多粒度不确定语言变量的多属性群决策方法及应用

Multi-attribute Group-Decision-Making with Multi-granularity Uncertain Linguistic Variables and Its Application

针对多粒度语言的多属性群决策问题,提出基于不确定语言变量的一致化新方法。利用虚拟术语指标不丢失信息的特点,选取连续性语言评价集,提出符合所给条件的多粒度不确定语言变量的转换函数,将多粒度语言一致化,并探讨了其优势之处,验证了其合理性;给出了不确定语言变量的有序加权平均(ULV-OWA)算子的定义,该算子具有计算简便且能充分考虑偏好信息分布的特点,通过加权算术平均(ULWA)算子和ULV-OWA算子对偏好信息进行集结,再提出优势可能度法对方案排序。所提出的方法不但避免了决策者偏好信息的丢失,而且减少了决策者的工作量,简便直观。最后,通过一个现代电子商务背景下物流供应商选择的案例研究,验证了方法的有效性和合理性。

In order to solve multi-attribute group-decision-making of multi-granularity language, a new method of unifying the multi-granularity linguistic term sets based on different uncertain linguistic variables has been proposed in this paper. Firstly, continuous language evaluation sets are selected in order to preserve the information by applying the virtual term indicators. Based on that, a conversion function of multi-granularity uncertain linguistic variables that comply with the given conditions is formulated to unify the multi-granularity linguistic term sets. We then discuss its advantages as well as its legitimacy. After that, an ordered weighted averaging operator （ULV- OWA ） of uncertain linguistic variables, which has desired characteristics of simple calculation and fully carrying on the distribution of preference information, is defined. By doing so, the preference information can be gathered by weighted arithmetic average （ULWA） and ULV-OWA operator. We then propose so-called advantage possibility degree method to sort programs. Our method can greatly simplify the decision-making process without loss of the preference information, and therefore reduce the workload of decision makers. Finally, the validity and rationality of our method are verified through a case study of logistics supplier selection problem in a modern electronic commerce environment.