Simulating an Ecosystem

Abstract

Simulations are computer programs based on models, used to imitate real-life scenarios. Simulating an ecosystem is a great challenge given the various factors that interact and animals that coexist. The thesis work presented in this paper aims to model an ecosystem and study what effect the number of resources, animal characteristics, and animal behaviors have on the natural selection of the simulated animals. The animals evolve through natural selection, implemented using genetic algorithms. Specifically, the project aims to investigate how the population and the traits of the animals change, based on interactions with the environment and with other species. This is done by gathering data while running the simulation. The ecosystem, with all its components, was built using the Unity game engine. The rabbits and wolves, that constitute the food chain, were modeled to be accurate but simplified versions of their real-life counterparts. The results show that reaching equilibrium is possible, albeit difficult if the right start values and correct world size are not chosen. Furthermore, the number of animals depends on the number of plants, and the traits of the animals change over time. The conclusions to be drawn from the results are that the instances of the simulation with the genetic algorithm implemented were able to run for longer compared to the instances without the genetic algorithm. The results also suggest even though it is greatly simplified, that the simulation proved to be stable for certain initial numbers of rabbits, wolves, and plants. This stability can be greatly improved in order for the simulation to handle greater variations in resources and the number of animals.