专家交互情境下不完备群组DEMATEL决策方法

为克服现有群组DEMATEL存在的尚未考虑群组专家之间的信息交互、对于不完备专家判断信息的推断机理不明确等缺陷,基于社交网络中的信任关系理论和凝聚层次聚类理论,通过给出不完备群组DEMATEL初始直接影响矩阵残缺值的推断方法和专家交互情境下群组DEMATEL直接影响矩阵信息修正方法,提出专家交互情境下不完备群组DEMATEL决策方法的实现步骤.最后,通过算例对比分析反映出通过专家多轮次交互,群组专家之间的共识度以及决策结果的可靠性持续提升,从而验证了所提出方法的科学性和可行性.

Incomplete group DEMATEL decision-making method under expert interaction context

Existing group decision-making trial and evaluation laboratory(DEMATEL) has drawbacks, such as information interaction between group experts is ignored and the inference mechanism for incomplete expert judgment information is unclear. To overcome the abovementioned drawbacks, based on the trust relationship theory in social networks and the agglomerative hierarchical clustering theory, the implementation steps for the incomplete group DEMATEL under expert interaction context are ultimately proposed by giving the inference method of missing values in the incomplete initial direct influence matrix (DIM) and the information correction method for DIM under the expert interaction context of the group DEMATEL. Finally, through the comparative analysis for a numerical example, it is reflected that the consensus degree between group experts and reliabilities of decision results are continuously improved through the experts'' multiple rounds of interaction, thus the proposed method is validated to be feasible and scientific.