基于证据距离和特征向量的仿真数据可信度评估方法  

Credibility Evaluation Method of Simulation Data Based on Evidence Distance and Eigenvector

作　　者：李文博 黄炎焱[1] 张寒 白逸凡 Li Wenbo;Huang Yanyan;Zhang Han;Bai Yifan(School of Automation,Nanjing University of Science and Technology,Nanjing 210094,China)

机构地区：[1]南京理工大学自动化学院,南京210094

出　　处：《兵工自动化》2024年第12期48-54,共7页Ordnance Industry Automation

基　　金：装备预先研究共用技术项目(50901010101)。

摘　　要：针对作战仿真模型数据的特点,从仿真模型输出的源数据入手,提出一种基于Jousselme证据距离的源数据可信度评估方法。根据仿真模型构建作战评估指标体系,确定输出的仿真数据类型,使用专家评估法将仿真数据进行量化处理作为证据源数据,根据Jousselme证据距离和特征向量法计算证据源数据的客观权重,利用特征向量法评估专家权重作为主观权重,根据2种权重得到综合权重,使用综合权重修正原证据源数据,经多次证据融合得到该仿真数据的可信度向量。通过仿真数据可信度评估实例,结果表明,该方法的可行性和准确度较高。According to the characteristics of combat simulation model data,this paper proposes a credibility evaluation method of source data based on Jousselme evidence distance.According to the simulation model,a combat evaluation index system is constructed,the type of output simulation data is determined,the simulation data is quantified as evidence source data by using an expert evaluation method,the objective weight of the evidence source data is calculated according to the Jousselme evidence distance and the eigenvector method,and the expert weight is evaluated by using the eigenvector method as a subjective weight.According to the two kinds of weights,the comprehensive weight is obtained,and the original evidence source data is modified by using the comprehensive weight,and the credibility vector of the simulation data is obtained through multiple evidence fusion.Through the example of simulation data credibility evaluation,the results show that the method is feasible and accurate.

关 键 词：Jousselme证据距离 特征向量法 仿真数据可信度评估 专家评估 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]