Simulating the spread and containment of COVID-19: an agent-based modelling approach

Abstract

Background:The ongoing pandemic, caused by the virus SARS-CoV-2, posed unprecedented challenges to decision makers worldwide. In a very short time, a variety of policies for disease mitigation were introduced and simulation models were applied to evaluate the effectiveness of these policies.Methods:We develop an agent-based model in Netlogo 6.1.1 capable of simulating common containment strategies, with a special focus on testing and tracing policies. Within this framework, we aim to analyse the impact of various parameters – mainly those influencing the testing process and the compliance to quarantine measures – on the performance of these strategies.Results:The calibration of unknown parameters regarding asymptomatic courses of the diseaseled to the conclusion, that around 35% of all agents undergoing the disease will remainasymptomatic and only be half as infectious as agents that develop symptoms. There sults further show that hygiene and closing measures are quite effective when the population behaves fully compliant. The effectiveness of household and individual tracinghighly depends on the testing process, only leading to a successful mitigation of thedisease if enough symptomatic people get tested and enough test capacity is available, to quickly evaluate the status of people queued for testing.