Iershov S. Theoretical basis of model-oriented construction of fuzzy intelligent multi-agent systems

The dissertation is devoted to solving problems of theoretical foundations of fuzzy intelligent multi-agent systems based on model-oriented approach. Model-based approach to the development of multi-agent systems is justified, which allows to solve problems related to the number of technology platforms for realization but can also accelerate the development of such systems. A process of system developing is proposed to specify by systems of transformations of fuzzy graphs, the result of which provides a number of most adequate platform-specific models of multi-agent systems. Systems of transformations of distributed fuzzy graphs are developed that can model dynamic network structure of intelligent agents. The conceptual basis for building fuzzy multi-agent systems is justified based on specification of transformations of fuzzy graphs and adaptation of fuzzy rules by coordinator component. Complex of multi-agent systems models based on type-2 fuzzy logic is constructed. A model-oriented method for generation of platform-specific models of fuzzy multi-agent systems that operate in high-performance environments is developed. Basic transformations that allow to generate fuzzy models of intelligent agents are constructed.