Calculate Weird Things. Twitter | Facebook | Email
Home | | Blog | Comments
Custom Search

Zombie Invasion Simulator

An epidemiological model of a zombie outbreak. For use in theoretical epidemiology.

Powered by NetLogo
View/download model file: zombies.nlogo


This model is open-source! I built this for a mathematical biology class in college (actually my final project for the class). As noted in the links above, you can download NetLogo and the code for the simulator, modify the code, and build your own, perhaps more intricate model. Or you can download everything and play with this tool offline, which might be good for long trips or something.

How it works

This zombie model is based on a basic predator-prey-food model, a variation of which I imagine is used as an example in most mathematical biology courses. The blocks on the background change colors to indicate whether food is available at that location or not, and the rate at which food regenerates can be controlled by the foodtime slider. Humans (black) eat food, whereas zombies (red) do not. Zombies, of course, eat people.

Where this model varies from a standard predator-prey model is in the assumption that humans will, on some scale, fight back. The model assumes that, in a given interaction between a functioning human and a zombie, one of two outcomes will result: either the zombie will win and turn the human into a zombie (the zombie's only means of reproduction), or the human will kill the zombie first. This is characterized as the win rate for zombies (0-100% of interactions) in the zombie-win slider.

By default, I have assumed that zombies will die off naturally (decay) faster than humans will, but you can change this to model different scenarios by adjusting the remove-percentage-zombies and remove-percentage-humans sliders. As the names imply, these sliders adjust the percentage of characters (what NetLogo calls 'turtles') of each type that die of causes external to the model. These causes include traditional disease and human-human violence for humans, and starvation and decay for zombies.


What I've found in playing around with this model is that our survival in a zombie scenario depends largely on the mortality of the zombies. The assumption that I tend to make about a zombie virus is that it would work by overheating the brain or something similar, reducing its functionality to primal kill/eat urges. In observed cases of such occurrences (both drug- and disease-induced), the brain usually doesn't last very long. A zombie would likely have only a few hours at most to find and infect as many victims as possible before the brain stem goes the way of the rest of the organ. If an infection started in a rural area, it's likely that the only victims would be in a single household before the disease died out. In a more urban setting, though (with a dense population as you can set with a slider), we can actually see many scenarios in which zombies outnumber humans for a while.

In cases where a zombie outbreak really gets out of control, our role as the most dangerous game becomes our greatest asset. The more often humans win in encounters with zombies, of course, the more likely we are to survive as a species. What's amazing to me is how crucial it is that we win a majority of our bouts with zombies. This does not necessarily mean employing firearms, but they would certainly be a welcome addition to our arsenals which would otherwise include cars, propane tanks, flares, chainsaws, machetes, and pretty much anything we can make explode.