Polyagent functionals in information technologies of intelligent monitoring

Abstract

The application of a multi-agent approach to the implementation of intelligent monitoring technologies in the form of monitoring information systems allows them to acquire new properties under replacing monitoring tasks. In crisis situations, when the properties of monitoring objects change dramatically, the monitoring information system is required to be able to obtain information about new properties of objects in the absence of new observations of the behavior of these objects. The informativeness of existing data sets is declining. To provide information on decision-making processes during the period of accumulation of the results of new observations requires a significant increase in the capacity of the means of extracting this information from existing data sets. Towards this, it is proposed to build monitoring information systems as a set of agents and methods of building relationships between them. It is proposed to provide the possibility of increasing the power of information transformation means by building supra-agent formations – agent functionalities. Since the requirements for agent functionals at different levels of the hierarchical structure of MIS are different, it is proposed to distinguish multiagent and polyagent functionals. They differ in purpose, structure, compatibility of arrays of input data on which their agents were built, methods of formation of relations between agents. What they have in common is the echelon structure, ensuring the adaptability of information technologies of intelligent monitoring to changes in the properties of objects and replacement of monitoring tasks. A new method of forming relations between agents in the form of ascending construction of the structure of agent functionals is proposed. The process of formation of polyagent space is described. The dependence of the emergence of the agent functional on the informativeness of the polyagent space has been experimentally proved.