Soroka M. Methods of building a multi-agent environment of intelligent training system for training air traffic controllers

Improved model of the training system of air traffic controllers, which, unlike the existing ones, is based on an agent-oriented approach, which allowed to create objects of environment with variable behavior and knowledge and give the training system signs of intelligence. Improved method of planning the behavior of agents in environment of the training system of air traffic controllers, which, in contrast to the known, is based on a modified method of forming a set of fuzzy binary conditions and continuing elementary plans using a set of fuzzy rules, which increased the efficiency of training air traffic controllers. The method of self-tuning the behavior of agents of the training environment of the intelligent initial system of air traffic controllers, which, unlike the known ones, is based on a certain set of fuzzy rules, knowledge of the results of agent interaction, which allowed to increase the variability of decision-making by air traffic controllers, was further developed. The practical significance of the obtained results is to bring the theoretical methods to their practical implementation. The use of a multi-agent environment and a full cycle of exercises reduced the number of air traffic controller’s errors by 10–18%, depending on the type of potentially conflict situations (PCS) and the circumstances of the PCS assumption, which proves the effectiveness of the developed methods. For different conditions of the reproducible air situation the developed environment of intelligent training system (ITS) of training of air traffic controllers prevails over existing simulators of training of air traffic controllers by 16 – 22%. The significance of the problem solved in the dissertation for science and practice is to further develop the theoretical and applied foundations of intelligent information technology to study the patterns of operators and their teams in navigation systems and traffic management.