SOFTWARE MODELS IN THE SAFE AND RELIABLE OPERATION OF A COMPLEX SPACE SYSTEM 

Severtsev Nikolay Alekseevich, doctor of technical sciences, professor, chief researcher, Federal research center «Computer science and control» of RAS (Dorodnitsyn computer center of the Russian Academy of Sciences) (119333, 40 Vavilova street, Moscоw, Russia), E-mail: severs@ccas.ru 

681.2.084-192 

10.21685/2307-4205-2019-4-1 

Background. An analysis is made of the distribution function of the observed software testing results that affect the reliability and safety of the spacecraft flight. The parameters used to evaluate the relative effectiveness of the safety assessment models and the degree of their adequacy are the current error flow rate, Brown statistics, likelihood function, and Kolmagorov distance.
Materials and methods. An analysis of the development of adaptive models for evaluating software reliability shows that a priori it is impossible to single out the most effective model for evaluating the reliability of software being worked out. The degree of adequacy of the safety assessment model can only be clarified a posteriori from software testing results. It is shown that a very specific measure of reliability and security is involved in the software, which is determined by the input data options.
Results. A measure of software reliability is determined through the density of time distribution. As indicators of software reliability at time t, we consider either the number of errors remaining in the software, or the time until the next error, or both indicators together.
Conclusions. the model under consideration is quite unified. It combines well-known approaches and makes it possible to further generalize on the basis of the theory of stochastic filtering both in the linear and nonlinear cases.

simulation, distribution function, software, reliability, security, space system