Simulation of the Spread of a Virus Throughout Interacting Populations with Varying Degrees and Methods of Vaccination Jack DeWeese Computer Systems Lab Abstract After doing some research into other virus simulation programs, I found Jill Dunham’s program which serves as a good base for my own simulation. While I began this quarter working on my original simulation from scratch, I moved to doing research on how other researchers worked with different variables and the differences between dynamic and agent-based modeling. It was during this research that I came upon Mrs. Dunham’s model (shown on the left) and began learning how the software works. During the year I was able to make extensive changes to the simulation software. These changes include, but are not limited to, implementing my own smallpox model, some tweaking of her pre-done viruses, the addition of hospitals, human behavior to go. To finalize the program to best work with my regression models, I added several methods to export, compile, and make computations with the data from each simulation run. Requirements The broad requirements of the project are to be able to realistically simulate the spread of a virus throughout a population, this population being one that can be easily modified and the virus being one that can likewise be easily modified. The program must also be able to show the effects of vaccination in prevention of the virus spread. Overview Using the software created by Jill Dunham (Dept. of Mathematics, George Mason), based off of the MASON platform, I have implemented additional variables, such as buildings, types of people, viruses, etc. in order to test vaccination methods. The software creates people, randomly assigning a type (i.e. profession and age group) and a home. In the case of a student, a school will be assigned, in the case of a doctor, a hospital will be assigned, and so on. The graphic above shows a test run of my simulation; the green represent uninfected, red – infected, and black – dead. In the my later models, blue represent vaccinated. Research I did a lot of research on different types of viruses mostly smallpox and the flu. I also did some on the common cold, which differs from influenza in that the symptoms are not as severe and the chance of death is virtually zero. The CDC provided most of the data on the viruses I tested. I had originally intended to model travel via air, but due to the scope of my populations and the research concluding that airliners are not as infectious as I had previously suspected, this will not be accounted for. Testing and Analysis By running hundreds of iterations covering ranges of variables including vaccination percentages and population sizes and having my software compile this data into averages, I was able to perform regression models on my results using the first equation shown on the right. The graph to the right depicts the results of the number of deaths due to a virus in a population of 300 and an independent variable of vaccination percentage. The second equation to the right is used to calculate the ideal vaccination percentage given that r v is the rate of death from a vaccine. By running iterations over population, the coefficients of the first equation can be generalized as a function of population size allowing mathematical modeling (the third equation shown to the right), which saves significant processing power if models need to be run several times.