AGENT BASED MODELLING OF COVID-19 SPREADING: FROM SIMPLE TO REFINED MODELS
Mikhail Kirillin,1, Aleksandr Khilov,1, Valeriya Perekatova,1, Ekaterina Sergeeva,1, Daria Kurakina,1, Ilya Fiks,1, Nikolay Saperkin,2, Ming Tang,3, Yong Zou,3, Elbert Macau,4, and Efim Pelinovsky,1,5
1 Institute of Applied Physics RAS, Nizhny Novgorod, Russia
2 Privolzhsky Research Medical University, Nizhny Novgorod, Russia
3 School of Physics and Electronic Science, East China Normal University, Shanghai, China
4 Instituto de Ciências e Tecnologia, Universidade Federal de São Paulo, São Paulo, Brasil
5 National Research University – Higher School of Economics, Nizhny Novgorod, Russia
Abstract
In view of a new wave of COVID-19, the ongoing pandemic remains one of the most significant current challenges. Development of novel models of COVID-19 spreading allow to understand regional features of the pandemic progression and work out prevention strategies aiming at minimization of population and economical losses. Traditionally employed SIR-type models or logistic equation based models provide quite good accuracy in simulation the general trends of COVID-19 pandemic in different countries, however, they lack accounting for different factors, such as population age structure, introduction of restrictive measures, individual agent behavior, population testing strategies, etc.
Agent-based models are a class of convenient tools that allow for a detailed study of the mechanisms of pandemic progression. These models consider interaction of agents within a chosen area governed by a predefined rules imitating their social activity. This approach is quite flexible and allow to account different aspects of social processes, including difference in activities in working days and weekends, introduction of restrictive measures and lockdowns. In this study we employ agent-based models to simulate the progress of the COVID-19 pandemic in Nizhny Novgorod region of Russian Federation. Based on our previous simulations of the first two waves of the pandemic, we consider four waves including the period of mass vaccination. For such an analysis, the simulation of the mass vaccination process was added to the previously developed “general pool” model. The study also considers a distributed pools approach which considers pendulum migration to the region capital from the surrounding towns.
The simulate dynamics of daily revealed cases and daily death cases demonstrate good agreement with the official data. The model reveals that the total numbers of new cases during the 2nd, the 3rd and the 4th waves significantly exceeds those that are revealed by the testing system, which is determined by both limited capacities of the testing system and social behavior. The simulation parameters providing the good fit of the official data implied that the number of contacts tested after revealing a new COVID-19 case decreases with the increase of the wave number.
The study is supported by RFBR (project no. 20-51-80004), CNPq (project no. 441016/2020-0), and NSFC.(project no. 82161148012).
Speaker
Mikhail Kirillin
Institute of Applied Physics RAS, Nizhny Novgorod
Russia
Discussion
Ask question
