Method of synthesis of a multilayer model of a monitoring software agent
Serhii Holub, D. V. Tolbatov · Mathematical machines and systems · 2023
The paper describes a new method of building multilayer models by a monitoring software agent and its usage for forecasting the number of COVID-19 cases in Ukraine. Monitoring agents are used to provide decision-making processes of different levels of urgency with information and knowledge through continuous observation of objects and intelligent analysis of the observation. The need to solve intellectual problems of classification, clustering, identification, forecasting, etc., in the process of performing monitoring tasks led to the usage of intelligent agents for this purpose. In addition, there is a special class of autonomous software systems – monitoring agents that have a special structure and individual functionality, utilize atypical methods of interaction with the external environment and specialized methods of processing the observation results, and require the creation of new methods of model synthesis. After the outbreak of COVID-19, scientists began to use modern technologies that could help fight and overcome this terrible disease. They started to teach the models to identify the disease and its various symptoms and to plan the ways of treatment. The insufficient informativeness of the observation results, on the basis of which models are trained, is typical for forecasting the pandemic development. Effective means of overcoming this problem are the use of multilayer models of observation objects. Our study describes a new method of building multilayer models by a monitoring software agent using the example of the task of forecasting the development of the incidence of COVID-19 in the population of Ukraine. This can make it possible to control the load on hospitals and plan the time and duration of social restrictions for the country’s population, in order to slow down the development of the pandemic in the conditions of the military aggression of the Russian Federation.