Boids Flocking Algorithms : Prey Survival Rate in Predator-Prey Simulations

Abstract: Virtual environments are only perceived as real as each part makes it out to be. One aspect of these simulated realities is the movement of the artificial life entities included in some of the virtual environments. These can for example be different kinds of so-called animats, simulations of animal behavior, which are types of artificial life programs to which the boid algorithm belongs. Boids, an algorithm for imitating animal flocking behavior, has different variations that have been introduced over the years. It has additionally been extended to simulate different scenarios, such as predator-prey simulations. Earlier work has been done in the field to generate results that provide the best possible parameters for the prey boids, in order to allow as many as possible to survive the predators. The parameters produced in that work are the ones belonging to the mathematical rules of the algorithm, making the so-called emergent behavior of the prey seemingly more intelligent. In this paper, the aim is to extend the work in the field, by examining how two different implementations, variations, of the boid flocking algorithm affect the emergent behavior, and in turn the survival rate of prey, in predator-prey simulations. It is also examined how changing the amount of neighboring boids taken into account in the calculations affect the survival rates. After the experiments, and evaluation of the results, it was found that a newer variation of the boids algorithm performed better than the original when it comes to emergent behavior and survival rates. It was also found that changing the amount of neighbors used in the calculations affects the emergent behavior, where more neighbors taken into account generally leads to higher survival rates.