Interactive Ecosystem Simulator
Abstract: As ecosystems are complex domains, both analytical and computer-aided modelscan aid in gaining insights about their dynamics. One such computer-aided modelis the concept of ecosystem simulation. This project aims to build an interactiveand visual ecosystem simulation in the Unity game engine. The purpose is to explore how modelling of animal behavior, trait evolution and dynamic terrain can becombined with a graphical representation to create an interactive ecosystem simulator. Implementation of these aspects includes exploration of machine learning andreactive behavior for animals, terrain generation, genetic reproductive algorithms aswell as run-time visualization and collection of data. The effects of these aspects areevaluated using comparisons between animal behavior models, impact of terrain andoutcomes of genetic evolution, in addition to software interactivity. The outcomeof this project indicated that the machine learning prioritization animals performednearly as well as reactive rule based animals in terms of survival, while the machinelearning steered animals performed sub-par in comparison to the others. Furthermore, it showed that terrain changes seemingly has a greater impact on the predatorpopulations compared to the prey populations in the simulator. Additionally, as aresult of the proposed evolution model, genetic traits of animals indicated to bepotentially adaptive to the environment. Finally, the graphical representation provided visual feedback and information to users. In total, the final product resultedin a working interactive ecosystem simulator. The implications of this thesis offers abaseline framework for modelling a visual interactive ecosystem simulator in regardsto future research, academic and entertainment applications.
AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)