基于D-S证据理论的测试性综合评估方法  

Comprehensive testability evaluation method based on D-S evidence theory

作　　者：余思奇[1] 景博[1] 黄以锋[1] 

机构地区：[1]空军工程大学航空航天工程学院,西安710038

出　　处：《计算机应用研究》2014年第7期2071-2073,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目(51201182);航空科学基金资助项目(20101996012)

摘　　要：基于Bayes理论的测试性评估方法能充分利用产品研制阶段先验信息,评估结论置信度相比经典方法有了很大提高。针对先验信息不确定性影响Bayes评估结论的问题,提出基于D-S证据理论的测试性综合评估方法。首先建立先验信息标准化模型,将先验信息转换成置信区间形式;然后根据D-S证据理论和Pignistic转换推导先验信息矩值,并解算先验分布超参数;最后基于Bayes定理给出后验分布和评估公式。实例证明该方法相比Bayes评估的置信度和评估精度更高。The testability evaluation method based on Bayes theory can make full use of the priori information in development phase and greatly improve assessment confidence compared to the classical method. But uncertainties in the priori information affect assessment conclusions. For this issue,this paper proposed the comprehensive testability evaluation method based on D-S evidence. Firstly,the paper established a priori information standardization model that priori information was transformed into the form of confidence interval. Then it derived the priori information moment based on D-S evidence theory and Pignistic transformation,and solved priori distribution over parameters. Finally,it gave the posteriori distribution and evaluation formula based on Bayes theory. Examples show that this method has a greater confidence and higher assessment accuracy compared to the Bayes method.

关 键 词：D-S证据理论 测试性评估 Bayes理论 Pignistic转换 

分 类 号：TP806.1[自动化与计算机技术—检测技术与自动化装置]