摘要
为了降低测试性验证试验的不确定性风险损失,提出了基于Bayesian理论的测试性验证试验方案。该方案综合考虑了故障注入成本以及弃真和采伪两类风险损失,建立了基于风险损失的试验方案模型。根据Bayesian理论,结合验前信息推导出试验最小样本和最大合格判定数的计算公式,并给出基于Bayesian平均风险准则的弃真和采伪两类风险参数的求解方法。在此基础上,按照试验期望风险损失最小的原则求解试验方案模型。结果表明,该方案能有效降低试验两类风险损失。
Aiming to reduce risk loss in testability demonstration tests,a scheme based on Bayesian theory is proposed. The cost of fault injection,risk loss of producer and consumer are considered syn-thetically,and a test model based on risk loss is established. According to Bayesian theory and prior information,the formulas of minimum test sample number and maximum qualification evaluation number are deduced,and a solving method of producer’s risk and consumer’s risk is provided based on Bayesian rule of average risk. The solving method of test model is presented under the principle of minimizing expect risk loss. Experimental results show that the scheme proposed here can reduce two types of risk loss effectively.
出处
《南京理工大学学报》
EI
CAS
CSCD
北大核心
2014年第6期775-780,共6页
Journal of Nanjing University of Science and Technology
关键词
不确定性风险损失
测试性验证试验
故障注入成本
弃真风险
采伪风险
Bayesian理论
验前信息
试验最小样本
最大合格判定数
期望风险损失
uncertainty risk loss
testability demonstration tests
fault injection cost
producer’s risk
consumer’ s risk
Bayesian theory
prior information
minimum test sample number
maximum&amp
nbsp
qualification evaluation number
expect risk loss