Agent Based Terrain Generator: Cruthú

Terrain generation models are applied in different industries and fields of study. Current techniques assist in planning transportation networks, visualizing population migrations, conducting epidemiology research, and training self-driving cars and drones. Many current applications and research models generate complex realistic artifacts, e.g., trees, rivers, coastlines, populations, and even cities. Unfortunately, most of these techniques are described and implemented separately. Techniques do not work together to generate a complete holistic view. This thesis proposes a model, Cruthú (Gaelic for creation ), that provides a novel platform allowing for complete world generation by integrating existing research and algorithms. The model is inspired by the Unix command-line pipeline where independent tools can be linked together to handle a complex workflow. Cruthú proposes a similar approach. This thesis uses the proposed model to create a framework application. To do this, we have investigated several industry standards to allow a generation module to present a robust, unified application program interface. Cruthú then uses dynamic loading and a pipelined architecture to generate a complete world. Cruthú is highly modular, allowing components to be swapped in, swapped out, or re-ordered to achieve desired results. A prototype model was built that implements the main features, and further work will allow for individual algorithms to be integrated in more complex interactions. In this prototype, common terrain generation techniques were used to show the ease of implementing algorithms