This simulation runs an epidemic modeled closely on COVID.
It allows interactively playing with different levels of countermeasures and different strategies.
The underlying model is a diffusive multi-region stochastic SEIR model for the disease. It includes seasonal variability and overdispersion, as well as a low constant influx of infected from abroad. The counter measures are modeled as affecting the reproduction rate R. Testing, contact tracing and isolating contacts is modeled as having a finite capacity to stop infections. Once too many people are infected tracing contacts becomes infeasible.
Note that the reduction in R is a combination of official measures and personal choice/compliance. An intermediate level of reduction might be the result of medium measures, strict compliance and individual responsibility, but could also result from stricter measures that are not adhered too due to lock down fatiguee for example.
A certain fraction of people infected are assumed to die, with the fraction going up when hospitals are overloaded. Vaccinations, once they become available about one year after the start of the simulation, reduce the death rate (up until the vulnerable 20% are vaccinated) and the reproduction rate (proportionally to the fraction of people vaccinated).
The underlying travel on which the virus spreads from region to region is generated by a stochastic model based on the data published by the Roland Koch Institute here: [https://www.covid-19-mobility.org/]
While many phenomena that were observed throughout the pandemic can be seen in these simulations, the simulations should not be treated as predictions. In particular we did not extensively tune the parameters real data.
A mathematical description of the underlying model will be made available eventually.
First, install dependencies and the build environment by running
npm installAfterwards, you can build the sources by running
npm run build