Abstract
In this paper, we propose a novel multi-attribute group decision making (MAGDM) approach under the p, q-quasirung orthopair fuzzy number (p, q-QOFN) environment. For this, we propose new multiplication operation and scalar power operation for p, q-QOFNs based on Yager’s norm. Then, by using the proposed multiplication operation and scalar power operation of p, q-QOFNs and the concept of prioritized geometric aggregation operator (AO), we propose the p, q-quasirung orthopair fuzzy Yager prioritized weighted geometric (p, q-QOFYPWG) AO for aggregating p, q-QOFNs. We also prove the different properties of the proposed p, q-QOFYPWG AO of p, q-QOFNs. However, based on the proposed p, q-QOFYPWG AO, we propose a new MAGDM approach in the context of p, q-QOFNs environment. Afterwards, we utilize the proposed MAGDM approach to solve the different MAGDM problems, and compare the preference orders (POs) obtained from the proposed MAGDM approach to POs obtained from other existing MAGDM approaches. The proposed MAGDM approach can overcome the shortcomings of the existing MAGDM approaches, where they cannot distinguish the POs of the alternatives in some cases. The proposed MAGDM approach provides a very useful approach to deal with MAGDM problems in the p, q-QOFNs environment.
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Avoid common mistakes on your manuscript.
1 Introduction
Multi-attribute group decision making (MAGDM) is the cognitive process of choosing a particular action from the several available alternatives. It is essential in everyday life, business, and governance because it enables individuals and organizations to manage challenges, make decisions, and achieve goals. Usually, effective MAGDM involves the assessment of possible choices on the basis of some criteria, goals, and constraints. Hence, it is evident that efficient MAGDM is crucial due to the necessity of successfully managing resources and addressing changes and objectives in situations where uncertainty is involved. It originates from incomplete, vague, or estimative information, which prevents reasonable foresight into its effects. To deal with such uncertainties, Zadeh (1965) introduced the theory of fuzzy sets (FSs) in 1965, where a variable can have membership grade (MG) instead of true or false values. Later, Atanassov (1986) defined the extension of the FS known as intuitionistic fuzzy sets (IFSs) that include the non-membership grade (NMG) with the MG. Following this, Yager (2013) generalized the IFSs to Pythagorean fuzzy sets (PFSs) to solve the uncertainties of the environment more effectively, and PFS provides more flexibility to decision making experts (DMExs). However, in certain instances, PFS may not adequately capture the evaluations of the DMExs. Therefore, Yager (2016) expanded on the ideas of IFS and PFS by creating the q-rung orthopair fuzzy set (q-ROFS) \(\langle \zeta _{\mathfrak {T}}, \varrho _{\mathfrak {T}}\rangle \) which satisfy the condition: \(0\le \zeta _{\mathfrak {T}}\le 1\), \(0\le \varrho _{\mathfrak {T}}\le 1\), \( 0\le \zeta _{\mathfrak {T}}^q + \varrho _{\mathfrak {T}}^q \le 1\) and \(q\ge 1\), which provides more range to express the information comparative to IFSs and PFSs. Many researchers have widely utilized the IFSs, PFSs and q-ROFS in various decision-making scenarios (Liu and Chen 2017; Chen et al. 2016; Chen and Niou 2011; Hussain et al. 2023; Alcantud 2023; Salimian and Mousavi 2022; Dutta and Borah 2022; Gao et al. 2021; Zhang et al. 2020; Çalı and Balaman 2019; Kumar and Chen 2022a; Xu and Wang 2012; Chen et al. 2014; Zhang et al. 2015; Kumar and Chen 2023; Garg 2021; Rahman and Ali 2020; Akram et al. 2020; Khan et al. 2019; Liu et al. 2024; Kumar and Chen 2022b; Zhang and Chen 2022; Garg and Chen 2020; Garg 2020; Liu et al. 2018; Pinar and Boran 2020; Wang et al. 2020; Zhong et al. 2019). Zhang et al. (2020) defined the MAGDM approach based on the multiplicative preference relations in the context of intuitionistic fuzzy numbers (IFNs). Kumar and Chen (2022a) proposed the advanced Heronian mean aggregation operator (AO) and MAGDM approach based on the proposed AO in the IFNs environment. Akram et al. (2020) proposed the MAGDM approach for the Pythagoran fuzzy numbers (PFNs) environment by using the ELECTRE technique. Kumar and Chen (2023) proposed the entropy measure of PFSs and AO for aggregation PFNs for MAGDM approach. Garg (2020) developed the AOs based on trigonometric functions and MAGDM approach based on the proposed AOs under the q-rung orthopair fuzzy numbers (q-ROFNs) environment. Liu et al. (2024) developed the AOs based on the Aczel-Alsina norm and power Heronian mean for MAGDM in the context of q-ROFNs.
In a q-ROFS, DMExs must assign equal values of q for both MG and NMG, a constraint that can significantly impact the overall decision-making process. To overcome this limitation, Seikh and Mandal (2022) introduced p, q-quasirung orthopair fuzzy set (p, q-QOFS) and introduced the p, q-quasirung orthopair fuzzy number (p, q-QOFN), where a p, q-QOFS \(\mathfrak {R}\) in the universal set Y is defined as \(\mathfrak {R}= \{\langle y,\zeta _{\mathfrak {R}}(y), \varrho _{\mathfrak {R}}(y) \rangle \mid y \in Y\}\), which satisfy the condition: \(0\le \zeta _{\mathfrak {R}}(y)\le 1\), \(0\le \varrho _{\mathfrak {R}}(y)\le 1\), \( 0\le \zeta _{\mathfrak {R}}^p + \varrho _{\mathfrak {R}}^q \le 1\), \(p\ge 1\) and \(q\ge 1\). The p, q-QOFS allows for a nuanced representation of uncertainty, which can be finely tuned by adjusting p and q. The p, q-QOFS becomes an IFS when \(p=q=1\) and becomes a PFS when \(p=q=2\). Similarly, when \(p=q\), p, q-QOFS is converted into q-ROFS. In last 3 years, researchers have used p, q-QOFSs widely to develop the different MAGDM method (Seikh and Mandal 2022; Rahim et al. 2023b, a, 2024b, a, c; Ahmad et al. 2024). Seikh and Mandal (2022) proposed the AOs for aggregating the p, q-QOFNs and MAGDM approach by using the proposed AOs to solve the problem of suitable site selection for electric vehicle charging. Rahim et al. (2023b) presented AOs based on confidence level technique and MAGDM approach by using the proposed AOs for the p, q-QOFNs environment. Rahim et al. (2023a) proposed the AOs based on sine trigonometric function for aggregating p, q-QOFNs and MAGDM approach based on the proposed AOs under the p, q-QOFNs environment. Rahim et al. (2024a) proposed the cosine similarity measure and distance measures for p, q-QOFSs and its application in MAGDM. Rahim et al. (2024c) introduced Dombi AOs for aggregating the p, q-QOFNs and MAGDM method based on the proposed AOs in the context of p, q-QOFNs. Ahmad et al. (2024) developed AOs based on the Hamacher norm and MAGDM approach based on proposed AOs in p, q-QOFNs environment. Rahim et al. (2024b) proposed the MAGDM approach based on the COPRAS technique for the p, q-QOFNs environment and its application in green supplier selection.
In this paper, we find that Seikh and Mandal’s MAGDM approach (Seikh and Mandal 2022), Ahmad et al.’s MAGDM approach (Ahmad et al. 2024), Garg’s MAGDM approach (Garg 2020), and Rahim et al.’s MAGDM approach (Rahim et al. 2023a) have the shortcomings, where they cannot distinguish the preference orders (POs) of the alternatives in some cases. Therefore, in order to overcome the shortcomings of Seikh and Mandal’s MAGDM approach (Seikh and Mandal 2022), Ahmad et al.’s MAGDM approach (Ahmad et al. 2024), Garg’s MAGDM approach (Garg 2020), and Rahim et al.’s MAGDM approach (Rahim et al. 2023a), it is necessary to propose a new MAGDM approach under the p, q-QOFNs environment.
In this paper, we propose new operations for p, q-QOFNs based on Yager’s norm (Yager 1994), namely, multiplication operation and scalar power operation. However, by using the proposed multiplication operation and scalar power operation, we propose the p, q-quasirung orthopair fuzzy Yager prioritized weighted geometric (p, q-QOFYPWG) AO for aggregating the p, q-QOFNs. We also prove the various properties of proposed p, q-QOFYPWG AO of p, q-QOFNs. Furthermore, by utilizing the p, q-QOFYPWG AO, we propose a novel MAGDM approach under the p, q-QOFNs environment. Afterwards, we solve a few MAGDM problems by using the proposed MAGDM approach and compare the preference orders (POs) obtained from the proposed MAGDM approach with POs obtained from other existing MAGDM approaches. The proposed MAGDM approach can overcome the shortcomings of Seikh and Mandal’s MAGDM approach (Seikh and Mandal 2022), Ahmad et al.’s MAGDM approach (Ahmad et al. 2024), Garg’s MAGDM approach (Garg 2020) and Rahim et al.’s MAGDM approach (Rahim et al. 2023a), where they cannot distinguish the POs of the alternatives in some cases.
The remaining part of this paper is organized as follows: Sect. 2 contains the elementary concepts relevant to this paper. In Sect. 3, we propose the multiplication operation and scalar power operation for p, q-QOFNs using Yager’s norm. Section 4 propose the p, q-QOFYPWG AO based on the proposed operational laws of p, q-QOFNs. In Sect. 5, we propose a new MAGDM approach in the p, q-QOF environment. Finally, Sect. 6 provides conclusion of the paper.
2 Preliminaries
This section presents the basic information related to this article.
Definition 1
(Yager 2016) A q-ROFS \(\mathfrak {T}\) in the universe of discourse Y is defined as:
where \(\zeta _{\mathfrak {T}}(y):Y \rightarrow [0,1]\) denotes the MG and \(\varrho _{\mathfrak {T}}(y):Y \rightarrow [0,1]\) denotes the NMG of \(y\in Y\), respectively, where \(0\le \zeta _{\mathfrak {T}}(y)\le 1\), \(0\le \varrho _{\mathfrak {T}}(y)\le 1\), \(0 \le (\zeta _{\mathfrak {T}}(y))^q+(\varrho _{\mathfrak {T}}(y))^q \le 1\) and \(q\ge 1\). The hesitancy degree of an element \(y \in Y\) is \((\pi _{\mathfrak {T}}(y))=(1-(\zeta _{\mathfrak {T}}(y))^q-(\varrho _{\mathfrak {T}}(y))^q)^{\frac{1}{q}}\).
Usually, the pair \(\langle \zeta _\mathfrak {T}(y), \varrho _\mathfrak {T}(y) \rangle \) in the q-ROFSs \(\mathfrak {T}= \{\langle y,\zeta _{\mathfrak {T}}(y), \varrho _{\mathfrak {T}}(y) \rangle \mid y \in Y\}\) called the q-ROFN.
Definition 2
(Yager 2016) Let \(\mathfrak {T}_{1}=\langle \zeta _{1}, \varrho _{1} \rangle \), \(\mathfrak {T}_{2}=\langle \zeta _{2}, \varrho _{2} \rangle \) and \(\mathfrak {T}=\langle \zeta , \varrho \rangle \) be three q-ROFNs, \(\kappa >0\). Then
-
(i)
\(\mathfrak {T}_{1}\oplus \mathfrak {T}_{2}=\left\langle \root q \of {\zeta ^q_{1}+{\zeta ^q_{2}}-\zeta _{1}^q\zeta _{2}^q},\varrho _{1}\varrho _{2} \right\rangle \);
-
(ii)
\(\mathfrak {T}_{1}\otimes \mathfrak {T}_{2}=\left\langle \zeta _{1}\zeta _{2},\root q \of {{\varrho ^q_{1}}+{\varrho ^q_{2}}-\varrho _{1}^q \varrho _{2}^q} \right\rangle \);
-
(iii)
\(\kappa \mathfrak {T}=\left\langle \root q \of {1-(1-\zeta ^q)^{\kappa }},\varrho ^{\kappa } \right\rangle \);
-
(iv)
\(\mathfrak {T}^{\kappa }=\left\langle \zeta ^{\kappa },\root q \of {1-(1-\varrho ^q)^{\kappa }}\right\rangle \).
Definition 3
(Seikh and Mandal 2022) A p, q-QOFS \(\mathfrak {R}\) in finite universe of discourse Y is defined as:
where \(\zeta _{\mathfrak {R}}(y)\) denotes the MG and \(\varrho _{\mathfrak {R}}(y)\) denotes the NMG of \(y\in Y\), respectively, where \(0\le \zeta _{\mathfrak {R}}(y)\le 1\), \(0\le \varrho _{\mathfrak {R}}(y)\le 1\), \(0 \le (\zeta _{\mathfrak {R}}(y))^p+(\varrho _{\mathfrak {R}}(y))^q \le 1\), \(p\ge 1\) and \(q\ge 1\). The hesitancy degree of an element \(y \in Y\) is \((\pi _{\mathfrak {R}}(y))^l=1-(\zeta _{\mathfrak {R}}(y))^p-(\varrho _{\mathfrak {R}}(y))^q\), where l is the least common multiple (LCM) of p and q.
In (Seikh and Mandal 2022), Seikh and Mandal called the pair \(\langle \zeta _\mathfrak {R}, \varrho _\mathfrak {R} \rangle \) in the p, q-QOFS \(\mathfrak {R}= \{\langle y,\zeta _{\mathfrak {R}}(y), \varrho _{\mathfrak {R}}(y) \rangle \mid y \in Y\}\) a p, q-QOFN.
Remark 1
Let us consider a case where we need to determine the minimum values of p and q, both greater than or equal to 1, for a given orthopair \(\langle \zeta _\mathfrak {R}, \varrho _\mathfrak {R} \rangle \), such that \(\zeta ^p_{\mathfrak {R}}+\varrho ^q_\mathfrak {R} \le 1\). Iterative computing approaches can provide unique solutions to issues that lack a closed-form solution. The minimal values of p and q that satisfy \(\zeta ^p_\mathfrak {R}+\varrho ^q_\mathfrak {R} \le 1\) are referred to as the p, q-niche of \(\langle \zeta _\mathfrak {R}, \varrho _\mathfrak {R} \rangle \). Note that if \(\hat{p},\hat{q}\) is the p, q-niche of \(\langle \zeta _\mathfrak {R}, \varrho _\mathfrak {R} \rangle \), then \(\langle \zeta _\mathfrak {R}, \varrho _\mathfrak {R} \rangle \) is valid for all \(p \ge \hat{p}\) and \(q \ge \hat{q}\).
Let \(Z=\{z_1,z_2,\ldots , z_n\}\) be some provided data and \(\digamma \) be a fuzzy concept. Assume an expert presents his preference as an orthopair \(\langle \zeta _\mathfrak {R}(z_{j}), \varrho _\mathfrak {R}(z_{j}) \rangle \) for each \(z_{j}\in Z\). Now the problem is to accurately portray the information by estimating the proper values of p and q. We may now proceed as follows:
-
(i)
Determine the p, q-niche for each orthopair \(\langle \zeta _\mathfrak {R}(z_{j}), \varrho _\mathfrak {R}(z_{j}) \rangle \), say \(p_j,q_j\).
-
(ii)
Determine the \(p^{*}\) and \(q^{*}\) niches where \(p^* = \max _{j}\{p_{j}\}\) and \(q^* = \max _{j}\{q_{j}\}\).
-
(iii)
Then we may denote E as \(p^{*},q^{*}\)-QOFS.
Definition 4
(Seikh and Mandal 2022) Let \(\mathfrak {R}_{1}=\langle \zeta _{1}, \varrho _{1} \rangle \), \(\mathfrak {R}_{2}=\langle \zeta _{2}, \varrho _{2} \rangle \) and \(\mathfrak {R}=\langle \zeta , \varrho \rangle \) be three p, q-QOFNs. Then,
-
(i)
\(\mathfrak {R}_{1}\oplus \mathfrak {R}_{2}=\left\langle \root p \of {{\zeta ^p_{1}}+{\zeta ^p_{2}}-\zeta _{1}^p\zeta _{2}^p},\varrho _{1}\varrho _{2}\right\rangle \).
-
(ii)
\(\mathfrak {R}_{1}\oplus \mathfrak {R}_{2}=\left\langle \zeta _{1}\zeta _{2},\root q \of {{\varrho ^q_{1}}+{\varrho ^q_{2}}-\varrho _{1}^q \varrho _{2}^q} \right\rangle \).
-
(iii)
\(\kappa \mathfrak {R}=\left\langle \root p \of {1-(1-\zeta ^p)^{\kappa }},\varrho ^{\kappa } \right\rangle \), where \(\kappa >0\).
-
(iv)
\(\mathfrak {R}^{\kappa }=\left\langle \zeta ^{\kappa },\root q \of {1-(1-\varrho ^q)^{\kappa }} \right\rangle \), where \(\kappa >0\).
Definition 5
(Seikh and Mandal 2022) Let \(\mathfrak {R}=\langle \zeta , \varrho \rangle \) be a p, q-QOFN. The score function \(S(\mathfrak {R})\) of \(\mathfrak {R}\) is defined as follows:
where \(S(\mathfrak {R}) \in [0,1]\), \( p\ge 1 \) and \(q\ge 1\).
Definition 6
(Seikh and Mandal 2022) Let \(\mathfrak {R}=\langle \zeta , \varrho \rangle \) be a p, q-QOFN. The accuracy function \({A}(\mathfrak {R})\) of \(\mathfrak {R}\) is defined as follows:
where \({A}(\mathfrak {R}) \in [0,1]\), \( p\ge 1 \) and \(q\ge 1\).
Definition 7
(Seikh and Mandal 2022) Let \(\mathfrak {R}_{1}=\langle \zeta _{1}, \varrho _{1} \rangle \) and \(\mathfrak {R}_{2}=\langle \zeta _{2}, \varrho _{2} \rangle \) be two p, q-QOFN. Then,
-
(i)
If \(S(\mathfrak {R}_{1})>S(\mathfrak {R}_{2})\) then \(\mathfrak {R}_{1}\succ \mathfrak {R}_{2}\).
-
(ii)
If \(S(\mathfrak {R}_{1})<S(\mathfrak {R}_{2})\) then \(\mathfrak {R}_{1}\prec \mathfrak {R}_{2}\).
-
(iii)
If \(S(\mathfrak {R}_{1})=S(\mathfrak {R}_{2})\) and,
-
(a)
If \({A}(\mathfrak {R}_{1})>({A}\mathfrak {R}_{2})\) then \(\mathfrak {R}_{1}\succ \mathfrak {R}_{2}\).
-
(b)
If \({A}(\mathfrak {R}_{1})<{A}(\mathfrak {R}_{2})\) then \(\mathfrak {R}_{1}\prec \mathfrak {R}_{2}\).
-
(c)
If \({A}(\mathfrak {R}_{1})={A}(\mathfrak {R}_{2})\) then \(\mathfrak {R}_{1}\sim \mathfrak {R}_{2}\).
Definition 8
(Yager 1994) Let \(\alpha \) and \(\beta \) be two real numbers and \(\lambda > 0\). The Yager’s t-norm \(Y_{TN}\) and t-conorm \(Y_{TCN}\) are defined as follows:
Definition 9
(Yager 2008) Let H be any alternative and let \(\Phi _{1},\Phi _{2},\ldots ,\) and \(\Phi _{n}\) be attributes with the linear priority order \(\Phi _{1}\succ \Phi _{2}\succ \ldots \succ \Phi _{n}\). If attribute \(\Phi _{e}\) has a higher priority order than attribute \(\Phi _{h}\) then \(e<h\), where \(e,h=1,2,\ldots ,n\) and \(e\ne h\). Let \(\Phi _{h}(H)\) represents the performance of the alternative H with respect to the attribute \(\Phi _{h}\), where \(\Phi _{h}(H) \in [0,1]\). The prioritized geometric (PG) AO of \(\Phi _{1}(H),\Phi _{2}(H),\ldots ,\) and \(\Phi _{n}(H)\) is defined as follows:
where \(T_{1}=1\), \(T_{h}=\prod _{k=1}^{h-1}\Phi _{k}(H)\) and \(h=2,3,\ldots ,n\).
3 The proposed p, q- quasirung orthopair fuzzy operations based on Yager’s norm
In this section, we propose new multiplication operation and scalar power operation for p, q-QOFNs based on Yager’s t-NM \(Y_{TN}\) and t-CNM \(Y_{TCN}\) defined in Definition 8.
Definition 10
Let \(\mathfrak {R}_{1}=\langle \zeta _{1}, \varrho _{1} \rangle \), \(\mathfrak {R}_{2}=\langle \zeta _{2}, \varrho _{2} \rangle \) and \(\mathfrak {R}=\langle \zeta , \varrho \rangle \) be three p, q-QOFNs. The proposed multiplication operation and proposed scalar power operation for p, q-QOFNs \(\mathfrak {R}_{1}=\langle \zeta _{1}, \varrho _{1} \rangle \), \(\mathfrak {R}_{2}=\langle \zeta _{2}, \varrho _{2} \rangle \) and \(\mathfrak {R}=\langle \zeta , \varrho \rangle \) based on the Yager’s norm are defined as follows:
-
(i)
Multiplication operation:
$$\begin{aligned} & \mathfrak {R}_{1}\otimes \mathfrak {R}_{2}\nonumber \\ & =\left\langle \root p \of {1-\min (1,((1-\zeta _{1}^{p})^\lambda +(1-\zeta _{2}^{p})^\lambda )^{\frac{1}{\lambda }})},\right. \nonumber \\ & \left. \root q \of {\min (1,(\varrho _{1}^{q\lambda }+\varrho _{2}^{q\lambda })^{\frac{1}{\lambda }})}\right\rangle , \end{aligned}$$(6)where \(p\ge 1\), \(q\ge 1\) and \(\lambda >0\).
-
(ii)
Scalar power operation:
$$\begin{aligned} \mathfrak {R}^{\kappa }= & \left\langle \root p \of {1-\min (1,(\kappa (1-\zeta ^{p})^\lambda )^{\frac{1}{\lambda }})},\right. \nonumber \\ & \left. \root q \of {\min (1,(\kappa \varrho ^{q\lambda })^{\frac{1}{\lambda }})} \right\rangle , \end{aligned}$$(7)where \(p\ge 1\), \(q\ge 1\), \(\kappa >0\) and \(\lambda >0\).
Example 1
Let \(\mathfrak {R}_{1}=\langle 0.7, 0.6 \rangle \) and \(\mathfrak {R}_{2}=\langle 0.8,0.4 \rangle \) be two p, q-QOFNs. Then,
-
(i)
By using Eq. (6), for \(p=3\), \(q=3\), and \(\lambda =3\), we obtain
$$\begin{aligned} \mathfrak {R}_{1}\otimes \mathfrak {R}_{2}= & \left\langle \root p \of {1-\min (1,((1-\zeta _{1}^{p})^\lambda +(1-\zeta _{2}^{p})^\lambda )^{\frac{1}{\lambda }})}, \right. \\ & \left. \root q \of {\min (1,(\varrho _{1}^{q\lambda }+\varrho _{2}^{q\lambda })^{\frac{1}{\lambda }})}\right\rangle \\= & \left\langle \root 3 \of {1-\min (1,((1-0.7^{3})^3 +(1-0.8^{3})^3)^{\frac{1}{3}})}, \right. \\ & \left. \root 3 \of {\min (1,(0.6^{9}+0.4^{9})^{\frac{1}{3}})}\right\rangle \\= & \langle 0.64, 0.60 \rangle . \end{aligned}$$ -
(ii)
By using Eq. (7), for \(p=3\), \(q=3,\lambda =3\), and \(\kappa =2\), we obtain
$$\begin{aligned} {\mathfrak {R}_{1}}^2= & \left\langle \root p \of {1-\min (1,(\kappa (1-\zeta ^{p}_{1})^\lambda )^{\frac{1}{\lambda }})}, \right. \\ & \left. \root q \of {\min (1,(\kappa \varrho ^{q\lambda }_{1})^{\frac{1}{\lambda }})} \right\rangle \\= & \left\langle \root 3 \of {1-\min (1,(2 (1-0.7^{3})^3)^{\frac{1}{3}})}, \right. \\ & \left. \root 3 \of {\min (1,(2(0.6)^{9})^{\frac{1}{3}})} \right\rangle \\= & \langle 0.56, 0.65 \rangle . \end{aligned}$$
Theorem 1
Let \(\mathfrak {R}_{1}\), \(\mathfrak {R}_{2}\) and \(\mathfrak {R}\) be three p, q-QOFNs. The proposed multiplication operation and scalar power operation, defined in Definition 10, satisfy the following properties:
-
(i)
\(\mathfrak {R}_{1}\otimes \mathfrak {R}_{2}=\mathfrak {R}_{2}\otimes \mathfrak {R}_{1}\),
-
(ii)
\((\mathfrak {R}_{1}\otimes \mathfrak {R}_{2})^{\kappa }=\mathfrak {R}_{1}^{\kappa }\otimes \mathfrak {R}_{2}^{\kappa }\),
-
(iii)
\(\mathfrak {R}^{\kappa _{1}}\otimes R ^{\kappa _{2}}= \mathfrak {R}^{(\kappa _{1}+\kappa _{2})}\),
where \(\kappa >0\), \( \kappa _{1}>0\) and \(\kappa _{2}>0\).
Proof
Let \(\mathfrak {R}_{1}=\langle \zeta _{1}, \varrho _{1} \rangle \), \(\mathfrak {R}_{2}=\langle \zeta _{2}, \varrho _{2} \rangle \) and \(\mathfrak {R}=\langle \zeta , \varrho \rangle \) be three p, q-QOFNs. Then,
-
(i)
By using Eq. (6), we have
$$\begin{aligned} \mathfrak {R}_{1}\otimes \mathfrak {R}_{2}= & \left\langle \root p \of {1-\min (1,((1-\zeta _{1}^{p})^\lambda +(1-\zeta _{2}^{p})^\lambda )^{\frac{1}{\lambda }})},\right. \\ & \left. \root q \of {\min (1,(\varrho _{1}^{q\lambda }+\varrho _{2}^{q\lambda })^{\frac{1}{\lambda }})}\right\rangle \\= & \left\langle \root p \of {1-\min (1,((1-\zeta _{2}^{p})^\lambda +(1-\zeta _{1}^{p})^\lambda )^{\frac{1}{\lambda }})}, \right. \\ & \left. \root q \of {\min (1,(\varrho _{2}^{q\lambda }+\varrho _{1}^{q\lambda })^{\frac{1}{\lambda }})}\right\rangle \\= & \mathfrak {R}_{2}\otimes \mathfrak {R}_{1}, \end{aligned}$$where \(p\ge 1\), \(q\ge 1\) and \(\lambda >0\).
-
(ii)
By using Eq. (6) and (7), we have
$$\begin{aligned} (\mathfrak {R}_{1}\otimes \mathfrak {R}_{2})^{\kappa }= & \left\langle \root p \of {1-\min (1,((1-\zeta _{1}^{p})^\lambda +(1-\zeta _{2}^{p})^\lambda )^{\frac{1}{\lambda }})},\right. \\ & \left. \root q \of {\min (1,(\varrho _{1}^{q\lambda }+\varrho _{2}^{q\lambda })^{\frac{1}{\lambda }})}\right\rangle ^ {\kappa }\\= & \left\langle \root p \of {1-\min (1,(\kappa ((1-\zeta _{1}^{p})^\lambda +(1-\zeta _{2}^{p})^\lambda ))^{\frac{1}{\lambda }})},\right. \\ & \left. \root q \of {\min (1,\kappa ((\varrho _{1}^{q\lambda }+\varrho _{2}^{q\lambda })^{\frac{1}{\lambda }}))}\right\rangle \\= & \left\langle \root p \of {1-\min (1,(\kappa (1-\zeta _{1}^{p})^\lambda +\kappa (1-\zeta _{2}^{p})^\lambda )^{\frac{1}{\lambda }})}, \right. \\ & \left. \root q \of {\min (1,(\kappa \varrho _{1}^{q\lambda }+\kappa \varrho _{2}^{q\lambda })^{\frac{1}{\lambda }})}\right\rangle \\= & \mathfrak {R}_{1}^{\kappa }\otimes \mathfrak {R}_{2}^{\kappa }, \end{aligned}$$where \(p\ge 1\), \(q\ge 1\), \(\lambda >0\) and \(\kappa >0\).
-
(iii)
By using Eq. (6) and (7), we have
$$\begin{aligned} \mathfrak {R}^{\kappa _{1}}\otimes \mathfrak {R}^{\kappa _{2}}= & \left\langle \root p \of {1-\min (1,(\kappa _{1}(1-\zeta ^{p})^\lambda +\kappa _{2}(1-\zeta ^{p})^\lambda )^{\frac{1}{\lambda }})},\right. \\ & \left. \root q \of {\min (1,(\kappa _{1}\varrho ^{q\lambda }+\kappa _{2} \varrho ^{q\lambda })^{\frac{1}{\lambda }})}\right\rangle \\= & \left\langle \root p \of {1-\min (1,((\kappa _{1} +\kappa _{2})(1-\zeta ^{p})^\lambda )^{\frac{1}{\lambda }})}, \right. \\ & \left. \root q \of {\min (1,(\kappa _{1}+\kappa _{2}) \varrho ^{q\lambda })^{\frac{1}{\lambda }})}\right\rangle \\= & \mathfrak {R}^{(\kappa _{1}+\kappa _{2})}, \end{aligned}$$where \(p\ge 1\), \(q\ge 1\), \(\lambda >0\), \(\kappa _1>0\) and \(\kappa _2>0\).
\(\square \)
4 The proposed p, q- quasirung orthopair fuzzy Yager prioritized weighted geometric aggregation operator of p, q-QOFNs
In this section, we propose the p, q-quasirung orthopair fuzzy Yager prioritized weighted geometric (p, q-QOFYPWG) AO for p, q-QOFNs based on the proposed multiplication operation, scalar power operation and the prioritized geometric AO given in Definition 9.
Definition 11
Let \(\mathfrak {R}_{1}=\langle \zeta _{1}, \varrho _{1}\rangle \), \(\mathfrak {R}_{2}=\langle \zeta _{2}, \varrho _{2}\rangle \), \(\ldots \), and \(\mathfrak {R}_{n}=\langle \zeta _{n}, \varrho _{n}\rangle \) be n p, q-QOFNs. The proposed p, q-QOFYPWG AO for aggregating the p, q-QOFNs \(\mathfrak {R}_{1}=\langle \zeta _{1}, \varrho _{1}\rangle \), \(\mathfrak {R}_{2}=\langle \zeta _{2}, \varrho _{2}\rangle \), \(\ldots \), and \(\mathfrak {R}_{n}=\langle \zeta _{n}, \varrho _{n}\rangle \) is defined as:
where \(p\ge 1\), \(q\ge 1\), \(\lambda >0\), \(w_{h}\) represents the weight of p, q-QOFN \(\mathfrak {R}_{h}\), \(w_{h} \ge 0\), \(h=1,2,\ldots , n\), \(\sum _{h=1}^{n}w_{h}=1\), \(T_{1}=1\), \(T_{h}=\prod _{e=1}^{h-1}S(\mathfrak {R}_{e})\), \(h=2, 3, \ldots ,n\), and \(S(\mathfrak {R}_{e})\) is the score value of the p, q-QOFN \(\mathfrak {R}_{e}=\langle \zeta _{e}, \varrho _{e} \rangle \) calculated by Eq. (3), \(S(\mathfrak {R}_{e})=\frac{1+\zeta ^p_{{e}}-\varrho ^q_{e}}{2}\) and \(e=1,2,\ldots ,h-1\).
Theorem 2
Let \(\mathfrak {R}_{1}=\langle \zeta _{1}, \varrho _{1}\rangle \), \(\mathfrak {R}_{2}=\langle \zeta _{2}, \varrho _{2}\rangle \), \(\ldots \), and \(\mathfrak {R}_{n}=\langle \zeta _{n}, \varrho _{n}\rangle \) be n p, q-QOFNs. The aggregated value of p, q-QOFNs \(\mathfrak {R}_{1}=\langle \zeta _{1}, \varrho _{1} \rangle \), \(\mathfrak {R}_{2}=\langle \zeta _{2}, \varrho _{2} \rangle \),, \(\ldots \), and \(\mathfrak {R}_{n}=\langle \zeta _{n}, \varrho _{n} \rangle \) by using the proposed p, q-QOFYPWG AO is a p, q-QOFN and given as follows:
where \(p\ge 1\), \(q\ge 1\), \(\lambda >0\), \(w_{h}\) represents the weight of p, q-QOFN \(\mathfrak {R}_{h}\), \(w_{h}\ge 0\), \(h=1,2,\ldots , n\), \(\sum _{h=1}^{n}w_{h}=1\), \(T_{1}=1\), \(T_{h}=\prod _{e=1}^{h-1}S(\mathfrak {R}_{e})\), \(h=2, 3, \ldots ,n\), and \(S(\mathfrak {R}_{e})\) is the score value of the p, q-QOFN \(\mathfrak {R}_{e}=\langle \zeta _{e}, \varrho _{e} \rangle \) calculated by Eq. (3), \(S(\mathfrak {R}_{e})=\frac{1+\zeta ^p_{e}-\varrho ^q_{e}}{2}\) and \(e=1,2,\ldots ,h-1\).
Proof
Let \(\mathfrak {R}_{1}=\langle \zeta _{1}, \varrho _{1}\rangle \), \(\mathfrak {R}_{2}=\langle \zeta _{2}, \varrho _{2}\rangle \), \(\ldots \), and \(\mathfrak {R}_{n}=\langle \zeta _{n}, \varrho _{n}\rangle \) be p, q-QOFNs and let \(u_{h}=\frac{w_{h}T_{h}}{\sum _{h=1}^{n}w_{h}T_{h}}\). To prove this theorem, we use the mathematical induction approach, as illustrated below:
-
(i)
Take \(h=2\), then by using Eq. (7), we obtain
$$\begin{aligned} {\mathfrak {R}_1}^{u_{1}}= & \left\langle \root p \of {1-\min \left\{ 1,\left( u_{1} (1-\zeta _{1}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} }, \right. \\ & \left. \root q \of {\min \left\{ 1,\left( u_{1}(\varrho _{1}^{q})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} } \right\rangle ,\\ {\mathfrak {R}_2}^{u_{2}}= & \left\langle \root p \of {1-\min \left\{ 1,\left( u_{2} (1-\zeta _{2}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} }, \right. \\ & \left. \root q \of {\min \left\{ 1,\left( u_{2}(\varrho _{2}^{q})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} } \right\rangle . \end{aligned}$$Then, by using Eq. (6), we obtain
$$\begin{aligned} & p, q-QOFYPWG(\mathfrak {R}_1,\mathfrak {R}_2)={\mathfrak {R}_1}^{u_{1}}\otimes {\mathfrak {R}_2}^{u_{2}}\\ & \quad = \left\langle \root p \of {1-\min \left\{ 1,\left( u_{1} (1-\zeta _{1}^{p})^{\lambda }+u_{2} (1-\zeta _{2}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} }\right. ,\\ & \qquad \left. \root q \of {\min \left\{ 1,\left( u_{1}(\varrho _{1}^{q})^{\lambda }+u_{2}(\varrho _{2}^{q})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} } \right\rangle \\ & \quad =\left\langle \root p \of {1-\min \left\{ 1,\left( \sum _{h=1}^{2}u_{h} (1-\zeta _{h}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} },\right. \\ & \qquad \left. \root q \of {\min \left\{ 1,\left( \sum _{h=1}^{2}u_{h}(\varrho _{h}^{q})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} }\right\rangle . \end{aligned}$$Hence, the result given in Eq. (9) is valid for \(h=2\).
-
(ii)
Suppose the result given in Eq. (9) is valid for \(h=n\), where
$$\begin{aligned} & p, q - QOFYPWG(\mathfrak {R}_{1},\mathfrak {R}_{2},\ldots ,\mathfrak {R}_{n})\\ & \quad =\left\langle \root p \of {1-\min \left\{ 1,\left( \sum _{h=1}^{n}u_h (1-\zeta _{h}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} },\right. \\ & \quad \left. \root q \of {\min \left\{ 1,\left( \sum _{h=1}^{n}u_h(\varrho _{h}^{q})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} } \right\rangle . \end{aligned}$$ -
(iii)
Now, take \(h=n+1\), we get
$$\begin{aligned} & p, q-QOFYPWG(\mathfrak {R}_1,\mathfrak {R}_2,\ldots \mathfrak {R}_{n+1})\\ & \quad =\left( \otimes _{h=1}^{n}{\mathfrak {R}_h}^{u_{h}}\right) \otimes {\mathfrak {R}_{n+1}}^{u_{n+1}}\\ & \quad = \left\langle \root p \of {1-\min \left\{ 1,\left( \sum _{h=1}^{n}u_{h} (1-\zeta _{h}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} },\right. \\ & \quad \left. \root q \of {\min \left\{ 1,\left( \sum _{h=1}^{n}u_{h}(\varrho _{h}^{q})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} }\right\rangle \\ & \quad \otimes \left\langle \root p \of {1-\min \left\{ 1,\left( u_{n+1} (1-\zeta _{{n+1}}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} },\right. \\ & \quad \left. \root q \of {\min \left\{ 1,\left( u_{n+1}(\varrho _{{n+1}}^{q})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} } \right\rangle \\ & \quad =\left\langle \root p \of {1-\min \left\{ 1,\left( \sum _{h=1}^{n+1}u_h (1-\zeta _{h}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} },\right. \\ & \quad \left. \root q \of {\min \left\{ 1,\left( \sum _{h=1}^{n+1}u_h(\varrho _{h}^{q})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} } \right\rangle . \end{aligned}$$Hence, the result given in Eq. (9) is valid for \(h=n+1\). Thus, the result is true for all natural numbers. Now, we shall prove that the result given in Eq. (9) is a p, q-QOFN. Let
$$\begin{aligned} \delta= & \root p \of {1-\min \left\{ 1,\left( \sum _{h=1}^{n}\frac{w_{h}T_{h}}{\sum _{h=1}^{n}w_{h}T_{h}} (1-\zeta _{h}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} }\\= & \root p \of {1-\min \left\{ 1,\left( \sum _{h=1}^{n}u_h (1-\zeta _{h}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} },\\ \gamma= & \root q \of {\min \left\{ 1,\left( \sum _{h=1}^{n}\frac{w_{h}T_{h}}{\sum _{h=1}^{n}w_{h}T_{h}}(\varrho _{h}^{q})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} } \\= & \root q \of {\min \left\{ 1,\left( \sum _{h=1}^{n}u_h(\varrho _{h}^{q})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} }. \end{aligned}$$Now, we will show that
-
(a)
\(0\le \delta \le 1\) and \(0\le \gamma \le 1\),
-
(b)
\(0\le \delta ^p+\gamma ^q\le 1\).
First, we prove that \(0\le \delta \le 1\). Because \(\mathfrak {R}_{1}=\langle \zeta _{1}, \varrho _{1}\rangle \), \(\mathfrak {R}_{2}=\langle \zeta _{2}, \varrho _{2}\rangle \), \(\ldots \), and \(\mathfrak {R}_{n}=\langle \zeta _{n}, \varrho _{n}\rangle \) are the p, q-QOFNs, we get \(0\le \zeta _h \le 1\), \(0\le \varrho _h \le 1\) and \(0\le \zeta _h^p +\varrho _h^q \le 1\), for all \(h=1,2,\ldots ,n\), \(p\ge 1\) and \(q\ge 1\). Therefore, we get \(0 \le {\zeta _{h} }^p\le 1\). Because \(\lambda >0\), we get \(0 \le (1-\zeta _{h}^p)^\lambda \le 1\). Now, let \(u_{h}=\frac{w_{h}T_{h}}{\sum _{h=1}^{n}w_{h}T_{h}}\), since \(w_h \ge 0\), \(S(\mathfrak {R}_h)\in [0,1], T_1=1\) and \(T_k=\prod _{e=1}^{k-1}S(\mathfrak {R}_e)\) \(\implies T_h\in [0,1]\) and \(w_{h}T_{h}\in [0,1]\). Therefore, we get \(0 \le \frac{w_{h}T_{h}}{\sum _{h=1}^{n}w_{h}T_{h}} \le 1\) \(\implies 0 \le u_h \le 1\). Thus, we get \(0 \le \left( \sum _{h=1}^{n}u_{h} (1-\zeta _{\mathfrak {R}_{h}}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\le 1\) and \( 0 \le 1-\min \left\{ 1,\left( \sum _{h=1}^{n}u_h (1-\zeta _{\mathfrak {R}_{h}}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} \le 1\). It implies that \( 0 \le \root p \of {1-\min \left\{ 1,\left( \sum _{h=1}^{n}u_h (1-\zeta _{\mathfrak {R}_{h}}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} }\le 1\). Hence \( 0 \le \delta \le 1\). Similarly, we can show that \(0\le \gamma \le 1\). Now, we prove that \(0 \le \delta ^p+\gamma ^q\le 1\). Since, \(0\le \zeta _{h}\le 1\), \(0\le \varrho _{h} \le 1\) and \(0\le \zeta _{h}^p+\varrho _{h}^q \le 1\), then we have,
$$\begin{aligned} & \varrho _{h}^q \le 1-\zeta _{h}^p.\\ & \quad \implies \sum _{h=1}^{n}u_{h}\varrho _{h}^{q\lambda }\le \sum _{h=1}^{n}u_{h}(1-\zeta _{h}^{p})^{\lambda } \\ & \quad \implies \min \left\{ 1,\sum _{h=1}^{n}u_{h}(\varrho _{h}^{q\lambda })^{\frac{1}{\lambda }}\right\} \le \min \left\{ 1,\left( \sum _{h=1}^{n}\delta _{h}(1-\zeta _{h}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} \\ & \quad \implies \min \left\{ 1,\sum _{h=1}^{n}u_{h}(\varrho _{h}^{q\lambda })^{\frac{1}{\lambda }}\right\} \\ & \qquad - \min \left\{ 1,\left( \sum _{h=1}^{n}u_{h}(1-\zeta _{h}^{p})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} \le 0\\ & \quad \implies 1-\min \left\{ 1,\sum _{h=1}^{n}u_{h}(1-\zeta _{h}^{p\lambda })^{\frac{1}{\lambda }}\right\} + \min \left\{ 1,\left( \sum _{h=1}^{n}u_{h}(\varrho _{h}^{q})^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} \le 1\\ & \quad \implies \delta ^p+\gamma ^q\le 1.\\ \end{aligned}$$Because \(\delta \ge 0\), \(\gamma \ge 0\), \(p\ge 1\) and \(q\ge 1\), we get \(\delta ^p\ge 0\), \(\gamma ^q\ge 0\) and \(\delta ^p+\gamma ^q\ge 0\). Hence, \(0 \le \delta ^p+\gamma ^q\le 1\).
-
(a)
\(\square \)
Example 2
Let \(\mathfrak {R}_{1}=\langle 0.6,0.8 \rangle \), \(\mathfrak {R}_{2}=\langle 0.4,0.6 \rangle \) and \(\mathfrak {R}_{3}=\langle 0.5,0.5 \rangle \) be three p, q-QOFNs with weights \(w_{1}=0.3, w_{2}=0.4\) and \(w_{3}=0.3\), respectively. First, we calculate the values of \(T_{1}=1\), \(T_{2}=S(\mathfrak {R}_{1})=\frac{1+\zeta ^p_{1}-\varrho ^q_{1}}{2}=\frac{1+0.6^{3}-0.8^{1}}{2}=0.2080\) and \(T_{3}=S(\mathfrak {R}_{1})\times S(\mathfrak {R}_{2})=0.2080 \times 0.2320=0.0483\). By using the proposed p, q-QOFYPWG AO of p, q-QOFNs shown in Eq. (8), we aggregate the p, q-QOFNs \(\mathfrak {R}_{1},\mathfrak {R}_{2}\) and \(\mathfrak {R}_{3}\), where \(p=3\), \(q=1\), \(\lambda =2\) and
In the following, we present some characteristics of the proposed p, q-QOFYPWG AO of p, q-QOFNs.
Property 1
(Idempotency) Let \(\mathfrak {R}_{1}=\langle \zeta _{1}, \varrho _{1}\rangle \), \(\mathfrak {R}_{2}=\langle \zeta _{2}, \varrho _{2}\rangle \), \(\ldots \), and \(\mathfrak {R}_{n}=\langle \zeta _{n}, \varrho _{n}\rangle \) be n p, q-QOFNs with weights \(w_{1}\), \(w_{2}\),\(\ldots \) and \(w_{n}\), respectively, where \(w_{h}\ge 0\), \(\sum _{h=1}^{n}w_{h}=1\) and \(h=1,2,\ldots ,n\). If \( \mathfrak {R}_{1}=\mathfrak {R}_{2}=\ldots =\mathfrak {R}_{n}=\mathfrak {R}\), then
Proof
Since the weights of the p, q-QOFNs \(\mathfrak {R}_{1},\mathfrak {R}_{2},\ldots , \mathfrak {R}_{n}\) are \(w_{1},w_{2},\ldots ,w_{n}\), respectively, where \(w_{h}\ge 0\) and \(\sum _{h=1}^{n}w_{h}=1\), if \(\mathfrak {R}_{1}=\mathfrak {R}_{2},\ldots ,= \mathfrak {R}_{n}=\mathfrak {R}\), then by using Eq.(8), we get
\(\square \)
Property 2
(Boundedness) Let \(\mathfrak {R}_{1}, \mathfrak {R}_{2},\ldots ,\) and \(\mathfrak {R}_{n}\) be p, q-QOFNs, \(\mathfrak {R}^{-}=\min \{\mathfrak {R}_{1}, \mathfrak {R}_{2},\ldots ,\mathfrak {R}_{n}\}\) and \(\mathfrak {R}^{+}=\max \{\mathfrak {R}_{1}, \mathfrak {R}_{2},\ldots ,\mathfrak {R}_{n}\}\). Then,
Proof
Since \(\mathfrak {R}^{-}=\min \{\mathfrak {R}_{1}, \mathfrak {R}_{2},\ldots ,\mathfrak {R}_{n}\}\) and \(\mathfrak {R}^{+}=\max \{\mathfrak {R}_{1}, \mathfrak {R}_{2},\ldots ,\mathfrak {R}_{n}\}\), therefore by using Eq. (8), we obtain
Similarly,
Thus, we get \(\mathfrak {R}^{-} \le p, q-QOFYPWG(\mathfrak {R}_{1},\mathfrak {R}_{2},\ldots ,\mathfrak {R}_{n}) \le \mathfrak {R}^{+}.\) \(\square \)
Property 3
(Monotonicity) Let \(\mathfrak {R}_{1},\mathfrak {R}_{2},\ldots ,\mathfrak {R}_{n}\) and \(\dot{\mathfrak {R}_{1}},\dot{\mathfrak {R}_{2}},\ldots ,\dot{\mathfrak {R}_{n}}\) be two families of p, q-QOFNs. If \(\mathfrak {R}_{h} \le \dot{\mathfrak {R}_{h}}\), where \(h=1,2,\ldots ,n\), then
Proof
By using Eq. (8), we obtain
Since \(\mathfrak {R}_{h} \le \dot{\mathfrak {R}_{h}}\), \(\forall h=1,2,\ldots , n,\) we obtain \(\otimes _{h=1}^{n}\mathfrak {R}_{h}^{\frac{w_{h}T_{h}}{\sum _{h=1}^{n}w_{h}T_{h}}} \le \otimes _{h=1}^{n}{\dot{\mathfrak {R}_{h}}}^{\frac{w_{h}T_{h}}{\sum _{h=1}^{n}w_{h}T_{h}}}\). Thus, we get \(p, q - QOFYPWG(\mathfrak {R}_{1},\mathfrak {R}_{2},\ldots ,\mathfrak {R}_{n}) \le p, q - QOFYPWG(\dot{\mathfrak {R}_{1}},\dot{\mathfrak {R}_{2}},\ldots ,\dot{\mathfrak {R}_{n}})\). \(\square \)
5 The proposed MAGDM approach based on the proposed p, q-QOFYPWG AO of p, q-QOFNs
In this section, we propose a novel MAGDM approach based on the proposed p, q-QOFYPWG AO under the p, q-QOFNs environment. Let \(H_{1},H_{2},\ldots ,\) and \(H_{m}\) are m alternatives and let \(\Phi _{1},\Phi _{2},\ldots ,\) and \(\Phi _{n}\) are n attributes. Let \(\Psi _{1},\Psi _{2},\ldots ,\Psi _{y}\) be the decision making experts (DMExs) with respective weights \(\varpi _{1},\varpi _{2},\ldots ,\varpi _{y}\), respectively, where \(\varpi _{j}\ge 0\), \(j=1,2,\ldots ,y\) and \(\sum _{j=1}^{y}\varpi _{j}= 1\). Each DMEx \(\Psi _{j}\) assesses the attribute \(\Phi _{h}\) of the alternative \(H_{e}\) by utilizing p, q-QOFN \(\tilde{\mathfrak {R}}^{j}_{eh}=\langle \tilde{\zeta }^{j}_{eh}, \tilde{\varrho }^{j}_{eh} \rangle \) to construct the decision matrix (DMx) \(\tilde{L}^{j}=(\tilde{\mathfrak {R}}^{j}_{eh})_{m\times n}\), shown as follows:
The proposed MAGDM approach involves the following steps:
-
Step 1:
Convert the DMXs \(\tilde{L}^1 = ({\tilde{\mathfrak {R}}^1_{eh}})_{m\times n} = (\langle {\tilde{\zeta }^1_{eh}},{\tilde{\varrho }^1_{eh}} \rangle )_{m\times n}, \tilde{L}^2 = ({\tilde{\mathfrak {R}}^2_{eh}})_{m\times n} = (\langle {\tilde{\zeta }^2_{eh}},{\tilde{\varrho }^2_{eh}} \rangle )_{m\times n}, \ldots \), \(\tilde{L}^y = ({\tilde{\mathfrak {R}}^y_{eh}})_{m\times n} = (\langle {\tilde{\zeta }^y_{eh}}, {\tilde{\varrho }^y_{eh}} \rangle )_{m\times n}\), into normalized DMxs (NDMxs) \({L}^1 = ({\mathfrak {R}^1_{eh}})_{m\times n} = (\langle {{\zeta }^1_{eh}},{{\varrho }^1_{eh}} \rangle )_{m\times n}, {L}^2 = ({{\mathfrak {R}}^2_{eh}})_{m\times n} = (\langle {{\zeta }^2_{eh}},{{\varrho }^2_{eh}} \rangle )_{m\times n}, \ldots \), \({L}^y = ({{\mathfrak {R}}^y_{eh}})_{m\times n} = (\langle {{\zeta }^y_{eh}}, {{\varrho }^y_{eh}} \rangle )_{m\times n}\) as follows:
$$\begin{aligned} {\mathfrak {R}^j_{eh}}= & {\left\{ \begin{array}{ll} \langle {\tilde{\zeta }^j_{eh}}, {\tilde{\varrho }^j_{eh}} \rangle :& \text { for benefit type attribute} \\ \langle {\tilde{\varrho }^j_{eh}}, {\tilde{\zeta }^j_{eh}}\rangle :& \text { for cost type attribute} \end{array}\right. }, \end{aligned}$$(10)where \(e = 1, 2, \ldots , m, h = 1, 2, \ldots , n\) and \( j = 1, 2, \ldots , y\).
-
Step 2:
Compute the values \(T^1_{eh},T^2_{eh}, \ldots \), and \(T^y_{eh}\) of p, q-QOFNs \(\mathfrak {R}^1_{eh},\mathfrak {R}^2_{eh},\ldots \), and \(\mathfrak {R}^y_{eh}\) appeared in NDMxs \({L}^1 = ({\mathfrak {R}^1_{eh}})_{m\times n} = (\langle {{\zeta }^1_{eh}},{{\varrho }^1_{eh}} \rangle )_{m\times n}, {L}^2 = ({{\mathfrak {R}}^2_{eh}})_{m\times n} = (\langle {{\zeta }^2_{eh}},{{\varrho }^2_{eh}} \rangle )_{m\times n}, \ldots \), and \({L}^y = ({{\mathfrak {R}}^y_{eh}})_{m\times n} = (\langle {{\zeta }^y_{eh}}, {{\varrho }^y_{eh}} \rangle )_{m\times n}\), respectively, to construct the matrices \(T^{1}=(T^{1}_{eh})_{m\times n}\), \(T^{2}=(T^{2}_{eh})_{m\times n}\), \(\ldots \), and \(T^{y}=(T^{y}_{eh})_{m\times n}\), as follows:
$$\begin{aligned} {T^j_{eh}}= & {\left\{ \begin{array}{ll} 1:& \text {if} j=1 \\ \prod _{a=1}^{j-1}S(\mathfrak {R}^{a}_{eh}):& \text {if} j=2,3, \ldots , y \end{array}\right. }, \end{aligned}$$(11)where \(S(\mathfrak {R}^{a}_{eh})=\frac{1+(\zeta ^a_{eh})^p-(\varrho ^a_{eh})^q}{2}\) is the score value of the p, q-QOFN \(\mathfrak {R}^a_{eh}\) which is obtained by Eq. (3), \(p\ge 1\), \(q\ge 1\), \(e = 1, 2, \ldots , m, h = 1, 2, \ldots , n\) and \( a = 1, 2, \ldots , y-1\).
-
Step 3:
Based on the obtained matrices \(T^{1}=(T^{1}_{eh})_{m\times n}\), \(T^{2}=(T^{2}_{eh})_{m\times n}\) and \(T^{y}=(T^{y}_{eh})_{m\times n}\) and the weights \(\varpi _{1},\varpi _{2},\ldots ,\varpi _{y}\) of the DMExs \(\Psi _{1},\Psi _{2},\ldots ,\Psi _{y}\), respectively, we compute the weights \(\varsigma ^{1}_{eh},\varsigma ^{2}_{eh},\ldots \), and \(\varsigma ^{y}_{eh}\) of p, q-QOFNs \(\mathfrak {R}^1_{eh},\mathfrak {R}^2_{eh},\ldots \), and \(\mathfrak {R}^y_{eh}\), respectively, to construct the weighted matrices \(W^{1}=(\varsigma ^{1}_{eh})_{m\times n}\), \(W^{2}=(\varsigma ^{2}_{eh})_{m\times n}, \ldots ,\) and \(W^{y}=(\varsigma ^{y}_{eh})_{m\times n}\), shown as follows:
$$\begin{aligned} \varsigma ^{j}_{eh}=\frac{\varpi _{j}T^{j}_{eh}}{\sum _{j=1}^{y}\varpi _{j}T^{j}_{eh}}, \end{aligned}$$(12)where \(e = 1, 2, \ldots , m\), \(h = 1, 2, \ldots , n\) and \( j = 1, 2, \ldots , y\).
-
Step 4:
Based on the obtained weights \(\varsigma ^{1}_{eh},\varsigma ^{2}_{eh},\ldots \), and \(\varsigma ^{y}_{eh}\) of p, q-QOFNs \(\mathfrak {R}^1_{eh},\mathfrak {R}^2_{eh},\ldots \), and \(\mathfrak {R}^y_{eh}\), respectively, and proposed p, q-QOFYPWG AO shown in Eq. (8), we aggregate the p, q-QOFNs \(\mathfrak {R}^1_{eh},\mathfrak {R}^2_{eh},\ldots \), and \(\mathfrak {R}^y_{eh}\) that appeared in NDMxs \({L}^1 = ({\mathfrak {R}^1_{eh}})_{m\times n}, {L}^2 = ({{\mathfrak {R}}^2_{eh}})_{m\times n}, \ldots \), \({L}^y = ({{\mathfrak {R}}^y_{eh}})_{m\times n} \) respectively, to get the aggregated p, q-QOFN \(\mathfrak {R}_{eh}=\langle \zeta _{eh}, \varrho _{eh} \rangle \) for constructing the collective DMx (CDMx) \({L}=({{\mathfrak {R}}_{eh}})_{m\times n}\), shown as follows:
$$\begin{aligned} \mathfrak {R}_{eh}= & p, q-QOFYPWG(\mathfrak {R}^1_{eh},\mathfrak {R}^2_{eh},\ldots ,\mathfrak {R}^y_{eh})\nonumber \\= & \left\langle \root p \of {1-\min \left\{ 1,\left( \sum _{j=1}^{y}\varsigma ^{j}_{eh}\left( 1-\left( \zeta ^{j}_{eh}\right) ^{p}\right) ^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} },\right. \nonumber \\ & \left. \root q \of {\min \left\{ 1,\left( \sum _{j=1}^{y}\varsigma ^{j}_{eh} \left( \varrho ^{j}_{eh}\right) ^{q\lambda }\right) ^{\frac{1}{\lambda }}\right\} } \right\rangle , \end{aligned}$$(13)where \(e = 1, 2, \ldots , m, h = 1, 2, \ldots , n\), \(p\ge 1\), \(q\ge 1\), and \(\lambda \in (0,\infty )\).
-
Step 5:
Calculate the value \(T_{eh}\) of the p, q-QOFNs \(\mathfrak {R}_{eh}\) appeared in CDMx \({L}=({{\mathfrak {R}}_{eh}})_{m\times n}\) to construct the matrix \(T=(T_{eh})_{m \times n}\), where
$$\begin{aligned} {T_{eh}}= & {\left\{ \begin{array}{ll} 1:& \text {if}\, t=1, \\ \prod _{t=1}^{h-1}S(\mathfrak {R}_{et}):& \text {if}\, h=2,3, \ldots , n, \end{array}\right. } \end{aligned}$$(14)\(S(\mathfrak {R}_{et})\) is the score value of the p, q-QOFN \(\mathfrak {R}_{et}\) obtained by using Eq. (3), \(e = 1, 2, \ldots , m; h=1,2,\ldots ,n; t = 1, 2, \ldots , n-1\).
-
Step 6:
Compute the entropy \(E_h\) of the attribute \(\Phi _h\) by using the p, q-QOFN \(\mathfrak {R}_{1\,h}\), \(\mathfrak {R}_{2\,h}\), \(\ldots \), \(\mathfrak {R}_{mh}\) appeared in \(h^{th}\) column of CDMx \({L}=({{\mathfrak {R}}_{eh}})_{m\times n}\), shown as follows:
$$\begin{aligned} E_h = {\frac{1}{m}}\sum _{e=1}^{m}\tan \left( \frac{\pi }{4}-\frac{|(\zeta _{eh})^p-(\varrho _{eh})^q|(1-(\pi _{eh})^l)}{4} \pi \right) ,\nonumber \\ \end{aligned}$$(15)where \(e=1,2,\ldots ,m\), \(h=1,2,\ldots ,n\), \((\pi _{eh})^l=1-(\zeta _{eh})^p-(\varrho _{eh})^q\), l is the LCM of p and q. Now by using the above entropy, we compute the weights \(w_1\), \(w_2\), \(\ldots \), \(w_n\) of the attributes \(\Phi _{1},\Phi _{2},\ldots ,\Phi _{n}\), respectively, as follows:
$$\begin{aligned} w_h=\frac{1-E_h}{n-\sum _{h=1}^{n}E_h}, \end{aligned}$$(16)where \(w_{h}\ge 0\), \(h=1,2,\ldots ,n\) and \(\sum _{h=1}^{n}w_{h}= 1\).
-
Step 7:
By utilizing the obtained weights \(w_1,w_2,\ldots ,w_n\) and obtained matrix \(T=(T_{eh})_{m \times n}\), we calculate the weights \(\varsigma _{e1},\varsigma _{e2},\ldots \), and \(\varsigma _{en}\) of p, q-QOFNs \(\mathfrak {R}_{e1},\mathfrak {R}_{e2},\ldots ,\) and \(\mathfrak {R}_{en}\), respectively, to construct the weighted matrix \(W=(\varsigma _{eh})_{m\times n}\), where
$$\begin{aligned} \varsigma _{eh}=\frac{w_{h}T_{eh}}{\sum _{h=1}^{n}w_{h}T_{eh}}, \end{aligned}$$(17)\(e = 1, 2, \ldots , m\) and \(h = 1, 2, \ldots , n\).
-
Step 8:
Based on the proposed p, q-QOFYPWG AO given in Eq. (8), we aggregate the p, q-QOFNs \(\mathfrak {R}_{e1},\mathfrak {R}_{e2},\ldots ,\) and \(\mathfrak {R}_{en}\) which appeared in the \(h^{th}\) row of the CDMx \({L}=({{\mathfrak {R}}_{eh}})_{m\times n}\) to obtain the overall p, q-QOFN \(\mathfrak {R}_{e}= \langle \zeta _{e}, \varrho _{e} \rangle \) of alternatives \(H_{e}\), shown as follows:
$$\begin{aligned} \mathfrak {R}_{e}= & p, q-QOFYPWG(\mathfrak {R}_{e1},\mathfrak {R}_{e2},\ldots ,\mathfrak {R}_{en})\nonumber \\= & \Bigg \langle \root p \of {1-\min \left\{ 1,\left( \sum _{h=1}^{n}\varsigma _{eh}\left( 1-(\zeta _{eh})^{p}\right) ^{\lambda }\right) ^{\frac{1}{\lambda }}\right\} },\nonumber \\ & \root q \of {\min \left\{ 1,\left( \sum _{h=1}^{n}\varsigma _{eh} \left( \varrho _{eh}\right) ^{q\lambda }\right) ^{\frac{1}{\lambda }}\right\} } \Bigg \rangle , \end{aligned}$$(18)where, \(e=1,2,\ldots ,m\), \(p,q\ge 1\) and \(\lambda \in (0,\infty ).\)
-
Step 9:
By using the Eq.(3), we calculate the score values \(S(\mathfrak {R}_{1})\), \(S(\mathfrak {R}_{2})\), \(\ldots \), and \(S(\mathfrak {R}_{m})\) of the overall p, q-QOFNs \(\mathfrak {R}_{1}= \langle \zeta _{1}, \varrho _{1} \rangle \), \(\mathfrak {R}_{2}= \langle \zeta _{2}, \varrho _{2} \rangle \), \(\ldots \), \(\mathfrak {R}_{m}= \langle \zeta _{m}, \varrho _{m} \rangle \) of the alternative \(H_{1}\), \(H_{2}\), ..., and \(H_{m}\), respectively, shown as follows:
$$\begin{aligned} S(\mathfrak {R}_{e})=\frac{1+(\zeta _{e})^p-(\varrho _{e})^q}{2}, \end{aligned}$$(19)where \(S(\mathfrak {R}_{e})\in [0,1]\) and \(e=1,2,\ldots ,m\).
-
Step 10:
If \(S(\mathfrak {R}_{a})>S(\mathfrak {R}_{b})\), then based on Definition 7, the preference order (PO) between the alternatives \(\mathfrak {R}_a\) and \(\mathfrak {R}_b\) is \(``\mathfrak {R}_a\succ \mathfrak {R}_b",\) where \(a=1,2,\ldots ,m\), \(b=1,2,\ldots ,m\) and \(a\ne b\). If \(S(\mathfrak {R}_{a})=S(\mathfrak {R}_{b})\), then, by using Eq.(4), we compute the accuracy values \(A(\mathfrak {R}_{a})=(\zeta _{a})^p+(\varrho _{a})^q\) and \(A(\mathfrak {R}_{b})=(\zeta _{b})^p+(\varrho _{b})^q\) of the overall p, q-QOFNs \(\mathfrak {R}_{a}= \langle \zeta _{a}, \varrho _{a} \rangle \) and \(\mathfrak {R}_{b}= \langle \zeta _{b}, \varrho _{b} \rangle \), respectively. If \(A(\mathfrak {R}_{a})>A(\mathfrak {R}_{b})\), then according to Definition 7, the PO between the alternatives \(\mathfrak {R}_a\) and \(\mathfrak {R}_b\) is \(``\mathfrak {R}_a\succ \mathfrak {R}_b"\). If \(S(\mathfrak {R}_{a})=S(\mathfrak {R}_{b})\) and \(A(\mathfrak {R}_{a})=A(\mathfrak {R}_{b})\), then alternatives \(\mathfrak {R}_a\) and \(\mathfrak {R}_b\) have the same PO, where \(a\ne b\). Thus, we get the PO of the alternatives \(\mathfrak {R}_1,\mathfrak {R}_2,\ldots ,\) and \(\mathfrak {R}_e\) and select the best choice.
Example 3
(Garg 2020) The government wants to prevent urban migration by selecting an ideal company for creating economic opportunities in rural areas of Jharkhand. Let the four attributes outlined by the government for selecting companies are: \(\Phi _{1}\) (“Focusing on technical capability”), \(\Phi _{2}\) (“Financial status”), \(\Phi _{3}\) (“Company background”) and \(\Phi _{4}\) (“References from previous projects”). Let the five companies \(H_{1},H_{2},H_{3},H_{4}\) and \(H_{5}\) as alternatives have shown keen interest in the project. Three DMExs \(\Psi _{1},\Psi _{2}\) and \(\Psi _{3}\) evaluate the companies \(H_{1},H_{2},H_{3},H_{4}\) and \(H_{5}\) towards the attributes \(\Phi _{1}\), \(\Phi _{2}\), \(\Phi _{3}\) and \(\Phi _{4}\). The weights of the DMExs \(\Psi _{1},\Psi _{2}\) and \(\Psi _{3}\) are \(\varpi _{1}=0.35, \varpi _{2}=0.40\) and \(\varpi _{3}=0.25\), respectively. Each DMEx \(\Psi _{j}\) assesses the attribute \(\Phi _{h}\) of the alternative \(H_{e}\) by utilizing p, q-QOFN \(\tilde{\mathfrak {R}}^{j}_{eh}=\langle \tilde{\zeta }^{j}_{eh}, \tilde{\varrho }^{j}_{eh} \rangle \), where \(j=1,2,3\), \(e=1,2,3,4,5\) and \(h=1,2,3,4\), to construct the DMx \(\tilde{L}^{j}=(\tilde{\mathfrak {R}}^{j}_{eh})_{5\times 4}\), shown as follows:
In the following, we utilize the proposed MAGDM approach to solve this MAGDM problem.
-
Step 1:
Since all the attributes \(\Phi _{1}\), \(\Phi _{2}\),\(\Phi _{3}\) and \(\Phi _{4}\) are benefit type, by using Eq. (10), we get NDMxs \({L}^1 = (\tilde{\mathfrak {R}}_{eh}^1)_{5 \times 4} =({\mathfrak {R}}_{eh}^1)_{5 \times 4} = (\langle {{\zeta }_{eh}^1}, {{\varrho }_{eh}^1}\rangle )_{5\times 4}\), \({L}^2 = (\tilde{\mathfrak {R}}_{eh}^2)_{5 \times 4} =({\mathfrak {R}}_{eh}^2)_{5 \times 4} = (\langle {{\zeta }_{eh}^2},{{\varrho }_{eh}^2}\rangle )_{5\times 4}\) and \({L}^3 = (\tilde{\mathfrak {R}}_{eh}^3)_{5 \times 4} =({\mathfrak {R}}_{eh}^3)_{5 \times 4} = (\langle {{\zeta }_{eh}^3},{{\varrho }_{eh}^3}\rangle )_{5\times 4}\).
-
Step 2:
By using Eq. (11), we calculate the values of \(T^{1}_{eh},T^{2}_{eh}\) and \(T^{3}_{eh}\) of p, q-QOFNs \(\mathfrak {R}^{1}_{eh},\mathfrak {R}^{2}_{eh}\) and \(\mathfrak {R}^{3}_{eh}\) of the NDMx \({L}^1 =({\mathfrak {R}}_{eh}^1)_{5 \times 4}\), \({L}^2 =({\mathfrak {R}}_{eh}^2)_{5 \times 4}\), and \({L}^3 =({\mathfrak {R}}_{eh}^3)_{5 \times 4}\), respectively, to obtain the matrices \(T^{1}=(T^{1}_{eh})_{5\times 4}\), \(T^{2}=(T^{2}_{eh})_{5\times 4}\) and \(T^{3}=(T^{3}_{eh})_{5\times 4}\), where, \(p=1\), \(q=4\), \(\lambda =1\),
-
Step 3:
By using Eq. (12), the obtained matrices \(T^{1}, T^{2}\) and \(T^{3}\) and the weights \(\varpi _{1}=0.35, \varpi _{2}=0.40\) and \(\varpi _{3}=0.25\) of the DMExs \(\Psi ^1,\Psi ^2\) and \(\Psi ^3\), respectively, we calculate the weights \(\varsigma ^{1}_{eh}, \varsigma ^{2}_{eh}\) and \(\varsigma ^{3}_{eh}\) of p, q-QOFNs \(\mathfrak {R}^{1}_{eh},\mathfrak {R}^{2}_{eh}\) and \(\mathfrak {R}^{3}_{eh}\), respectively, to construct the weighted matrices \(W^{1}=(\varsigma ^{1}_{eh})_{5\times 4}\), \(W^{2}=(\varsigma ^{2}_{eh})_{5\times 4}\) and \(W^{3}=(\varsigma ^{3}_{eh})_{5\times 4}\), where \(p=1\), \(q=4\), \(e=1,2,3,4,5\) and \(h=1,2,3,4\),
-
Step 4:
By using Eq. (13), we obtain the aggregated p, q-QOFN \(\mathfrak {R}_{eh}=\langle {{\zeta }_{eh}},{{\varrho }_{eh}}\rangle \) by aggregating the p, q-QOFNs \(\mathfrak {R}^{1}_{eh}=\langle {{\zeta }_{eh}^1},{{\varrho }_{eh}^1}\rangle \), \(\mathfrak {R}^{2}_{eh}=\langle {{\zeta }_{eh}^2},{{\varrho }_{eh}^2}\rangle \) and \(\mathfrak {R}^{3}_{eh}=\langle {{\zeta }_{eh}^3},{{\varrho }_{eh}^3}\rangle \) that appeared in the NDMxs \(L^{1}=(\mathfrak {R}^{1}_{eh})_{5\times 4}\), \(L^{2}=(\mathfrak {R}^{2}_{eh})_{5\times 4}\) and \(L^{3}=(\mathfrak {R}^{3}_{eh})_{5\times 4}\), respectively, to construct the CDMx \(L=(\mathfrak {R}_{eh})_{5\times 4}=\langle \zeta _{eh}, \varrho _{eh} \rangle _{5\times 4}\), where \(\lambda =1\), \(p=1\) and \(q=4\),
-
Step 5:
By using Eq. (14), we calculate the value \(T_{eh}\) of the aggregated p, q-QOFN \(\mathfrak {R}_{eh}\), to get the matrix \(T=(T_{eh})_{5\times 4}\), where \(e=1,2,3,4,5\) and \(h=1,2,3,4\),
-
Step 6:
By using Eq. (15), we calculate the entropies \(E_1\), \(E_2\), \(E_3\) and \(E_4\) of the attributes \(\Phi _1\), \(\Phi _2\), \(\Phi _3\) and \(\Phi _4\), respectively, where \(E_1= 0.6890\), \(E_2= 0.7785\), \(E_3=0.6444\) and \(E_4=0.6428\). Then, by using Eq. (16), we calculate the weights \(w_1\), \(w_2\), \(w_3\), \(w_4\) of the attributes \(\Phi _1\), \(\Phi _2\), \(\Phi _3\) and \(\Phi _4\), respectively, where \(w_{1}= 0.2497,w_{2}= 0.1779,w_{3}=0.2855\) and \(w_{4}=0.2869\).
-
Step 7:
By using Eq. (17), the obtained matrix T and the weights \(w_{1}= 0.2497,w_{2}= 0.1779,w_{3}=0.2855\) and \(w_{4}=0.2869\) of the attributes \(\Phi _{1}, \Phi _{2}, \Phi _{3}\) and \(\Phi _{4}\), respectively, we calculate the weight \(\varsigma _{eh}\) of p, q-QOFN \(R_{eh}\), to construct the weighted matrix \(W=(\varsigma _{eh})_{5 \times 4}\), where \(e=1,2,3,4,5\), \(h=1,2,3,4\),
-
Step 8:
By using Eq. (18) and obtained weight matrix \(W=(\varsigma _{eh})_{5 \times 4}\), we obtain the overall aggregated p, q-QOFN \(\mathfrak {R}_{e}=\langle \zeta _{e}, \varrho _{e} \rangle \) of the alternative \(H_{e}\), where \(e=1,2,3,4,5\), \(\zeta _{1} = 0.3304\), \(\varrho _{1} = 0.5466\), \(\zeta _{2} = 0.5194\), \(\varrho _{2} = 0.2573\), \(\zeta _{3} = 0.4835\), \(\varrho _{3} = 0.3177\),\(\zeta _{4} = 0.4535\), \(\varrho _{4} = 0.4286\), \(\zeta _{5} = 0.5731\), \(\varrho _{5} = 0.3172\), \(\mathfrak {R}_{1}= \langle 0.3304, 0.5466 \rangle \), \(\mathfrak {R}_{2}= \langle 0.5194, 0.2573 \rangle \), \(\mathfrak {R}_{3}= \langle 0.4835, 0.3177 \rangle \), \(\mathfrak {R}_{4}= \langle 0.4535, 0.4286 \rangle \) and \(\mathfrak {R}_{5}= \langle 0.5731,0.3172 \rangle \).
-
Step 9:
By using Eq. (19), we calculate the score values \(S(\mathfrak {R}_{1})\),\(S(\mathfrak {R}_{2})\), \(S(\mathfrak {R}_{3})\), \(S(\mathfrak {R}_{4})\) and \(S(\mathfrak {R}_{5})\) of the overall aggregated p, q-QOFNs \(\mathfrak {R}_{1}= \langle 0.3304, 0.5466 \rangle \), \(\mathfrak {R}_{2}= \langle 0.5194, 0.2573 \rangle \), \(\mathfrak {R}_{3}= \langle 0.4835, 0.3177 \rangle \), \(\mathfrak {R}_{4}= \langle 0.4535, 0.4286 \rangle \) and \(\mathfrak {R}_{5}= \langle 0.5731, 0.3172 \rangle \), respectively, where \(S(\mathfrak {R}_{1})=0.6206\),\(S(\mathfrak {R}_{2})=0.7575\), \(S(\mathfrak {R}_{3})=0.7366\), \(S(\mathfrak {R}_{4})=0.7099\) and \(S(\mathfrak {R}_{5})=0.7815\).
-
Step 10:
Because \(S(\mathfrak {R}_{5})>S(\mathfrak {R}_{2})>S(\mathfrak {R}_{3})>S(\mathfrak {R}_{4})>S(\mathfrak {R}_{1})\), where \(S(\mathfrak {R}_{1})=0.6206\),\(S(\mathfrak {R}_{2})=0.7575\), \(S(\mathfrak {R}_{3})=0.7366\), \(S(\mathfrak {R}_{4})=0.7099\) and \(S(\mathfrak {R}_{5})=0.7815\), the PO of the alternatives \(H_{1}, H_{2}, H_{3}, H_{4}\) and \(H_{5}\) is \(``H_{5}\succ H_{2}\succ H_{3}\succ H_{4}\succ H_{1}"\). Thus, \(H_{5}\) is the best alternative.
Table 1 presents a comparison of the POs of the alternatives \(H_{1}, H_{2}, H_{3}, H_{4}\) and \(H_{5}\) obtained by various MAGDM approaches for Example 3. From Table 1, it is clear that Garg’s MAGDM approach (Garg 2020), Seikh and Mandal’s MAGDM approach (Seikh and Mandal 2022), Rahim et al.’s MAGDM approach (Rahim et al. 2023a), Ahmad et al.’s MAGDM approach (Ahmad et al. 2024) and the proposed MAGDM approach obtain the same PO \(``H_{5}\succ H_{2}\succ H_{3}\succ H_{4}\succ H_{1}"\) of the alternatives \(H_{1}, H_{2}, H_{3}, H_{4}\) and \(H_{5}\).
Example 4
Let \(H_{1}, H_{2}, H_{3}\) and \(H_{4}\) be four alternatives and \(\Phi _{1}\), \(\Phi _{2}\), \(\Phi _{3}\) and \(\Phi _{4}\) be four attributes. The weights of the DMExs \(\Psi _{1},\Psi _{2}\) and \(\Psi _{3}\) are \(\varpi _{1}=0.40, \varpi _{2}=0.20\) and \(\varpi _{3}=0.40\), respectively. Each DMEx \(\Psi _{j}\) assesses the attribute \(\Phi _{h}\) of the alternative \(H_{e}\) by utilizing p, q-QOFN \(\tilde{\mathfrak {R}}^{j}_{eh}=\langle \tilde{\zeta }^{j}_{eh}, \tilde{\varrho }^{j}_{eh} \rangle \), where \(j=1,2,3\), \(e=1,2,3,4\) and \(h=1,2,3,4\), to construct the DMx \(\tilde{L}^{j}=(\tilde{\mathfrak {R}}^{j}_{eh})_{4\times 4}\), shown as follows:
In the following, we utilize the proposed MAGDM approach to solve this MAGDM problem.
-
Step 1:
Since all the attributes \(\Phi _{1}\), \(\Phi _{2}\), \(\Phi _{3}\) and \(\Phi _{4}\) are benefit type, by using Eq. (10), we get NDMxs \({L}^1 = (\tilde{\mathfrak {R}}_{eh}^1)_{4 \times 4} =({\mathfrak {R}}_{eh}^1)_{4 \times 4} = (\langle {{\zeta }_{eh}^1}, {{\varrho }_{eh}^1}\rangle )_{4\times 4}\), \({L}^2 = (\tilde{\mathfrak {R}}_{eh}^2)_{4 \times 4} =({\mathfrak {R}}_{eh}^2)_{4 \times 4} = (\langle {{\zeta }_{eh}^2},{{\varrho }_{eh}^2}\rangle )_{4\times 4}\) and \({L}^3 = (\tilde{\mathfrak {R}}_{eh}^3)_{4 \times 4} =({\mathfrak {R}}_{eh}^3)_{4 \times 4} = (\langle {{\zeta }_{eh}^3},{{\varrho }_{eh}^3}\rangle )_{4\times 4}\).
-
Step 2:
By using Eq. (11), we calculate the values of \(T^{1}_{eh}, T^{2}_{eh}\) and \(T^{3}_{eh}\) of p, q-QOFNs \(\mathfrak {R}^{1}_{eh}, \mathfrak {R}^{2}_{eh}\) and \(\mathfrak {R}^{3}_{eh}\) of the NDMx \(L^1=(\mathfrak {R}^1_{eh})_{4\times 4}\), \(L^2=(\mathfrak {R}^2_{eh})_{4\times 4}\) and \(L^3=(\mathfrak {R}^3_{eh})_{4\times 4}\), respectively, to obtain the matrices \(T^{1}=(T^{1}_{eh})_{4\times 4}\), \(T^{2}=(T^{2}_{eh})_{4\times 4}\) and \(T^{3}=(T^{3}_{eh})_{4\times 4}\), where \(p=3\), \(q=3\), \(\lambda =1\),
-
Step 3:
By using Eq. (12), the obtained matrices \(T^{1}, T^{2}\) and \(T^{3}\) and the weights \(\varpi _{1}=0.40, \varpi _{2}=0.20\) and \(\varpi _{3}=0.40\) of the DMExs \(\Psi ^1, \Psi ^2\) and \(\Psi ^3\), respectively, we calculate the weights \(\varsigma ^{1}_{eh}, \varsigma ^{2}_{eh}\) and \(\varsigma ^{3}_{eh}\) of p, q-QOFNs \(\mathfrak {R}^{1}_{eh}, \mathfrak {R}^{2}_{eh}\) and \(\mathfrak {R}^{3}_{eh}\), respectively, to construct the weighted matrices \(W^{1}=(\varsigma ^{1}_{eh})_{4\times 4}\), \(W^{2}=(\varsigma ^{2}_{eh})_{4\times 4}\) and \(W^{3}=(\varsigma ^{3}_{eh})_{4\times 4}\), where \(p=3,q=3, e=1,2,3,4\) and \(h=1,2,3,4\),
-
Step 4:
By using Eq. (13), we obtain the aggregated p, q-QOFN \(\mathfrak {R}_{eh}=\langle {{\zeta }_{eh}},{{\varrho }_{eh}}\rangle \) by aggregating the p, q-QOFNs \(\mathfrak {R}^{1}_{eh}=\langle {{\zeta }_{eh}^1},{{\varrho }_{eh}^1}\rangle \), \(\mathfrak {R}^{2}_{eh}=\langle {{\zeta }_{eh}^2},{{\varrho }_{eh}^2}\rangle \) and \(\mathfrak {R}^{3}_{eh}=\langle {{\zeta }_{eh}^3},{{\varrho }_{eh}^3}\rangle \) that appeared in the NDMxs \(L^{1}=(\mathfrak {R}^{1}_{eh})_{4\times 4}\), \(L^{2}=(\mathfrak {R}^{2}_{eh})_{4\times 4}\) and \(L^{3}=(\mathfrak {R}^{3}_{eh})_{4\times 4}\), respectively, to construct the CDMx \(L=(\mathfrak {R}_{eh})_{4\times 4}=\langle \zeta _{eh}, \varrho _{eh} \rangle _{4\times 4}\), where \(\lambda =1, p=3\) and \(q=3\),
-
Step 5:
By using Eq. (14), we calculate the value \(T_{eh}\) of the aggregated p, q-QOFN \(\mathfrak {R}_{eh}\), to get the matrix \(T=(T_{eh})_{4\times 4}\), where \(e=1,2,3,4\) and \(h=1,2,3,4\),
-
Step 6:
By using Eq. (15), we calculate the entropies \(E_1\), \(E_2\), \(E_3\) and \(E_4\) of the attributes \(\Phi _1\), \(\Phi _2\), \(\Phi _3\) and \(\Phi _4\), respectively, where \(E_1= 0.9336\), \(E_2= 0.9868\), \(E_3=0.9480\) and \(E_4=0.9614\). Then, by using Eq. (16), we calculate the weights \(w_1\), \(w_2\), \(w_3\), \(w_4\) of the attributes \(\Phi _1\), \(\Phi _2\), \(\Phi _3\) and \(\Phi _4\), respectively, where \(w_{1}= 0.3902,w_{2}= 0.0775,w_{3}=0.3054\) and \(w_{4}=0.2268\).
-
Step 7:
By using Eq. (17), the obtained matrix T and the weights \(w_{1}= 0.3902,w_{2}= 0.0775,w_{3}=0.3054\) and \(w_{4}=0.2268\) of the attributes \(\Phi _{1}, \Phi _{2}, \Phi _{3}\) and \(\Phi _{4}\), respectively, we calculate the weight \(\varsigma _{eh}\) of p, q-QOFN \(R_{eh}\), to construct the weighted matrix \(W=(\varsigma _{eh})_{4 \times 4}\), where \(e=1,2,3,4\), \(h=1,2,3,4\),
-
Step 8:
By using Eq. (18) and obtained weight matrix \(W=(\varsigma _{eh})_{4 \times 4}\), we obtain the overall aggregated p, q-QOFN \(\mathfrak {R}_{e}=\langle \zeta _{e}, \varrho _{e} \rangle \) of the alternative \(H_{e}\), where \(e=1,2,3,4\), \(\zeta _{1} = 0.4830\), \(\varrho _{1} = 0.4381\), \(\zeta _{2} = 0.2770\), \(\varrho _{2} = 0.6639\), \(\zeta _{3} = 0.3985\), \(\varrho _{3} = 0.3443\),\(\zeta _{4} = 0.6231\), \(\varrho _{4} = 0.2617\), \(\mathfrak {R}_{1}= \langle 0.4830, 0.4381 \rangle \), \(\mathfrak {R}_{2}= \langle 0.2770, 0.6639 \rangle \), \(\mathfrak {R}_{3}= \langle 0.3985, 0.3443 \rangle \) and \(\mathfrak {R}_{4}= \langle 0.6231, 0.2617 \rangle \).
-
Step 9:
By using Eq. (19), we calculate the score values \(S(\mathfrak {R}_{1})\), \(S(\mathfrak {R}_{2})\), \(S(\mathfrak {R}_{3})\) and \(S(\mathfrak {R}_{4})\) of the overall aggregated p, q-QOFNs \(\mathfrak {R}_{1}= \langle 0.4830, 0.4381 \rangle \), \(\mathfrak {R}_{2}= \langle 0.2770, 0.6639 \rangle \), \(\mathfrak {R}_{3}= \langle 0.3985, 0.3443 \rangle \) and \(\mathfrak {R}_{4}= \langle 0.6231, 0.2617 \rangle \), respectively, where \(S(\mathfrak {R}_{1})=0.5143\),\(S(\mathfrak {R}_{2})=0.3643\), \(S(\mathfrak {R}_{3})=0.5112\) and \(S(\mathfrak {R}_{4})=0.6120\).
-
Step 10:
Because \(S(\mathfrak {R}_{4})>S(\mathfrak {R}_{1})>S(\mathfrak {R}_{3})>S(\mathfrak {R}_{2})\), where \(S(\mathfrak {R}_{1})=0.5143\),\(S(\mathfrak {R}_{2})=0.3643\), \(S(\mathfrak {R}_{3})=0.5112\) and \(S(\mathfrak {R}_{4})=0.6120\), the PO of the alternatives \(H_{1}, H_{2}, H_{3}\) and \(H_{4}\) is \(``H_{4}\succ H_{1}\succ H_{3}\succ H_{2}"\). Thus, \(H_{4}\) is the best alternative.
Table 2 presents a comparison of the POs of the alternatives \(H_{1}, H_{2}, H_{3}\) and \(H_{4}\) obtained by various MAGDM approaches for Example 4. From Table 2, it is clear that Seikh and Mandal’s MAGDM approach (Seikh and Mandal 2022) and Ahmad et al.’s MAGDM approach (Ahmad et al. 2024) cannot handle this MAGDM problem because it get the indeterminant form in the intermediate steps while solving this MAGDM problem. However, Garg’s MAGDM approach (Garg 2020), Rahim et al.’s MAGDM approach (Rahim et al. 2023a) and the proposed MAGDM approach obtain the same PO \(``H_{4}\succ H_{1}\succ H_{3}\succ H_{2}"\) for the alternatives \(H_{1}, H_{2}, H_{3}\) and \(H_{4}\). Therefore, the proposed MAGDM approach can overcome the shortcomings of Seikh and Mandal’s MAGDM approach (Seikh and Mandal 2022) and Ahmad et al.’s MAGDM approach (Ahmad et al. 2024) in this case.
Example 5
Let \(H_{1}, H_{2}, H_{3}\) and \(H_{4}\) be four alternatives and \(\Phi _{1}\), \(\Phi _{2}\), \(\Phi _{3}\) and \(\Phi _{4}\) be four attributes. The weights of the DMExs \(\Psi _{1}, \Psi _{2}\) and \(\Psi _{3}\) are \(\varpi _{1}=0.40, \varpi _{2}=0.40\) and \(\varpi _{3}=0.20\), respectively. Each DMEx \(\Psi _{j}\) assesses the attribute \(\Phi _{h}\) of the alternative \(H_{e}\) by utilizing p, q-QOFN \(\tilde{\mathfrak {R}}^{j}_{eh}=\langle \tilde{\zeta }^{j}_{eh}, \tilde{\varrho }^{j}_{eh} \rangle \), where \(j=1,2,3\), \(e=1,2,3,4\) and \(h=1,2,3,4\), to construct the DMx \(\tilde{L}^{j}=(\tilde{\mathfrak {R}}^{j}_{eh})_{4\times 4}\), shown as follows:
In the following, we utilize the proposed MAGDM approach to solve this MAGDM problem.
-
Step 1:
Since all the attributes \(\Phi _{1}\), \(\Phi _{2}\), \(\Phi _{3}\) and \(\Phi _{4}\) are benefit type, by using Eq. (10), we get NDMxs \({L}^1 = (\tilde{\mathfrak {R}}_{eh}^1)_{4 \times 4} =({\mathfrak {R}}_{eh}^1)_{4 \times 4} = (\langle {{\zeta }_{eh}^1}, {{\varrho }_{eh}^1}\rangle )_{4\times 4}\), \({L}^2 = (\tilde{\mathfrak {R}}_{eh}^2)_{4 \times 4} =({\mathfrak {R}}_{eh}^2)_{4 \times 4} = (\langle {{\zeta }_{eh}^2},{{\varrho }_{eh}^2}\rangle )_{4\times 4}\) and \({L}^3 = (\tilde{\mathfrak {R}}_{eh}^3)_{4 \times 4} =({\mathfrak {R}}_{eh}^3)_{4 \times 4} = (\langle {{\zeta }_{eh}^3},{{\varrho }_{eh}^3}\rangle )_{4\times 4}\).
-
Step 2:
By using Eq. (11), we calculate the values of \(T^{1}_{eh}, T^{2}_{eh}\) and \(T^{3}_{eh}\) of p, q-QOFNs \(\mathfrak {R}^{1}_{eh}, \mathfrak {R}^{2}_{eh}\) and \(\mathfrak {R}^{3}_{eh}\) of the NDMx \({L}^1 =({\mathfrak {R}}_{eh}^1)_{4 \times 4}\), \({L}^2 =({\mathfrak {R}}_{eh}^2)_{4 \times 4}\) and \({L}^3 =({\mathfrak {R}}_{eh}^3)_{4 \times 4}\) respectively, to obtain the matrices \(T^{1}=(T^{1}_{eh})_{4\times 4}\), \(T^{2}=(T^{2}_{eh})_{4\times 4}\) and \(T^{3}=(T^{3}_{eh})_{4\times 4}\), where \(p=3\), \(q=3\), \(\lambda =1\),
-
Step 3:
By using Eq. (12), the obtained matrices \(T^{1}, T^{2}\) and \(T^{3}\) and the weights \(\varpi _{1}=0.40, \varpi _{2}=0.40\) and \(\varpi _{3}=0.20\) of the DMExs \(\Psi ^1,\Psi ^2\) and \(\Psi ^3\), respectively, we calculate the weights \(\varsigma ^{1}_{eh}, \varsigma ^{2}_{eh}\) and \(\varsigma ^{3}_{eh}\) of p, q-QOFNs \(\mathfrak {R}^{1}_{eh},\mathfrak {R}^{2}_{eh}\) and \(\mathfrak {R}^{3}_{eh}\), respectively, to construct the weighted matrices \(W^{1}=(\varsigma ^{1}_{eh})_{4\times 4}\), \(W^{2}=(\varsigma ^{2}_{eh})_{4\times 4}\) and \(W^{3}=(\varsigma ^{3}_{eh})_{4\times 4}\), where \(p=3,q=3, e=1,2,3,4\) and \(h=1,2,3,4\),
-
Step 4:
By using Eq. (13), we obtain the aggregated p, q-QOFN \(\mathfrak {R}_{eh}=\langle {{\zeta }_{eh}},{{\varrho }_{eh}}\rangle \) by aggregating the p, q-QOFNs \(\mathfrak {R}^{1}_{eh}=\langle {{\zeta }_{eh}^1},{{\varrho }_{eh}^1}\rangle \), \(\mathfrak {R}^{2}_{eh}=\langle {{\zeta }_{eh}^2},{{\varrho }_{eh}^2}\rangle \) and \(\mathfrak {R}^{3}_{eh}=\langle {{\zeta }_{eh}^3},{{\varrho }_{eh}^3}\rangle \) that appeared in the NDMxs \(L^{1}=(\mathfrak {R}^{1}_{eh})_{4\times 4}\), \(L^{2}=(\mathfrak {R}^{2}_{eh})_{4\times 4}\) and \(L^{3}=(\mathfrak {R}^{3}_{eh})_{4\times 4}\), respectively, to construct the CDMx \(L=(\mathfrak {R}_{eh})_{4\times 4}=\langle \zeta _{eh}, \varrho _{eh} \rangle _{4\times 4}\), where \(\lambda =1, p=3\) and \(q=3\),
-
Step 5:
By using Eq. (14), we calculate the value \(T_{eh}\) of the aggregated p, q-QOFN \(\mathfrak {R}_{eh}\), to get the matrix \(T=(T_{eh})_{4\times 4}\), where \(e=1,2,3,4\) and \(h=1,2,3,4\),
-
Step 6:
By using Eq. (15), we calculate the entropies \(E_1\), \(E_2\), \(E_3\) and \(E_4\) of the attributes \(\Phi _1\), \(\Phi _2\), \(\Phi _3\) and \(\Phi _4\), respectively, where \(E_1= 0.7313\), \(E_2= 0.9963\), \(E_3=0.9781\) and \(E_4=0.9634\). Then, by using Eq. (16), we calculate the weights \(w_1\), \(w_2\), \(w_3\), \(w_4\) of the attributes \(\Phi _1\), \(\Phi _2\), \(\Phi _3\) and \(\Phi _4\), respectively, where \(w_{1}= 0.8119,w_{2}= 0.0113,w_{3}=0.0662\) and \(w_{4}=0.1106\).
-
Step 7:
By using Eq. (17), the obtained matrix T and the weights \(w_{1}= 0.8119,w_{2}= 0.0113,w_{3}=0.0662\) and \(w_{4}=0.1106\) of the attributes \(\Phi _{1}, \Phi _{2}, \Phi _{3}\) and \(\Phi _{4}\), respectively, we calculate the weight \(\varsigma _{eh}\) of p, q-QOFN \(\mathfrak {R}_{eh}\), to construct the weighted matrix \(W=(\varsigma _{eh})_{4 \times 4}\), where \(e=1,2,3,4\) and \(h=1,2,3,4\),
-
Step 8:
By using Eq. (18) and obtained weight matrix \(W=(\varsigma _{eh})_{4 \times 4}\), we obtain the overall aggregated p, q-QOFN \(\mathfrak {R}_{e}=\langle \zeta _{e}, \varrho _{e} \rangle \), of the alternative \(H_{e}\), where \(e=1,2,3,4\), \(\zeta _{1} = 0.3722\), \(\varrho _{1} = 0.3619\), \(\zeta _{2} = 0.1249\), \(\varrho _{2} = 0.5336\), \(\zeta _{3} = 0\), \(\varrho _{3} = 1\),\(\zeta _{4} = 0.5399\), \(\varrho _{4} = 0.1361\), \(\mathfrak {R}_{1}= \langle 0.3722, 0.3619 \rangle \), \(\mathfrak {R}_{2}= \langle 0.1249, 0.5336 \rangle \), \(\mathfrak {R}_{3}= \langle 0, 1 \rangle \) and \(\mathfrak {R}_{4}= \langle 0.5399, 0.1361 \rangle \).
-
Step 9:
By using Eq. (19), we calculate the score values \(S(\mathfrak {R}_{1})\),\(S(\mathfrak {R}_{2})\), \(S(\mathfrak {R}_{3})\) and \(S(\mathfrak {R}_{4})\) of the overall aggregated p, q-QOFNs \(\mathfrak {R}_{1}= \langle 0.3722, 0.3619 \rangle \), \(\mathfrak {R}_{2}= \langle 0.1249, 0.5336 \rangle \), \(\mathfrak {R}_{3}= \langle 0, 1 \rangle \) and \(\mathfrak {R}_{4}= \langle 0.5399, 0.1361 \rangle \), respectively, where \(S(\mathfrak {R}_{1})=0.5021\),\(S(\mathfrak {R}_{2})=0.4250\), \(S(\mathfrak {R}_{3})=0\) and \(S(\mathfrak {R}_{4})=0.5774\).
-
Step 10:
Because \(S(\mathfrak {R}_{4})>S(\mathfrak {R}_{1})>S(\mathfrak {R}_{2})>S(\mathfrak {R}_{3})\), where \(S(\mathfrak {R}_{1})=0.5021\),\(S(\mathfrak {R}_{2})=0.4250\), \(S(\mathfrak {R}_{3})=0\) and \(S(\mathfrak {R}_{4})=0.5774\), the PO of the alternatives \(H_{1}, H_{2}, H_{3}\) and \(H_{4}\) is \(``H_{4}\succ H_{1}\succ H_{2}\succ H_{3}"\). Thus, \(H_{4}\) is the best alternative.
Table 3 presents a comparison of the POs of the alternatives \(H_{1},H_{2},H_{3}\) and \(H_{4}\) obtained by various MAGDM approaches for Example 5. From Table 3, it is clear that Seikh and Mandal’s MAGDM approach (Seikh and Mandal 2022) cannot handle this MAGDM problem because it get the indeterminant form in the intermediate steps while solving this MAGDM problem. However, Garg’s MAGDM approach (Garg 2020) obtain the PO \(``H_{1}= H_{4}\succ H_{2}\succ H_{3}"\) the alternatives \(H_{1}, H_{2}, H_{3}\) and \(H_{4}\), where it cannot distinguish the PO between the alternatives \(H_1\) and \(H_4\) in this particular case. While, Rahim et al.’s MAGDM approach (Rahim et al. 2023a) obtain the PO \(``H_{1}= H_{2}= H_{4}\succ H_{3}"\) for the alternatives \(H_{1}, H_{2}, H_{3}\), and \(H_{4}\), where it cannot distinguish the PO among the alternatives \(H_1\), \(H_2\) and \(H_4\) in this particular case. Moreover, Ahmad et al.’s MAGDM approach (Ahmad et al. 2024) and the proposed MAGDM approaches obtain the same PO \(``H_{4}\succ H_{1}\succ H_{2}\succ H_{3}"\) of the alternatives \(H_{1},H_{2},H_{3}\), and \(H_{4}\). Therefore, the proposed MAGDM approach can overcome the shortcomings of Seikh and Mandal’s MAGDM approach (Seikh and Mandal 2022), Garg’s MAGDM approach (Garg 2020) and Rahim et al.’s MAGDM approach (Rahim et al. 2023a) in this case.
6 Conclusion
In this paper, we have proposed new multiplication operation and scalar power operation for p, q-quasirung orthopair fuzzy numbers (p, q-QOFNs) based on Yager’s norm. Then, by using the proposed multiplication operation and scalar power operation of p, q-QOFNs and the concept of prioritized geometric aggregation operator (AO), we have proposed the p, q-quasirung orthopair fuzzy Yager prioritized weighted geometric (p, q-QOFYPWG) AO for aggregating p, q-QOFNs. We have also proved several properties of the proposed p, q-QOFYPWG AO of p, q-QOFNs. However, based on the proposed p, q-QOFYPWG AO, we have proposed a new MAGDM approach under the p, q-QOFNs environment. Afterwards, we have utilized the proposed MAGDM approach to solve different numerical MAGDM problems and compare the preference orders (POs) obtained from the proposed MAGDM method with POs obtained from Garg’s MAGDM approach (Garg 2020), Seikh and Mandal’s MAGDM approach (Seikh and Mandal 2022), Rahim et al.’s MAGDM approach (Rahim et al. 2023a) and Ahmad et al.’s MAGDM approach (Ahmad et al. 2024). From Example 3, Example 4 and Example 5, it is clear that the proposed MAGDM method can overcome the shortcomings of Garg’s MAGDM approach (Garg 2020), Seikh and Mandal’s MAGDM approach (Seikh and Mandal 2022), Rahim et al.’s MAGDM approach (Rahim et al. 2023a) and Ahmad et al.’s MAGDM approach (Ahmad et al. 2024), where they can not distinguish between the POs of available alternatives. The proposed MAGDM approach offers a useful approach to deal with MAGDM problems in the p, q-QOFNs environment.
Data availability
The numerical data used to support the findings of this study are available from the corresponding author upon request.
References
Ahmad T, Rahim M, Yang J, Alharbi R, Khalifa HAEW (2024) Development of p, q- quasirung orthopair fuzzy Hamacher aggregation operators and its application in decision-making problems. Heliyon. https://doi.org/10.1016/j.heliyon.2024.e24726
Akram M, Ilyas F, Garg H (2020) Multi-criteria group decision making based on ELECTRE I method in Pythagorean fuzzy information. Soft Comput 24:3425–3453
Alcantud JCR (2023) Multi-attribute group decision-making based on intuitionistic fuzzy aggregation operators defined by weighted geometric means. Granular Comput 8(6):1857–1866
Atanassov KT (1986) Intuitionistic fuzzy sets. Fuzzy Sets Syst 20(1):87–96
Çalı S, Balaman ŞY (2019) A novel outranking based multi criteria group decision making methodology integrating ELECTRE and VIKOR under intuitionistic fuzzy environment. Expert Syst Appl 119:36–50
Chen SM, Niou SJ (2011) Fuzzy multiple attributes group decision-making based on fuzzy preference relations. Expert Syst Appl 38(4):3865–3872
Chen SM, Lin TE, Lee LW (2014) Group decision making using incomplete fuzzy preference relations based on the additive consistency and the order consistency. Inform Sci 259:1–15
Chen SM, Cheng SH, Chiou CH (2016) Fuzzy multiattribute group decision making based on intuitionistic fuzzy sets and evidential reasoning methodology. Inform Fusion 27:215–227
Dutta P, Borah G (2022) Multicriteria group decision making via generalized trapezoidal intuitionistic fuzzy number-based novel similarity measure and its application to diverse COVID-19 scenarios. Artif Intell Rev 56(4):3543–3617. https://doi.org/10.1007/s10462-022-10251-z
Gao Y, Liu C, Zhao L, Zhang K (2021) Multi-attribute group decision-making method based on time-series q-rung orthopair fuzzy sets. J Intell Fuzzy Syst 41(1):2161–2170
Garg H (2020) A novel trigonometric operation-based q-rung orthopair fuzzy aggregation operator and its fundamental properties. Neural Computing and Applications 32(18):15,077–15,099
Garg H (2021) Sine trigonometric operational laws and its based Pythagorean fuzzy aggregation operators for group decision-making process. Artif Intell Rev 54(6):4421–4447
Garg H, Chen SM (2020) Multiattribute group decision making based on neutrality aggregation operators of q-rung orthopair fuzzy sets. Inform Sci 517:427–447
Hussain A, Wang H, Ullah K, Garg H, Pamucar D (2023) Maclaurin symmetric mean aggregation operators based on novel Frank T-norm and T-conorm for intuitionistic fuzzy multiple attribute group decision-making. Alexandria Eng J 71:535–550
Khan MSA, Abdullah S, Ali A, Amin F (2019) Pythagorean fuzzy prioritized aggregation operators and their application to multi-attribute group decision making. Granular Computing 4:249–263
Kumar K, Chen SM (2022) Group decision making based on advanced intuitionistic fuzzy weighted Heronian mean aggregation operator of intuitionistic fuzzy values. Inform Sci 601:306–322
Kumar K, Chen SM (2022) Group decision making based on q-rung orthopair fuzzy weighted averaging aggregation operator of q-rung orthopair fuzzy numbers. Inform Sci 598:1–18
Kumar K, Chen SM (2023) Group decision making based on entropy measure of Pythagorean fuzzy sets and Pythagorean fuzzy weighted arithmetic mean aggregation operator of Pythagorean fuzzy numbers. Inform Sci 624:361–377
Liu P, Chen SM (2017) Group decision making based on Heronian aggregation operators of intuitionistic fuzzy numbers. IEEE Trans Cybern 47(9):2514–2530
Liu P, Chen SM, Wang P (2018) Multiple-attribute group decision-making based on q-rung orthopair fuzzy power Maclaurin symmetric mean operators. IEEE Trans Syst Man Cybern Syst 50(10):3741–3756
Liu P, Khan Q, Jamil A, Haq IU, Hussain F, Ullah Z (2024) A novel MAGDM technique based on q-rung orthopair fuzzy Aczel-Alsina power Heronian mean for sustainable supplier selection in organ transplantation networks for healthcare devices. Int J Fuzzy Syst 26(1):121–153
Pinar A, Boran FE (2020) A q-rung orthopair fuzzy multi-criteria group decision making method for supplier selection based on a novel distance measure. Int J Mach Learn Cybern 11:1749–1780
Rahim M, Garg H, Khan S, Alqahtani H, Khalifa HAEW (2023) Group decision-making algorithm with sine trigonometric p, q-quasirung orthopair aggregation operators and their applications. Alexandria Eng J 78:530–542
Rahim M, Shah K, Abdeljawad T, Aphane M, Alburaikan A, Khalifa HAEW (2023) Confidence levels-based p, q-quasirung orthopair fuzzy operators and its applications to criteria group decision making problems. IEEE Access 11:109983–109996
Rahim M, Abosuliman SS, Alroobaea R, Shah K, Abdeljawad T (2024) Cosine similarity and distance measures for p, q,-quasirung orthopair fuzzy sets: Applications in investment decision-making. Heliyon. https://doi.org/10.1016/j.heliyon.2024.e32107
Rahim M, Akhtar Y, Yang MS, Ali HE, Elhag AA (2024) Improved COPRAS method with unknown weights under p, q-quasirung orthopair fuzzy environment: Application to green supplier selection. IEEE Access 12:69783–69795
Rahim M, Eldin EM, Khan S, Ghamry NA, Alanzi AM, Khalifa HAEW (2024) Multi-criteria group decision-making based on Dombi aggregation operators under p, q-quasirung orthopair fuzzy sets. J Intell Fuzzy Syst 46(1):53–74
Rahman K, Ali A (2020) New approach to multiple attribute group decision-making based on Pythagorean fuzzy Einstein hybrid geometric operator. Granular Computing 5(3):349–359
Salimian S, Mousavi SM (2022) The selection of healthcare waste treatment technologies by a multi-criteria group decision-making method with intuitionistic fuzzy sets. J Indus Syst Eng 14(1):205–220
Seikh MR, Mandal U (2022) Multiple attribute group decision making based on quasirung orthopair fuzzy sets: Application to electric vehicle charging station site selection problem. Eng Appl Artif Intell 115:105299
Wang J, Wei G, Wei C, Wei Y (2020) MABAC method for multiple attribute group decision making under q-rung orthopair fuzzy environment. Defence Technol 16(1):208–216
Xu Y, Wang H (2012) The induced generalized aggregation operators for intuitionistic fuzzy sets and their application in group decision making. Appl Soft Comput 12(3):1168–1179
Yager RR (1994) Aggregation operators and fuzzy systems modeling. Fuzzy Sets Syst 67(2):129–145
Yager RR (2008) Prioritized aggregation operators. Int J Approx Reason 48(1):263–274
Yager RR (2013) Pythagorean membership grades in multicriteria decision making. IEEE Trans Fuzzy Syst 22(4):958–965
Yager RR (2016) Generalized orthopair fuzzy sets. IEEE Trans Fuzzy Syst 25(5):1222–1230
Zadeh LA (1965) Fuzzy sets. Inform Control 8(3):338–353
Zhang Z, Chen SM (2022) Group decision making based on multiplicative consistency and consensus of Pythagorean fuzzy preference relations. Inform Sci 601:340–356
Zhang X, Liu P, Wang Y (2015) Multiple attribute group decision making methods based on intuitionistic fuzzy Frank power aggregation operators. J Intell Fuzzy Syst 29(5):2235–2246
Zhang Z, Chen SM, Wang C (2020) Group decision making with incomplete intuitionistic multiplicative preference relations. Inform Sci 516:560–571
Zhong Y, Gao H, Guo X, Qin Y, Huang M, Luo X (2019) Dombi power partitioned Heronian mean operators of q-rung orthopair fuzzy numbers for multiple attribute group decision making. PLoS One 14(10):e0222007. https://doi.org/10.1371/journal.pone.0222007
Funding
Authors have no funding.
Author information
Authors and Affiliations
Contributions
Each author has equal contribution.
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no Conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Redhu, A., Kumar, K. Multi-attribute group decision making based on p, q-quasirung orthopair fuzzy Yager prioritized weighted geometric aggregation operator of p, q-quasirung orthopair fuzzy numbers. Granul. Comput. 9, 75 (2024). https://doi.org/10.1007/s41066-024-00495-1
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s41066-024-00495-1