Definition of characteristics of reliability of radio-electronic systems

There are two main kinds of tests for reliability. One of them is standard test, which is used for large-scale products. The aim of the standard tests is to evaluate the parameters of reliability. Other type of the tests is control tests, which check the compliance to specifications of an indicator of reliability of the system. The paper studies the second type of the tests. The aim of the study is to decide whether these the characteristics of reliability of a product (made by radioelectronic systems) correspond to the set requirements provided by specifications on production of a product. For the solution of this task the mathematical apparatus of the statistical theory of hypotheses is used. Two hypotheses are considered: the hypothesis $H_0$ — the average time of operation for failure $t^*=T_0$ — is set by the requirements of specifications (a product is good); a hypothesis $H_1$ — the average time of operation for failure $t^*=T_1<T_0$ — alternative (a product is bad). The procedure of checking the hypotheses has a shortcoming, i. e. the quality of the solution is defined after carrying out the tests. Such procedure of checking the hypotheses is not optimum. The paper presents the consecutive procedure of checking the hypotheses (Wald's procedure), which assumes decision-making after each refusal and a stop of tests, if decision-making with the set quality is possible. The algorithm of checking the compliance of the law of distribution of the received selection to indicative or other law of distribution by criterion $\chi^2$ is shown.