Managing the Composition of Training Reflexive- active System of Artificial Heterogeneous Intelligent Agents

The paper discusses an approach to constructing virtual reality training simulators based on reflexive-active systems of artificial heterogeneous intelligent agents. Thanks to the dynamic composition and structure of agent connections, virtual reality training simulators based on reflexive-active systems of artificial heterogeneous intelligent agents will be able to flexibly adapt to the user's capabilities, taking into account his abilities and progress in learning. Agents of such systems are characterized by the ability to reflexively model each other's behavior, including an interface agent that ensures user interaction with the system and represents the user's agent in the virtual system of agents. Thus, the most relevant use of such simulators is to train personnel in control mechanisms in environments where control objects show signs of subjectivity. This paper examines in detail the mechanisms for managing the composition of agents of the proposed systems.