Abstract:

The report is devoted to the architecture, structural organization and development of a multi-agent system of representation and processing of knowledge (MASKRP), as well as its use in the educational process. Models of reactive and cognitive software learning agents are presented. Examples of interaction diagrams, states and transitions of training agents, rule systems and queries of a problem-oriented multi-agent solver are given. A classical logic solver and a multi-agent solver developed on the basis of models of training software agents are compared. The results of fuzzy queries to the knowledge base, which were implemented in accordance with the sets of fuzzy rules and membership functions specified in the example, are shown. The results demonstrate the practical viability and effectiveness of the use MASKRP in the educational process. The basis of design decisions when creating MASKRP laid down the requirement to support users with different levels of competency. A problem-oriented visual design toolkit has been developed to support knowledge acquisition and user-friendly configuration of automated reasoning mechanisms.

Keywords:

Intelligent Learning Systems, Multi-Agent Systems, Knowledge-Base Systems, Intelligent Software Agents, Fuzzy Systems, Adaptive Leaning Systems.