This project was aiming to create a simulation involving multiple agents all working towards collecting enough food.
I created it with my partner Mary Stufflebeam during the 2011 Facebook hackathon. We wrote the backend in Java, which controlled the agents as they moved around and collected food. We wrote the frontend in python to give a graphical view of what was going on in the simulation.
We implemented several simple reaction agents to control the gatherer agents, and then we started to work on adding predators. However we found it extremely difficult to have a stable population of both gatherers and predators.
The reason, I believe, is that we simulated with a constant amount of grass every round, but the number of gatherers would fluctuate. Also we worked only with discrete generations, which limited the meaningfulness of our simulation.
The next steps would be to allow the algorithm to change itself. Also, we should change the grass to be able to grow back, so there would hopefully be some equilibrium that maintains the populations of both the predators and the gatherers.
I built this project with Mary Stufflebeam.
You can view the code for this project here.