区间信息下的主客方协作式群体评价方法及其应用

目前关于主客方协作式评价问题的研究相对较少且主要是基于点值评价信息的,考虑到评价环境的复杂性与不确定性,本文将点值信息向区间信息方向拓展,探讨了一种新的区间信息下的主客方协作式群体评价方法。本文首先探讨了一种能够较好融合评价信息"质与量"的区间诱导密度加权合成算子-IIDWA;然后以主方信息完备度及客方信息诚信度为诱导分量分别对主客方评价信息进行聚类分组,并从规模和属性两个角度出发分别确定相应的密度加权向量;最后在主客方协作规则下,利用IIDWA算子对主客方区间信息分别进行二维集结,以得出最终综合评价结果。文章最后给出了一个应用算例,算例表明了方法的可行性与有效性。

A Group Evaluation Method and Application Based on Collaboration of Subject and Object under Interval Information

In classical single criterion group evaluation methods, the ratings that experts (the subject) have given for the object evaluated are usually exact data. In addition, the evaluation process is also generally dominated by the subject with lack of the object participation. However, the ratings given are human judgements including preferences that may be vague; using interval data should be more suitable. Otherwise, with the circumstance of emphasizing democracy and freedom, participation of the object (especially when the object is human) in evaluation process is rather essential. Therefore, a group evaluation method based on collaboration of subject and object is put forward under interval information. In this paper, an Interval Induced Density Weighted Algorithm-IIDWA is presented to aggregate interval data. Firstly, completeness of subject information-μand integrity of object information-ωare calculated, by which original interval data are clustered into right group. Secondly, ultimate weight vector of each group are synthesized by their attribute weight vector-ξ^? and scale weight vector-ξ^e, so it will possess the superiority of containing the characteristics of attribute and scale of each group. Finally, information of subject and object are aggregated respectively by IIDWAto obtain interval comprehensive results and then possibility degree approach of interval data is conducted for ranking. In the end, a numerical example is given to illustrate the feasibility and validity of this paper. Meanwhile, the result based on the TOPSIS method with interval data is also calculated in order to compare the existing difference with ranking of this paper. As the result of the two methods show, the ranking is different, which indicates that participation of the object can make contribution to the ranking result. In this paper, the complementation of the subject and object information has been implementated.