TL;DR
This paper introduces a novel decision-making framework for selecting sustainable Unmanned Aerial Vehicle (UAV) suppliers, crucial for applications like power line inspection. It tackles the inherent uncertainty in expert evaluations by employing a "two-dimension linguistic intuitionistic fuzzy variable" (2DLIFV) approach, integrating it with the Best-Worst Method (BWM) for attribute weighting and the MULTIMOORA method for supplier ranking.
What Problem Does This Paper Address?
The increasing adoption of UAVs for critical tasks like power line inspection necessitates reliable and sustainable supplier selection. However, this process is fraught with challenges:
- Uncertainty and Fuzziness: Expert evaluations are often subjective, vague, and expressed in linguistic terms, making them difficult to quantify.
- Attribute Importance: Different attributes (e.g., price, quality, reliability, environmental impact) carry varying levels of importance, requiring a robust method to determine their weights.
- Expert Subjectivity: Multiple experts contribute to the decision, and their individual experiences and potential biases need to be accounted for.
- Data Representation: Existing methods struggle to fully capture the nuanced and uncertain nature of expert judgments.
This paper aims to address these issues by proposing a more flexible and comprehensive decision-making framework.
Methods
The core of this paper lies in extending existing decision-making methodologies to handle complex, uncertain, and linguistic data. The key components are:
- Two-Dimension Linguistic Intuitionistic Fuzzy Variables (2DLIFVs): This novel representation scheme is used to capture expert evaluations. It goes beyond standard fuzzy sets by incorporating both the evaluation itself (e.g., "good") and the reliability of that evaluation (e.g., "very reliable"). This allows for a richer representation of uncertainty.
- 2DLIF-BWM (Best-Worst Method): The authors adapt the Best-Worst Method (BWM) to work with 2DLIFVs. BWM is known for its efficiency in determining attribute weights by focusing on pairwise comparisons between the best and worst attributes and others. The 2DLIF extension allows experts to express these preferences using the 2DLIFV framework.
- Novel Expert Weight Calculation: A unique method is introduced to determine the weights of individual experts. This method considers both the "uncertainty degree" (how hesitant an expert is) and the "consensus degree" (how much an expert agrees with others), aiming for a more balanced and reliable weighting of expert opinions.
- 2DLIF-MULTIMOORA: The MULTIMOORA method, a comprehensive approach combining ratio system, reference point, and full multiplicative forms, is extended to work with 2DLIFVs. This allows for the aggregation of evaluations across multiple attributes and the final ranking of suppliers.
Datasets
Not discussed in the paper. The study uses a simulated case study with expert evaluations.
Key Insights and Ideas
- Richer Uncertainty Representation: 2DLIFVs offer a significant advantage over traditional fuzzy sets by explicitly modeling the reliability of an evaluation alongside the evaluation itself. This is crucial for real-world decision-making where expert confidence can vary.
- Integrated Decision-Making: Combining BWM for attribute weighting and MULTIMOORA for ranking within the 2DLIFV framework creates a holistic approach that addresses multiple facets of the decision problem.
- Expert Weighting Matters: The proposed method's novel approach to expert weighting, considering both uncertainty and consensus, is a significant contribution. It acknowledges that not all expert opinions are equally valuable or reliable, and provides a data-driven way to account for this.
- Robustness in Supplier Selection: The sensitivity analysis demonstrates that the proposed method's ranking results are stable even when attribute weights fluctuate, suggesting its reliability in practical applications.
Key Results
The paper presents a case study for selecting a sustainable UAV supplier for power line inspection. The results indicate:
- Optimal Supplier Identified: Alternative A4 was consistently ranked as the optimal supplier across all tested methods, including the proposed approach.
- Ranking Stability: The proposed method showed robust performance, with ranking orders remaining stable despite variations in attribute weights.
- Comparative Advantage: The proposed method's ranking of alternatives A2 and A3 differed slightly from some other methods (TOPSIS, PROMETHEE), which the authors attribute to the richer information captured by 2DLIFVs compared to simpler linguistic intuitionistic fuzzy information.
Limitations and Future Work
The authors acknowledge the following limitations and suggest future research directions:
- Static Assumption: The current model assumes a static decision-making environment. Future work could explore dynamic characteristics of supplier selection.
- Expert Subjectivity: The inherent subjectivity of expert evaluations remains a limitation.
- Group Consensus Enhancement: Future research could incorporate group consensus methods to further improve the reliability of evaluation information.
- Large-Scale Decision-Making: Extending the method to handle large-scale group decision-making problems would enhance its practical applicability.
Summary updated