论文部分内容阅读
针对现有测试性验证试验方案需要较大故障样本量的问题,提出了一种利用研制信息制定测试性验证试验方案的新方法。首先,在二项分布经典抽样模型的基础上,运用证据理论方法,研究了基于测试性试验数据、测试性预计结果和专家经验等研制信息的基本信任分配函数的构造方法,建立了基于融合不同种类研制信息的测试性验证试验方案,并确定了该方案的存在性,给出了方案的确定方法,最后开展了案例应用研究。由于充分考虑了研制信息,因此在确保有较好的验证效果的条件下,与传统的试验方案相比,该方法可以明显减少故障样本量,或在样本量保持不变的情况下有效降低双方风险。
Aiming at the problem that the existing test verification scheme requires a large amount of fault samples, a new method of using the development information to develop a test verification scheme is proposed. First of all, on the basis of the binomial classical sampling model, the author constructs the basic trust assignment function based on the test data, the test predictive result and the expert experience by using the evidence theory method. Based on this, Type of test information to verify the test program, and to determine the existence of the program, given the program to determine the method, and finally carried out a case study. Due to the full consideration of the development information, the method can significantly reduce the number of fault samples or ensure that the sample size can be effectively reduced compared with the traditional experimental scheme under the condition of ensuring good verification results risk.