Resumen
The paper analyzes the spread of COVID-19 in 2020-2021 in 8 regions of the Russian Federation: Moscow, Saint-Petersburg, Krasnodar Territory, Stavropol Territory, Bashkortostan Republic, Sverdlovsk Region, Krasnoyarsk Territory, and Primorye Territory. Agent-based modelling is the main tool of our research. Information on lethal outcomes in the populations of the regions chosen for study was used as initial data. Models simulated the states of 39666133 residents, who communicated the infection through daily social contacts within their communities. The current states of agents in the adopted infection description method included susceptible, exposed, immune and dead. An adequate epidemiological model for each RF region was found with a numerical procedure based on the solution of the inverse problem with regularization. The procedure gave a virtual epidemic process which implemented a random scenario of infection transmission for all agents of a region in accord with actual retrospective data. This helped obtain additional, non-observable parameters such as transmissivity of infectious and immune agents. From the change of transmissivity we obtained coefficients for the growth of infection probability due to the seasonal factor and virus variation. Our results are illustrated by tables and plots. They show the epidemic processes in Moscow and Saint-Petersburg to be close with one another and much different from those in the other RF regions. The seasonal growth of transmissivity was found to be equal to about 1.55-1.77 times and independent of the locality latitude. For Moscow and Saint-Petersburg, the probability of infection through daily contacts is, on average, twice as high as that in the other regions.