Simulation Model of the Neurocognitive System Controlling an Intellectual Agent Displaying Exploratory Behavior in the Real World

The paper presents the neurophysiological mechanisms of human exploratory behavior. The possibility of modeling such behavior in an intelligent agent based on a multi-agent neurocognitive architecture is shown. An autonomous software agent has been developed, in the control multi-agent neurocognitive architecture of which a system of intrinsic motivation is provided for the implementation of exploratory behavior. This behavior is aimed at forming the completeness of the functional representation of the fragments of the “intelligent agent - environment” system observed with the help of the agent's sensors. Intrinsic motivation is an algorithm for stimulating an intelligent agent to perform certain behavioral programs by forming an internal stimulus (reward) to perform this program in a multi-agent neurocognitive architecture. It is shown that in the process of exploratory behavior, an intelligent agent immersed in a communicative environment forms a speech and information model of the world due to the dynamic formation of functional systems based on the cooperation of neural agents.