基于验前试验信息熵的测试性验证试验方案

Method for Testability Demonstration Test Design Based on Entropy of Prior Test Information

摘要:

产品在研制阶段存在大量的试验数据,为有效利用验前数据,降低测试性验证试验样本量,提出一种基于验前试验信息熵的测试性验证试验方案。该方案利用信息熵来度量研制阶段多次验前试验数据对测试性验证试验所起的作用,依据平均互信息熵和信息总量相等的原则,将多次验前试验数据等效成一次成败型数据；在此基础上,通过相容性检验方法确定验前数据与试验数据的相容性水平,并以Beta分布为验前分布,利用加权混合贝叶斯理论建立混合验后分布,之后,基于贝叶斯平均风险理论求解满足双方风险要求的试验方案；最后,以某型雷达发射分机为例,对其进行测试性验证试验研究,研究结果验证了该方案的有效性。

Abstract:

In the development stage, there is a large number of test data. Aiming to making good use of the prior test data and reducing the sample number of testability demonstration test, a method for testability demonstration test design based on information entropy of prior test information is proposed. In the method information entropy is used to check the role of different prior test data for the testability demonstration test, and many times prior test data is equivalent to success or failure data at a time according to the principles of equal mean interactive entropy and equal total information. On this basis, compatibility of prior data and test data is established by the method of compatibility test, while Beta distribution is made use as prior distribution and the mixed posteriori distribution is established by weighted hybrid Bayesian theory. Then, testing scheme satisfying the risk requirement of producer and consumer is solved based on Bayesian average risk theory. At last, to verify the proposed method, an example of a certain type of transmitter is given, and the result shows the validity of the proposed method.