Model for Verification of Agent Commutations of Distributed Multiagent Systems

Abstract

The development of multi-agent technologies and their widespread implementation in distributed computer systems is due to the properties of autonomy, mobility, intelligence and cooperativity, which provide power and flexibility in solving distributed, dynamic, fuzzy problems. At the same time, it is precisely these characteristic properties of multi-agent systems, that require special measures to ensure their operability and reliability of functioning, especially in the case of critical use. The proposed model for verifying the properties of dynamic communication is based on extended Petri nets and automata behavioral check and recognition experiments. Verification of two-tier and multi-tier communications of arbitrary spatial and temporal topology includes behavioral check of the mechanisms of conditions, preparation, transmission, buffering, broadcasting, taking into account probabilistic-temporal, group and concurrent characteristics. The verification model of agent commutation is designed to determine the conditions for analyze, check and recognizing of agent interactions in diagnostic methods for distributed multi-agent systems and allows to reduce the computational complexity of analysis, check and recognition of behavior of multi-agent systems.