Procedurally Generated Cities for Game Environments

Year long project (2014)
This research uses procedurally generated cities as an example to combine Evolutionary Algorithms with Autonomous Agents as digital cities are otherwise well researched and understood.

People involved:
Jan Kruse – Researcher
Dr. Andy Connor  – Supervisor
Dr. Charles Walker – Supervisor

Description:
Convincing 3D models are a prerequisite for the creation of immersive games, movies and virtual realities. Creating high-quality content requires substantial skill and time and is subject to high costs and multiple iterations. Certain objects from areas such as architecture or biology necessitate fine detail and a very high number of small building blocks, which intensify the problem even further.

This project seeks to address these issues by using Human-centric Evolutionary Algorithms to support the designer and therefore reduce the human labour required in the process.

This research uses procedurally generated cities as an example to combine Evolutionary Algorithms with Autonomous Agents as digital cities are otherwise well researched and understood. This supports the researcher’s hypothesis that the combined effort of algorithms and agents provide a more effective, efficient and practically employable approach to digital asset creation.

Contact:
Jan Kruse –  jkruse@aut.ac.nz

Images