Development of methods for evaluating complex organizational and technical systems using the theory of artificial intelligence

Complex organizational and technical systems are the object of research. The problem that is solved in the study is an increase in the efficiency of the assessment of the process of functioning of complex organizational and technical systems (OTS) while maintaining a given level of reliability. Methods of evaluating complex organizational and technical systems using the theory of artificial intelligence were developed. The originality of the research is: – in full coverage of critical events occurring during the functioning of the OTS. This is achieved due to the use of the Dempster-Schafer theory, which achieves the completeness of the assessment of the entire spectrum of critical events in the OTS; – in a comprehensive description of the process of functioning of the OTS. This makes it possible to increase the accuracy of OTS modeling for subsequent management decisions; – adapt to the type and duration of abnormalities due to multi-level adaptation of the artificial immune system; – in the ability to carry out initial adjustment of OTS knowledge bases using an improved genetic algorithm. This allows to reduce the computational complexity during the further formation of the OTS knowledge base by reducing the metric of rule formation in the OTS knowledge base; – in the ability to model the nature of the development of atypical events in the OTS due to the use of time series, which achieves the possibility of developing preventive measures to minimize the impact of the specified events on the process of operation of the OTS; – in the gradual reduction of the metric of the formation of the knowledge base about the states of OTS, due to the training of agents of the improved genetic algorithm. This allows to reduce the number of computing resources of the subsystem for assessing the state of operation of the OTS; The proposed methods provide an average increase in efficiency from 16 % to 23 %, while ensuring high convergence of the obtained results at the level of 93.17%.