H = (h1 , h2 , ..., hm ) be a distinct set of m probable alternatives and ce = (ce1 , ce2 , ..., cen ) be a
finite set of n criteria, where hi indicates the ith alternatives and cej indicates the jth criteria. Let
D = (d1 , d2 , ..., dt ) be a finite set of t experts, where dk indicates the kth expert. The expert dk supplies
her appraisal of an alternative hi on an attribute cej as a PFN(i = 1, 2, ..., m; j = 1, 2, ..., n). The expert
h i
(s)
information is represented by the Pythagorean fuzzy set decision-making matrix D s = Eip .
m×n
n
Assume that β p ( p = 1, 2, ..., m) is weight vector of the attribute cej such that 0 ≤ β p ≤ 1, ∑ β p = 1 and
p =1
n
ψ = (ψ1 , ψ2 , ..., ψm ) is the weight vector of the decision makers dk such that ψk ≤ 1, ∑ ψk = 1.
h i k =1
s (s)
We construct the Pythagorean fuzzy decision-making matrices, D = Eip for decisions. Basically,
m×n
criteria have two types: one is benefit criteria and the other
h one
i is cost criteria. If the Pythagorean fuzzy
decision matrices have cost type criteria matrices, D s = Eip
s can be converted into the normalized
m×n (
s , for benefit criteria A
Eip p
h i
(s)
Pythagorean fuzzy decision matrices, Rs = rip s
, where rip = s j=
m×n Eip , for cost criteria A p ,
s s . If all the criteria have the same type, then there is no need
1, 2, ..., n, and Eip is the complement of Eip
for normalization.
Step 1: In this step, we get the collective Pythagorean information and using proposed Dombi operators
to evolute the alternative preference values with associated weights, which are ω = (ω1 , ω2 , ..., ωn )
with ω p ≥ 0, Σnp=1 ω p = 1.
Step 2: We find the score value Se(C p ) and the accuracy value A e(C p ) of the cumulative overall
preference value hi (i = 1, 2, ..., m).
Step 3: By the definition, rank the alternatives hi (i = 1, 2, ..., m) and choose the best alternative which
has the maximum score value.

4.1. Numerical Example

Assume that a fund manager Mr. M in a wealth management firm is assessing five potential
investment opportunities <1 , <2 , <3 , <4 , <5 (let < = {r1 , r2 , r3 , r4 , r5 }). The firm mandates that
the fund manager has to evaluate the following five parameters (criteria) Y1, Y2 , Y3 , Y4 , Y5 (let Y =
Symmetry 2019, 11, 383 16 of 19

{y1 , y2 , y3 , y4 , y5 }). For each parameter, yq ∈ Y (q = 1, 2, 3, 4, 5). Construct the Pythagorean fuzzy set
decision-making matrices are shown in Table 1.

Table 1. Investment opportunities in a wealth management firm D1 .

Y1 Y2 Y3 Y4 Y5
<1 (0.5, 0.4) (0.5, 0.3) (0.2, 0.6) (0.3, 0.4) (0.3, 0.4)
<2 (0.7, 0.3) (0.7, 0.3) (0.6, 0.2) (0.6, 0.2) (0.7, 0.2)
<3 (0.5, 0.4) (0.6, 0.4) (0.6, 0.2) (0.5, 0.3) (0.6, 0.5)
<4 (0.7, 0.2) (0.7, 0.2) (0.4, 0.2) (0.5, 0.2) (0.4, 0.4)
<5 (0.4, 0.3) (0.5, 0.2) (0.4, 0.5) (0.4, 0.6) (0.3, 0.4)

In this problem, all the entries are benefit type. Therefore, we do not normalize the decision matrix.
(Case:1)
Step 1: Now, we use PFDWG to evolute collective performance with weight vectors β =
(0.15, 0.25, 0.35, 0.10, 0.15)T and σ = 1 as follows in Table 2:

Table 2. Aggregated Pythagorean information matrix (PFDWG).

<1 (0.06568, 0.98699)

<2 (0.57562, 0.99870)
<3 (0.42181, 0.99777)
<4 (0.24911, 0.99898)
<5 (0.16873, 0.99701)

Similarly, we can find for σ = 2, 3, 4, 5, 6, 7, 8, 9, 10.

Steps 2 and 3: Now, we find the score value of each alternative and their ranking as shown in Table 3.

Table 3. Ranking using PFDWG operator.

σ Se (<1 ) Se (<2 ) Se (<3 ) Se (<4 ) Se (<5 ) Ranking

1 −0.9698 −0.6660 −0.8176 −0.9359 −0.9655 <2 > <3 > <4 > <5 > <1
2 −0.9789 −0.6917 −0.8375 −0.9498 −0.9749 <2 > <3 > <4 > <5 > <1
3 −0.9824 −0.7090 −0.8527 −0.9545 −0.9796 <2 > <3 > <4 > <5 > <1
4 −0.9841 −0.7202 −0.8632 −0.9568 −0.9821 <2 > <3 > <4 > <5 > <1
5 −0.9851 −0.7275 −0.8703 −0.9582 −0.9836 <2 > <3 > <4 > <5 > <1
6 −0.9857 −0.7326 −0.8752 −0.9590 −0.9845 <2 > <3 > <4 > <5 > <1
7 −0.9862 −0.7363 −0.8787 −0.9597 −0.9852 <2 > <3 > <4 > <5 > <1
8 −0.9865 −0.7391 −0.8813 −0.9601 −0.9856 <2 > <3 > <4 > <5 > <1
9 −0.9867 −0.7412 −0.8833 −0.9605 −0.9860 <2 > <3 > <4 > <5 > <1
10 −0.9869 −0.7429 −0.8848 −0.9608 −0.9863 <2 > <3 > <4 > <5 > <1

(Case 2)
Step 1: Now, we use PFDWA to evolute collective performance with weight vectors β =
(0.15, 0.25, 0.35, 0.10, 0.15)T and σ = 1 as follows in Table 4:

Table 4. Aggregated Pythagorean information matrix (PFDWA).

<1 (0.99784, 0.17076)

<2 (0.99784, 0.16076)
<3 (0.90668, 0.06663)
<4 (0.96847, 0.04495)
<5 (0.98566, 0.077166)
Symmetry 2019, 11, 383 17 of 19

Similarly, we can find for σ = 2, 3, 4, 5, 6, 7, 8, 9, 10.

Steps 2 and 3: Now, we find the score value of each alternative and their ranking as follows:
Looking at the illustration above, it is evident that, though overall ranking values of the
alternatives are dissimilar, due to the usage of two Dombi aggregation operators, the ranking order
regarding the alternatives are analogous, and the most desirable alternative is <2 in order to analyze
the consequence of parameter σ ∈ [1, 10] on the ranking of the alternatives in the PFDWG and PFDWA
operators, which are exposed in Tables 3 and 5.

Table 5. Ranking using a PFDWA operator.

σ Se (<1 ) Se (<2 ) Se (<3 ) Se (<4 ) Se (<5 ) Ranking

1 0.9664 0.9698 0.81763 0.9359 0.9655 <2 > <1 > <5 > <4 > <3
2 0.9789 0.6917 0.8375 0.9498 0.9749 <1 > <5 > <4 > <3 > <2
3 0.9824 0.7090 0.8527 0.9545 0.9796 <1 > <5 > <4 > <3 > <2
4 0.9841 0.7202 0.8632 0.9568 0.9821 <1 > <5 > <4 > <3 > <2
5 0.9851 0.7275 0.8703 0.9582 0.9836 <1 > <5 > <4 > <3 > <2
6 0.9857 0.7326 0.8752 0.9590 0.9845 <1 > <5 > <4 > <3 > <2
7 0.9862 0.7363 0.8787 0.9597 0.9852 <1 > <5 > <4 > <3 > <2
8 0.9865 0.7391 0.8813 0.9601 0.9856 <1 > <5 > <4 > <3 > <2
9 0.9867 0.7412 0.8833 0.9605 0.9860 <1 > <5 > <4 > <3 > <2
10 0.9869 0.7429 0.8848 0.9608 0.9863 <1 > <5 > <4 > <3 > <2

4.2. Analyzing the Consequence of Parameter σ on Decision-Making Results

To describe the effect of the parameters σ on multi attribute decision-making outcomes, we have
utilized dissimilar values of σ to rank the alternatives. The results of score function and ranking order of
the alternatives Hi (i = 1, 2, 3, 4, 5) in the range of 1 ≤ σ ≤ 10 based on PFDWG and PFDWA operators
are presented in Tables 3 and 5 correspondingly. When σ ∈ [1, 10] using an PFDWG aggregation
operator, we obtained a rank of alternatives as <2 > <3 > <4 > <5 > <1 ; here, <2 is the best choice,
but, when using a PFDWA aggregation operator, we obtained two different ranks. When σ = 1, we
obtained <2 > <1 > <5 > <4 > <3 and when σ ∈ [2, 10] , we get <1 > <5 > <4 > <3 > <2 . Hence,
the overall best rank is <2 .
To these MADM problems based on PFDWG and PFDWA operators, we realize that the different
values of parameters σ can change corresponding ranking orders of the alternatives for PFDWA
operator, which is more reactive to σ in this MADM procedure, even though for numerous values of
the parameters σ might be reformed making arrangements corresponding to PFDWG operators, which
is less responsive to σ in this multi-attribute decision-making (MADM) procedure.

4.3. Comparison Analysis

This section deals with comparison analysis of the proposed Dombi aggregation operators under
Pythagorean fuzzy numbers with other well known aggregation operators. We compared this proposed
Dombi aggregation Operators with O-PFWA [3], O-PFPWA [3], PFWA [8], PFOWA [8], SPFWA [24],
PFEWA [23], PFEOWA [23], CPFWA [25], L-PFWA [26], and L-PFOWA [26,27]. Their results are
summarized as follows.
Symmetry 2019, 11, 383 18 of 19

Comparison analysis of existing operators:

Averaging Operators Ranking Geometric Operators Ranking

O-PFWA [3] <2 > <3 > <4 > <5 > <1 O-PFWG <2 > <3 > <4 > <5 > <1
O-PFWPA [3] <2 > <4 > <3 > <5 > <1 O-PFWPG <2 > <4 > <3 > <5 > <1
PFWA [8] <2 > <4 > <3 > <5 > <1 PFWG <2 > <3 > <4 > <5 > <1
PFOWA [8] <2 > <4 > <3 > <5 > <1 PFOWG <2 > <4 > <3 > <5 > <1
SPFWA [24] <2 > <4 > <3 > <5 > <1 SPFWG <2 > <4 > <3 > <5 > <1
PFEWA [23] <2 > <4 > <3 > <5 > <1 PFEWG <2 > <3 > <4 > <5 > <1
PFEOWA [23] <2 > <4 > <3 > <5 > <1 PFEOWG <2 > <4 > <3 > <5 > <1
CPFWA [25] <4 > <2 > <3 > <5 > <1 CPFWG <2 > <3 > <4 > <5 > <1
L-PFWA [26] <2 > <4 > <3 > <5 > <1 L-PFWG <2 > <3 > <4 > <5 > <1
L-PFOWA [26] <2 > <4 > <3 > <5 > <1 L-PFOWG <2 > <3 > <4 > <5 > <1

Comparision analysis of proposed operators:

Averaging Operators Ranking Geometric Operators Ranking

PFDWA <2 > <1 > <5 > <4 > <3 PFDWG <2 > <3 > <4 > <5 > <1
PFDOWA <2 > <1 > <5 > <3 > <4 PFDOWG <2 > <3 > <4 > <5 > <1
PFDHWA <2 > <1 > <5 > <4 > <3 PFDHWG <2 > <3 > <4 > <5 > <1

From the above comparative analysis table, we say that our proposed Dombi Pythagorean fuzzy
aggregation operators are more effective and reliable than previous aggregation operators.

5. Conclusions
In this paper, we have proposed novel aggregation operators, namely, Dombi weighted
average/geometric, ordered weighted average/geometric and Dombi hybrid weighted average/geometric
for Pythagorean fuzzy numbers. In addition, we gave the comparison of proposed and existing aggregation
operators and discussed how our proposed technique is more effective than other existing operators for
aggregation. Finally, we provided an approach to deal with the decision problems using the proposed
Dombi operators. A numerical example shows how our proposed technique helped us with being effective
in decision-making problems.

Author Contributions: Conceptualization, S.A. (Shahzaib Ashraf) and S.A. (Saleem Abdullah); methodology,
S.A. (Shahzaib Ashraf); software, S.A. (Shahzaib Ashraf); validation, A.A.k. (Arshad Ahmad Khan), S.A.
(Shahzaib Ashraf), S.A. (Saleem Abdullah), M.Q. (Muhammad Qiyas) and J.L. (Jianchao Luo); investigation, S.A.
(Shahzaib Ashraf); writing—original draft preparation, S.A. (Shahzaib Ashraf); writing—review and editing, S.A.
(Saleem Abdullah), M.Q.; visualization, S.A. (Saleem Abdullah); supervision, S.A. (Saleem Abdullah); funding
acquisition, A.A.K., J.L.
Funding: This paper is supported by “Research on the pilot effect evaluation, operational pattern, supporting
policies of the contracted management of farm land mortgage finance”, the National Natural Science Foundation
of China (NSFC), Jan, 2016–2019, No.71573210: Sponsor and host Jianchao luo. This paper is supported by
“Research on the Effectiveness Evaluation, Risk Control and System Construction of the Agricultural Credit
Guarantee Policy”, National Natural Science Foundation of China (NSFC NO: 71873100).
Acknowledgments: The authors would like to thank the editor in chief, associate editor and the anonymous
referees for detailed and valuable comments which helped to improve this manuscript.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Zadeh, L.A. Fuzzy sets. Inf. Control 1965, 8, 338–353.
2. Atanasov, K.T. Intuitionistic fuzzy sets. Fuzzy Sets Syst. 1986, 20, 87–96.
3. Yager, R.R. Pythagorean membership grades in multicriteria decision-making. IEEE Trans. Fuzzy Syst. 2014,
22, 958–965.
Symmetry 2019, 11, 383 19 of 19

4. Yager, R.R.; Abbasov, A.M. Pythagorean membership grades, complex numbers, and decision-making. Int. J.
Intell. Syst. 2013, 28, 436–452.
5. Ren, P.; Xu, Z.; Gou, X. Pythagorean fuzzy TODIM approach to multi-criteria decision-making. Appl. Soft Comput.
2016, 42, 246–259.
6. Wei, G. Pythagorean fuzzy interaction aggregation operators and their application to multiple attribute
decision-making. J. Intell. Fuzzy Syst. 2017, 33, 2119–2132.
7. Wei, G.; Lu, M. Pythagorean fuzzy power aggregation operators in multiple attribute decision-making. Int. J.
Intell. Syst. 2018, 33, 169–186.
8. Xu, Q.; Yu, K.; Zeng, S.; Liu, J. Pythagorean fuzzy induced generalized OWA operator and its application to
multi-attribute group decision-making. Int. J. Innov. Comput. Inf. Control 2017, 13, 1527–1536.
9. Xue, W.; Xu, Z.; Zhang, X.; Tian, X. Pythagorean fuzzy LINMAP method based on the entropy theory for
railway project investment decision making. Int. J. Intell. Syst. 2018, 33, 93–125.
10. Rahman, K.; Abdullah, S.; Ahmed, R.; Ullah, M. Pythagorean fuzzy Einstein weighted geometric aggregation
operator and their application to multiple attribute group decision-making. J. Intell. Fuzzy Syst. 2017, 33,
635–647.
11. Zhang, X.; Xu, Z. Extension of TOPSIS to Multiple Criteria Decision-Making with Pythagorean Fuzzy Sets.
Int. J. Intell. Syst. 2014, 29, 1061–1078.
12. Dombi, J. A general class of fuzzy operators, the DeMorgan class of fuzzy operators and fuzziness measures
induced by fuzzy operators. Fuzzy Sets Syst. 1982, 8, 149–163.
13. Liu, P.; Liu, J.; Chen, S.M. Some intuitionistic fuzzy Dombi Bonferroni mean operators and their application
to multi-attribute group decision-making. J. Oper. Res. Soc. 2018, 69, 1–24.
14. Chen, J.; Ye, J. Some single-valued neutrosophic Dombi weighted aggregation operators for multiple attribute
decision-making. Symmetry 2017, 9, 82.
15. Shi, L.; Ye, J. Dombi aggregation operators of neutrosophic cubic sets for multiple attribute decision-making.
Algorithms 2018, 11, 29.
16. Lu, X.; Ye, J. Dombi aggregation operators of linguistic cubic variables for multiple attribute decision-making.
Information 2018, 9, 188.
17. He, X. Typhoon disaster assessment based on Dombi hesitant fuzzy information aggregation operators.
Nat. Hazards 2018, 90, 1153–1175.
18. Jana, C.; Senapati, T.; Pal, M.; Yager, R.R. Picture fuzzy Dombi aggregation operators: Application to MADM
process. Appl. Soft Comput. 2019, 74, 99–109.
19. Jana, C.; Pal, M.; Wang, J.Q. Bipolar fuzzy Dombi aggregation operators and its application in
multiple-attribute decision-making process. J. Ambient. Intell. Humaniz. Comput. 2018, 1–17. [CrossRef]
20. Wei, G.; Wei, Y. Some single-valued neutrosophic Dombi prioritized weighted aggregation operators in
multiple attribute decision-making. J. Intell. Fuzzy Syst. 2018, 1–13. [CrossRef]
21. Klement, E.P.; Mesiar, R. Logical, Algebraic, Analytic, and Probabilistic Aspects of Triangular Norms; Elsevier:
New York, NY, USA, 2005.
22. Nguyen, H.T.; Walker, R.A. A First Course in Fuzzy Logic; CRC Press: Boca Raton, FL, USA, 1997.
23. Garg H, A new generalized Pythagorean fuzzy information aggregation using einstein operations and its
application to decision making. Int. J. Intell. Syst. 2016, 31, 886–920.
24. Ma, Z.M.; Xu, Z.S. Symmetric Pythagorean fuzzy weighted geometric/averaging operators and their
application in multi-criteria decision-making problems. Int. J. Intell. Syst. 2016, 31, 1198–1219.
25. Garg, H. Confidence levels based Pythagorean fuzzy aggregation operators and its application to
decision-making process. Comput. Math. Organ. Theory 2017, 1–26. [CrossRef]
26. Garg, H. New logarithmic operational laws and their aggregation operators for Pythagorean fuzzy set and
their applications. Int. J. Intell. Syst. 2018, 1–25. [CrossRef]
27. Hussain, A.; Mahmood, T.; Ali, M.I. Rough Pythagorean fuzzy ideals in semigroups. Comput. Appl. Math.
2019. [CrossRef]

c 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).