群体决策可靠性的分层反馈与多级度量

Hierarchical feedback and multi-level measurement of groupdecision reliability

为了提高群体决策质量,获得更加可靠的决策结果,结合语言环境下考虑专家自信度的异构多属性群决策问题,建立了群体决策可靠性的分层反馈机制与多级度量模型。利用三支决策思想将各方案各属性下对应子群体中的专家按照自信度的高、中、低划分成三类,进而,依次从子群体的专家构成、自信度以及共识度3个层次对决策可靠性进行量化分析,建立由粗到细、逐层细化的反馈机制,为决策过程的可靠性调整和控制提供途径;综合考虑子群体中高、中自信度专家的占比以及自信度和共识度等影响因素,构建多级可靠性度量模型,并形成针对方案级决策和整体决策的两级可靠性判定规则,有效提升决策结果的可解释性。在此基础上,提出一种可靠性驱动的异构多属性群决策方法,通过供应商选择实例分析并验证了所提方法的可行性和有效性。

In order to improve the quality of group decision making and obtain more reliable decision results,this paper establishes a hierarchical feedback mechanism and multi-level measurement model for the reliability problem of heterogeneous multi-attribute group decision making considering self-confidence in linguistic environment.First,experts in the subgroups under each attribute of each alternative are divided into three categories according to the high,medium and low self-confidence by combining the three-way decision-making ideas.Then,the decision reliability is quantitatively analyzed from the three levels of expert composition,self-confidence and consensus of subgroups successively.A feedback mechanism is established from coarse to fine and refined layer by layer,providing a way for reliability adjustment and control of decision-making process.Considering the proportion of high and medium confidence experts in the sub-groups as well as factors such as confidence level and consensus level,a multi-level reliability measurement model is constructed,and two-level reliability judgment rules for scheme-level decision-making and overall decision-making are formed to effectively improve the interpretability of decision-making results.On this basis,a reliability driven heterogeneous multi-attribute group decision-making method is proposed.Finally,the feasibility and effectiveness of the proposed method are analyzed and verified by the case of supplier selection.