Monitoring of multi-agent systems under conditions of uncertainty

The current stage in the development of information systems is characterized by the use of multi-component distributed architectures and methods of intellectual processing and analysis of large amounts of heterogeneous data. At present, the problem of ensuring the reliability and safety of such systems has not been sufficiently investigated. The purpose of the work is to develop a method for computer analysis of algorithms for monitoring the operation of multi-agent systems based on the proposed member. Cor. Academy of Sciences of the USSR A.A. Vavilov principle of successive disclosure of topological, structural, parametric and signal uncertainty. Algorithms for analyzing multi-agent systems using probabilistic methods and fuzzy logic, formal models and methods of the theory of queuing systems, Petri nets are considered. The results of the experiments demonstrate an increase in the efficiency of analysis of multi-agent systems using the developed algorithms and software for monitoring and reconfiguration under uncertainty. The network software package is intended for use in teaching students and in conducting research.