Basic architecture of reflexive-active systems of artificial heterogeneous intelligent agents

The paper is devoted to the development of the basic architecture of reflexive-active systems of artificial heterogeneous intelligent agents including a standard functional structure of the system and a generalized architecture of agents. Due to the combination of heterogeneous agents in the system, it hybridizes various specialists' knowledge and methods of formal representation of the systems taking into account the instrumental and functional heterogeneity of emerging problems. The reflexive capabilities of agents allow them to model other agents, themselves, and the system as a whole reducing the intensity of conflicts and the duration of building interaction when the composition of the system changes. As a result, self-organization of the system in a strong sense becomes possible during which for each new practical problem, a solution method relevant to it is developed.