Agent-based modelling of the first and the second waves of COVID-19 spreading in Russian Federation regions
The COVID-19 outbreak that has grown into a full-scale world crisis remains one of the current largest humanity challenges. Development of strategies for the disease spread prevention aiming at minimization of population and economical losses requires adequate models allowing to explain and predict the trends in the pandemic progression. Traditional simulation approaches based on the derivatives of a SIR model allow to predict general trends lacking, however, account for random factors. Agent-based models provide a more flexible tool accurately accounting for different factors, such as population age structure, introduction of restrictive measures, individual behavioral response, and population testing strategies.
In this paper we report on the development of an agent-based model of COVID-19 spreading and its application for several representative regions of Russian Federation based on available data on population structure, self-isolation index, and testing strategies. The approach is based on an agent-based model with a general pool with a day-scale time resolution that incorporates the rules of the population testing strategy. The simulation covers the time period from February 2020 till April 2021 that contains two waves of the pandemic. The considered areas include Moscow, Niznhy Novgorod region, and Novosibirsk region. The tuning parameters from the model are self-isolation index and testing strategy parameters.
The simulated dependencies of daily new cases and COVID-19-associated deaths are in good agreement with official statistics. Simulation results allow to reveal the differences between official statistics determined by the testing strategy and real situation with the disease spreading. It is demonstrated, that the relative number of infected agents that were not detected by the testing system during the first wave of the pandemic is significantly lower than that during the second wave owing to the limited capacity of the testing system.
The study is supported by RFBR (project no. 20-51-80004), CNPq, and NSFC (project no. 82161148012).
Link to video presentation: https://cloud.mail.ru/public/rfKS/dp47fLi1E
Institute of Applied Physics RAS