Optimization of the Structure of the Intelligent Active System as a Necessary Condition for the Harmonization of Creative Solutions

Abstract

This article is devoted to the problem of identifying the most effective intelligent agents suitable in their personal qualities for solving non-standard, complex, creative tasks. The effectiveness of the decisions made by intelligent agents determines the productivity of the entire system. The possibilities of generating creative solutions to complex, non-standard problems largely depend on the state of intelligent agents. An intelligent agent, taking into account its inherent anthropomorphic properties and qualities, can be in various psycho-emotional states: from a depressed mood to creative uplift and insight. The states of the agent change over time, which he informs the system about at specified points in time. It is assuming that all agents provide reliable information about their states to the system. Based on the data obtained, the system should determine the most and least stable agents and redistribute the functions and responsibilities between them to maximize own utility function. The task is characterizing by the need to analyze large volumes of subjective information and belongs to the class of difficulty formalized tasks. To solve it, fuzzy clustering algorithms that are well implementing in the Matlab software environment can be using.