An agent model in which commuting, compliance, testing and contagion parameters drive infection in a population of thousands of millions. Agents follow Ornstein-Uhlenbeck processes in the plane and collisions drive transmission. Results are stored at SwarmPrediction.com for further analysis, and can be retrieved by anyone.
Covered in this post with the followup here where possible improvements are also discussed and acknowledgements are made. See also the SwarmPrediction.Com list of articles. The author is not an epidemiologist. The model expresses no opinions on the health aspects of COVID-19. The model offers a novel motion model with some interesting analytic properties also discussed in the article referenced above and presented in a more technical working paper shared on Overleaf.
pip install pandemic
>> from pandemic import run
>> run()
See also examples of library use and examples of using the public database of simulations generated by this model.
- pandemic.simulation.simulate is the main routine
- pandemic.client offers an alternative in object oriented style
See pandemic/client.py for examples of extending the newer style to include data storage, lockdown and so forth.
Pandemic can be run in a docker container.
docker run xtellurian/pandemic
See SwarmPrediction.com for an explanation of a SETI-like project to crowd-source a surrogate model.
Covered in detail in this article. As noted some analytical properties of special cases of the model are discussed in a working paper.
Opinions and issues are most welcome.