At the beginning of the COVID-19 pandemic, an idea came to my mind to model the spread of the SARS-CoV-2 virus in different populations. This led to a preprint that I publish in medRxiv back in March 2020. And eventually, I added a new section to it showing how we can optimise a limited number of vaccines in a population to maximise their impact, which I published as a paper in PLOS ONE.
I also made some predictions on how the virus can spread in different
populations based on the data I could gather of the current state of the spread at the
time. You can see those results below. I decided to leave the analysis as it
was, as a casestudy to see what can go right and wrong in mathematical
epidemiology.
If you want to run your own simulations, you can find the code in this GitHub Repository. The code includes the results of the optimisation scheme. If you want to run the optimisation scheme yourself, please let me know and I can send you the code. But please note that the code needs Global Optimisation Toolbox in Matlab and if run on a typical PC, might take hours. I ran it on 40 hyper-threaded E5 cores and it can take around 15 minutes on such a cluster to run.
The above-mentioned manuscript uses Germany as a case study. You can look into
that for all the details of the simulations and results, including the results on how to
optimise the limited number of vaccines in the population of Germany to obtain the
best outcome.
I have used the mathematical model to simulate the spread of COVID-19 in Iran too. I wrote a report in March 2020 which predicted a grim picture of the situation in Iran in the months to come. You can read the report here. In November 2020, I updated that report and looking at how the situation has unfolded, or better to say an estimate of that, given the complete lack of transparency in Iran. You can read the updated report here.
Last updated on November 9, 2021.