Formation of Artificial Collective Intelligence Based on the Theory of Active Systems

The object in this paper is the process (its content, structure, dynamics) of formation of artificial collective intelligence at a variety of enterprises united by socio-economic and/or production relations. Production clusters and self-regulating organizations (SRO) are identified as forms of such relations. The research of concepts related to the category of “intelligence” has been carried out. It is substantiated that the tools for organizing and functioning of clusters and self-regulating organizations should be more actively used in the formation of transport processes. In particular, it is shown that transport and logistics chains of freight delivery have both the properties of clusters and the properties of self-regulating organizations. The mechanism of formation of transport and logistics chains of freight delivery is considered. The spectrum of mathematical tools for the synthesis of artificial intelligence systems is analyzed. In particular, the theory of pattern recognition, fuzzy sets, games, percolation, active systems. It is substantiated that the models and methods of the theory of active systems, designed to stimulate the joint work of enterprises and coordinate their divergent economic and production interests, are an effective tool for the formation of collective intelligence of a variety of interacting agents (enterprises and top-level management structures). The advantages and disadvantages of the basic mechanism for coordinating conflicting interests of economic agents are analyzed. The author's results of the development of basic models are described regarding: a) compensating for the inadequacy of models in practical applications, b) expanding the scope of methods application. In particular: according to the point a) the requirements for the quadratic of the cost function, the absence of initial production costs, and the full sale of manufactured products have been removed; according to point) models for sequential and parallel operation of enterprises in clusters and self-regulating organizations have been clarified. New directions of the research on the topic are substantiated: taking into account several criteria for the quality of enterprises operation and the multidimensionality of influencing factors, checking the adequacy of criterion models (the adequacy interval should be wider than the area of agreed solutions), organizational basis for the formation and use of the obtained collective intelligence.