Additive Consistency-Based Decision-Making with Incomplete Probabilistic Linguistic Preference Relations
Zi-yu Chen,Fei Xiao,Min-hui Deng,He-wei Liu,Jian-qiang Wang
DOI: https://doi.org/10.1007/s40815-021-01144-4
IF: 4.085
2021-07-25
International Journal of Fuzzy Systems
Abstract:Probabilistic linguistic term set (PLTS), an efficient tool to describe decision information, can sufficiently express decision makers' hesitation and preference. Probabilistic linguistic preference relation (PLPR) is based on PLTSs to describe the preference information of experts for paired alternatives. However, in practice, due to the complexity of the problem, the incompleteness of information and the lack of professional knowledge, the incomplete PLPR (InPLPR) with missing information often appears. Therefore, this paper proposes a decision-making method under InPLPR. Firstly, in order to fully consider the specific situation of missing values, missing linguistic term-InPLTS (MLT-InPLTS) is subdivided into missing single linguistic term-InPLTS (MSLT-InPLTS) and missing multiple linguistic terms-InPLTS (MMLT-InPLTS). Then, a two-stage mathematical optimization model of missing information estimation based on additive consistency, fuzzy entropy and hesitation entropy is established. Subsequently, aiming at the unacceptable consistency of complete PLPR (CPLPR) after filling in the missing values, a consistency improvement method based on the idea of gradient descent is proposed. Afterward, probabilistic linguistic weighted averaging (PLWA) operator is used to rank alternatives. Finally, medical supplier selection is taken as an example to verify the effectiveness of the proposed decision-making method, and the robustness and advantages of this method are illustrated by sensitivity analysis and comparison with other methods.
computer science, information systems,automation & control systems, artificial intelligence