This project's purpose is to evaluate the efficiency of six different types of broadly defined collaborative algorithms for the task of a predator chasing a prey (possible applications in law enforcement?)
1- Fully centralized collaboration.
2- Fully centralized collaboration of independent subteams.
3- Encrypted communication of actions between predators.
4- Communication of actions between predators that can be overheard by a prey.
5- Individual actions while taking into account positions of other predators.
6- No collaboration at all.
We also evaluate two different types of stigmergic communication algorithms, one in which predators avoid each other's trails and one in which they follow each other's trails.
We used the Berkeley pacman AI framework for graphics.