TL;DR
This paper introduces new aggregation operators for single-valued neutrosophic numbers (SVNNs) that can handle uncertainty and interdependencies in decision-making. These operators are then applied to develop a novel method for multiple-attribute group decision-making (MAGDM), demonstrating improved robustness in handling potentially unreasonable data.
What Problem Does This Paper Address?
Real-world decision-making often involves complex scenarios with uncertain, incomplete, or even inconsistent information. Traditional fuzzy set (FS) and intuitionistic fuzzy set (IFS) theories, while useful, struggle with indeterminate and inconsistent data. Neutrosophic sets (NSs), and specifically single-valued neutrosophic sets (SVNSs), offer a more powerful framework to represent such information by including truth, indeterminacy, and falsity membership degrees. However, existing aggregation operators for SVNNs may not fully capture the nuances of decision-making, such as the interrelationship between attributes or the impact of outlier data. This paper addresses this gap by proposing new aggregation operators that combine the strengths of Power Average (PA) and Heronian Mean (HM) operators, specifically designed to handle SVNNs and improve MAGDM processes.
Methods
The core of this paper lies in the development of novel aggregation operators for SVNNs. The authors introduce:
- Single-Valued Neutrosophic Power Heronian Aggregation (SVNPHA) operators: These operators extend the concept of Power Heronian Aggregation (PHA) to SVNNs. They leverage the PA operator's ability to account for the support degree between input arguments (mitigating the impact of unreasonable data) and the HM operator's capacity to capture interrelationships among arguments.
- Single-Valued Neutrosophic Weighted Power Heronian Aggregation (SVNWPHA) operators: These are weighted versions of the SVNPHA operators, allowing for differential importance of individual SVNNs.
- Single-Valued Neutrosophic Geometric Power Heronian Aggregation (SVNGPHA) operators: These operators incorporate a geometric mean aspect into the PHA framework for SVNNs.
- Single-Valued Neutrosophic Weighted Geometric Power Heronian Aggregation (SVNWGPHA) operators: These are weighted versions of the SVNGPHA operators.
The paper also defines operational rules for SVNNs, comparison methods, and distance measures, which are foundational for developing and proving the properties of these new operators. A MAGDM method is then proposed using the SVNWPHA and SVNWGPHA operators to aggregate decision information from multiple decision-makers.
Datasets
The paper uses synthetic data for its illustrative example (Example 2) to assess air quality. This data is presented as three normalized standardized decision matrices from three different experts (decision-makers), representing air quality data from Guangzhou for 2006-2009. No new benchmark datasets are introduced.
Key Insights and Ideas
- Bridging Uncertainty and Interdependencies: The proposed SVNPHA and SVNWGPHA operators effectively combine the ability to handle uncertain information (via SVNNs) with the capacity to model both the support between individual data points and the interrelationships among attributes. This is a significant advancement over operators that focus on only one aspect.
- Robustness to Unreasonable Data: By incorporating the "support degree" concept from the PA operator, these new operators can mitigate the undue influence of extreme or outlier values in the decision-making process, leading to more reliable outcomes.
- Flexibility through Parameters: The inclusion of parameters
pandqin the proposed operators allows for flexibility in how the aggregation is performed, enabling decision-makers to tailor the method to specific problem contexts. - Enhanced MAGDM: The developed MAGDM method, utilizing the SVNWPHA and SVNWGPHA operators, provides a structured approach to fuse information from multiple experts with varying weights and attribute importance, leading to a more comprehensive and robust final decision.
Key Results
The paper successfully defines and proves several properties for the proposed SVNPHA, SVNWPHA, SVNGPHA, and SVNWGPHA operators, including idempotency, commutativity, and boundedness. An illustrative example (Example 2) demonstrates the application of the proposed MAGDM method using SVNWPHA. The results show a consistent ranking of alternatives, with the best alternative being E4, regardless of variations in the p and q parameters within a reasonable range. A sensitivity analysis (Tables 4 and 5) indicates that while different parameter values can lead to minor shifts in ranking, the overall best alternative remains stable. A comparison with existing methods (Table 6) shows that the proposed method achieves the same ranking results, demonstrating its effectiveness. Table 7 highlights that the proposed operators possess desirable properties (considering interrelationship and support degree) that are not all present in other compared operators.
Limitations and Future Work
The authors acknowledge that the proposed operator calculation is slightly more complicated due to the simultaneous consideration of PA and HM operators. They suggest that future research could extend these operators to other domains such as medical diagnosis and pattern recognition. The sensitivity analysis also notes that extreme values of p and q can lead to more complex calculations and potentially less stable ranking results, although the method remains relatively stable for moderate parameter values.
Summary updated