基于乘积偏好关系的专家模糊核聚类赋权方法

Method of Expert Fuzzy Kernel Clustering Weighting Based on Multiplicative Preference Relations

多属性、多目标性决策中,针对专家给出各方案偏好关系下的决策问题,提出一种基于乘积偏好关系的专家模糊核聚类赋权方法。该方法运用模糊核聚类的思想实现对决策专家的聚类,并通过放宽归一化约束条件,克服了传统模糊核聚类算法中离群点对聚类结果的影响。同时,在专家类内赋权过程中,运用CI-IOWG算子集结同类专家的意见,依据不同专家对于形成类别一致性意见的贡献程度来确定专家权重;克服了传统基于熵权或判断矩阵一致性的赋权方法的局限性。算例表明,该方法可行、有效。

Within the multiple attributes and multi-objective decision making problems, for the case that each decision maker has a preference relation referring to alternatives, a method of expert fuzzy kernel clustering weighting based on multiplicative preference relations is proposed, in which the experts are classified by using fuzzy kernel clustering principle. By loosening the normalization constraints, the effects of outliers on the clustering results could be overcome. At the same time, this paper presents CI-IOWG operator for group decision-making with muhiplicative preference relation in the process of determining the intra class weight. And the weighting method can determine the experts＇ weight according to the contribution degree for clustering which overcomes the limitations of the traditional weighting method based on entropy and consistency. The example shows that the method is feasible and effective.