Probability Distribution-Based Processing Model of Probabilistic Linguistic Term Set and Its Application in Automatic Environment Evaluation

• Published in: International journal of fuzzy systems Vol. 23; no. 6; pp. 1697 - 1713
• Main Authors: Li, Yu, Wang, Xiao-kang, Wang, Jian-qiang, Li, Jun-bo, Li, Lin
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.09.2021; Springer Nature B.V
• Subjects: Artificial Intelligence; Attitudes; Computational Intelligence; Data integration; Decision making; Engineering; Environmental monitoring; Linguistics; Management Science; Methods; Multiple criterion; Operations Research; Probability; Probability distribution; Project evaluation; Sensitivity analysis; Statistical analysis; Multi-criteria decision-making (MCDM); Pre-processing technique; Probabilistic linguistic term set (PLTS); Probabilistic linguistic earth-mover (PLEM) distance; Probabilistic linguistic probability distribution (PLPD)
• ISSN: 1562-2479; 2199-3211
• DOI: 10.1007/s40815-021-01060-7

The probabilistic linguistic term set (PLTS) is an effective tool for describing linguistic value in the process of decision-making. However, due to its complex structure, there remains numerous defects in utilizing the PLTS information. Therefore, by connecting the PLTS with discrete probability distribution, this study investigates a novel probability distribution-based processing model to process PLTS and further develops a practical multi-criteria decision-making (MCDM) approach for addressing automatic environmental monitoring evaluation problem. First, we propose two types of pre-processing methods to complete probability normalization. Accordingly, the normalized PLTS is represented in the form of probabilistic linguistic probability distribution (PLPD). Then, probabilistic linguistic earth-mover (PLEM) distance is constructed to measure the divergency between PLPDs, and a probabilistic linguistic probability distribution weighted mean (PLPDWM) is also developed to facilitate information fusion. On this basis, an aggregation operator-based MCDM approach is established that integrates the probability distribution-based PLTS processing methods and the proposed worst-priority weight (WPW) method. Finally, the developed approach is demonstrated by solving a practical atmospheric pollutant evaluation problem and its strengths are verified through further sensitivity analysis and comparative analysis.