基于最佳聚类准则的多级模糊态势评估方法

摘要

In order to solve the problems of the uncertainty and fuzziness and of the fuzzy partition and membership functions to be given in advance in situation assessment and the influence of ignored data distribution characteristics to the evaluation results , this paper proposed a multi-level fuzzy comprehensive evaluation method of situation assessment based on the best clustering criteria. First, it completed automatically the fuzzy partition and determined the membership functions after getting the number and the centers of the best clustering according to the criteria. Second, it established the first level evaluation models of the weighted average and the main factors for a single factor. After, it obtained the two level models of the above two models based on the all factors' weights under AHP. Finally, it was applied to the process of the civil aviation disaster situation assessment and got the better results. It shows that the results of the weighted average model are more accurate by the evaluating the two models.%针对态势评估中评估因素的不确定性、模糊性和模糊集划分、隶属函数需事先给定以及忽略了数据分布特点对评估结果影响的问题,提出了基于最佳聚类准则的多级模糊综合评判态势评估方法.根据最佳聚类准则得到最佳聚类数和聚类中心后完成数据属性的模糊集划分及隶属函数的确定,建立了基于单因素的主因素和加权平均的一级模糊评估模型,利用层次分析法得到所有因素对评估结果的影响权值,并建立所有因素的主因素和加权平均的二级模糊评估模型,将其应用到民航灾难态势评估过程,得到了较好的态势评估结果.通过对两种模型的评价,得出加权平均模型的态势评估结果更准确.