基于双重驱动的实数型群体评价随机聚合算法

Random Aggregation Algorithm of Real Number Group Evaluation Based on Double Drive

针对实数型群体评价信息有效集结的问题,提出了一种凝聚群体共识的随机聚合模型。考虑实数型群体评价信息直接集结运算并不能有效反映决策环境的复杂多变及不确定性因素的影响,为此将离散的实数型群体评价信息转化为连续型随机变量,以凸显现实中的决策环境、更加细腻地刻画群体价值表达;在此基础上,提出了双重驱动的指标赋权方法,即根据各指标随机信息之间的差异性(差异驱动)和群体共识的一致性程度(共识驱动)进行组合赋权;最后,聚合求解被评价对象之间带有概率特征的可能性排序。该方法可以较全面地集结群体评价信息,从可能性的视角给出合理的排序结论,算例分析验证了方法的有效性和合理性。

This paper addresses the issue of effectively aggregating real-valued group evaluation information.In order to achieve this,a stochastic aggregation model for consolidating group consensus is proposed.In light of the limitations of direct aggregation of discrete real-valued group evaluation information in effectively reflecting the complexity and variability of decision-making environments,as well as the influence of uncertain factors,this study employs a transformation of such discrete information into continuous random variables.This transformation is intended to elucidate the nuances of real-world decision-making environments and portray group value expressions with greater precision.Subsequently,a dual-driven approach for indicator weighting is introduced,which combines weights based on the differences among random information pertaining to various indicators(difference-driven)and the degree of consensus among the group(consensus-driven).Ultimately,a probabilistic ranking with possibility characteristics among the evaluated objects is derived through aggregation.This methodology enables a comprehensive aggregation of group evaluation information and offers reasonable ranking conclusions from a probabilistic perspective.The effectiveness and rationality of the proposed approach are demonstrated through the use of an illustrative example.