Calibration of a regional agent-based travel demand model to simulate the spread of COVID-19

Abstract

The recent COVID-19 pandemic has shown the need of readily available tools to simulate the spreading of infectious diseases and assess the potential impact of policy measures aimed at the containment of the disease.

The most common approach in epidemiology is to use compartmental models which model disease spreading as a series of stocks (compartments) and flows. The most basic version considers three compartments: susceptible, infectious and recovered. One of the core assumptions of these models, however, is a homogeneous population which is a serious limitation when it comes to clustered outbreaks. The research at hand uses an agent-based travel demand model (MATSim) coupled with a recently developed extension (EpiSim) to simulate the spread of the pandemic by tracking the interactions of agents und subsequently identifying infections by following their contact network. This overcomes the compartmental assumption of the previous models.

In this presentation, we summarize the calibration results of the EpiSim model which is based on the regional activity- and agent-based travel demand model of the trinational region around Basel. The calibration procedure is based on data on case numbers and hospitalisations from three countries (Switzerland, Germany and France) at different spatial resolution. The underlying models account for the effect of border closures and spatially varying restrictions to limit the spread of COVID-19. In contrast to existing studies, this model pays particular attention to disease import through external traffic from outside the model area as well as cross-border travel demand.