关键词: 主客方协作式群体评价, 异构偏好, 不确定语言, 区间灰数, 调节因子

Abstract: Due to the increasingly complex decision-making environment, the resolution to some large-scale decision-making problems often requires group evaluation methods to improve the scientificity and decision-making quality. However, in previous relevant studies, most of them have not considered the participation of object side, and the expansion of applicable pattern is relatively insufficient. At the same time, most of the existing group evaluation methods based on subject-object collaboration have not considered the situation where both sides are willing to express their opinions by using heterogeneous preference information. Also, the interaction between the subject and object sides is relatively inadequate and the comprehensive correction effect of information is not obvious. Aiming at these problems, in this paper, a new collaborative group evaluation method is proposed based on heterogeneous preference information. After using uncertain language and interval grey number to express the evaluation opinion of each side respectively, the interaction and fusion of their opinions are realized through two collaborative processes.
The goal of the “First Collaboration” between the subject and object is to correct the evaluation information. Firstly, the collaborative initial score of each evaluation value is given, and the principles of both sides are consistent. In detail, the calculations are realized based on the deviation degree of uncertain language and the relatively accurate integrated value of interval grey number respectively. Secondly, the “Completeness Degree” and “Trustworthy Degree” of corresponding information are calculated based on collaborative initial scores accordingly, and the adjustment factors are used to correct the corresponding evaluation information of both sides,so as to complete the first collaboration. Subsequently, the “Second Collaboration” between two sides is launched with the goal of evaluation information gathering. Firstly, based on the TOPSIS method, the corrected evaluation information of the subject and object sides is preliminarily fused respectively to obtain the comprehensive proximity degree, forming a comprehensive proximity matrix. Secondly, an ordered group clustering based on row division is proposed, and the grey relational clustering method is applied into the ordered group clustering based on the comprehensive proximity matrix of each side. Finally, based on the TDWA operator, the two-dimensional information of the comprehensive proximity is aggregated to obtain the comprehensive evaluation result of each side respectively. Then, the revised comprehensive opinions of two sides are aggregated to obtain the comprehensive evaluation results based on subject-object collaboration, and the second collaboration is completed.
Through a practical example of medical equipment supplier partners’ evaluation and selection, the proposed method is effectively applied. And the partner with the highest score is the best medical equipment supplier that decision maker should select. On this basis, in order to illustrate the rationality and effectiveness of this method, the collaboration process between the subject and object is removed and the method internal comparisons are conducted. The results show that the two collaborations effectively realize the interaction and fusion of heterogeneous preference opinions compared to the independent evaluation results of the subject and object without collaboration. Moreover, in order to illustrate its advantages, the group evaluation method based on subject-object collaboration in previous studies are further introduced to conduct the method external comparisons. The results show that the two collaborative processes could effectively enhance the accuracy and depth of the evaluation information interaction between two sides, the rationality and stability of the evaluation results are also improved. Nevertheless, it should be pointed out that this method is mainly applicable to situations where the evaluation object has the ability ofactive conscious behavior. For this issue, the further in-depth research will be conducted in future to achieve its expansion in different application scenarios.

Key words: group evaluation based on subject-object collaboration, heterogeneous preferences, uncertain linguistic information, interval gray number, adjustment factor