Research On Covering-based Variable Precision Fuzzy Rough Sets Based On Fuzzy Logical Operators And Corresponding Multi-attribute Decision-making

As a combination of rough set theory and fuzzy set theory,fuzzy rough sets have been obtained extensive attentions from plenty of scholars.Based on the merits of fuzzy logical operators in information fusion and the basic paradigm of variable precision rough set models,this thesis mainly studies covering-based variable precision fuzzy rough set models and their applications in multi-attribute decision-making(MADM).More precisely,this thesis includes:(1)Based on four fuzzy neighborhood operators proposed by D'eer et al.on fuzzy coverings,two types of covering-based variable precision fuzzy rough set models are defined by means of fuzzy logical operators and a parameter ?,which are called covering-based light and loose variable precision fuzzy rough sets,respectively.Meanwhile,we investigate some significant properties and relationships of the two proposed rough set models.In addition,we combine the proposed rough set models with VIKOR method and present an MADM method to medicine selection problems in the background of medical diagnosis.Through case study,comparative analysis and experiment evaluation,we verify the reasonability and validity of the proposed method.(2)A reflexive fuzzy ?-neighborhood operator on fuzzy ?-coverings by virtue of fuzzy R-implicators is defined.Meanwhile,its properties and the relationships with existing fuzzy neighborhoods are discussed.Then,we define covering-based generalized variable precision fuzzy rough sets based on fuzzy logical operators and a parameter ?.Afterwards,by combining the proposed rough set model with the PROMETHEE II and EDAS methods,we establish a novel MADM method.In the end,the rationality and effectiveness of the established decision-making method are discussed via case study,comparative analysis and experiment evaluation.