Resumen
The subject of the study is the methods and means for identifying the state of a computer system of critical purpose. The aim of this article is to conduct an analysis and development of a general scheme for the synthesis of identification measurements in the system of identification of the state of a computer system of critical purpose. The article solves the following tasks. The methods of selection of informative indicators of state identification and methods for the synthesis of identification measurements in a computer system of critical purpose are analyzed. As a result of the analysis, the main optimization problems are formulated and an example of the possible use of the known mathematical device in their solution is given. When solving the tasks, methods of multi-criteria evaluation, discriminant and cluster analysis, mathematical statistics and comparative approach were used. Results obtained. Studies have shown that the general scheme for identifying the state of a computer system of critical purpose should include methods for identifying abnormalities and methods for identifying abuses. The task of identifying anomalies should be solved under very difficult constraints on the authenticity of the results of identification, and the efficiency of the solution to this problem should be worse than the requirements set forth in the guideline documents. As a result, a general scheme for identifying the state of a computer system of critical purpose that differs by the complex use of advanced identification methods and their adaptation to possible changes in input data is synthesized. The set of possible input indicators for state identification was experimentally determined. A comparative analysis of state identification methods was conducted. The timetables for identifying abuses in a computer system of critical purpose, depending on the number of recorded data were received. Conclusion: the complex use of the methods of identification of anomalies allowed to increase the authenticity of identification to 1,9 times, and the complex use of methods for identifying abuses in computer systems of the critical purpose allowed to increase the efficiency of identification up to 2 times.