Abstract

Most procedural content generation techniques heavily rely on input parameter design. However, few consider the role of human designers in the process of applying procedural generation. The variability of configuration parameters further complicates the integration of procedural algorithms into the game design process. This thesis proposes a unified representation of gamespaces capable of modeling the properties of spaces and objects in games. The defined framework illustrates a design process that uses these properties to automate the configuration of input parameters for procedural content generators. The introduced Gamespace Graph offers an interface to intuitively design gamespaces while simultaneously configuring procedural inputs. It establishes a hierarchical hypergraph structure that formally represents gamespaces and provides simplified perspectives and design tools. The concept of a Data Transformer is presented as a tool for interpreting the gamespace graph’s structure and data into configurations for content generators. It acts as an intermediary layer between the content designer and the developer of procedural algorithms and streamlines the content design process by significantly reducing the parameter design effort. A Unity framework is developed to implement the theoretical concepts. It validates the system’s properties within a well-known commercial game engine, evaluating the resulting design process concerning various procedural content generation algorithms. Additionally, specialized procedural algorithms are developed to leverage the gamespace graph structure and produce refined generation results.

Results/Implementation/Project Description

Conclusion

[ PDF (optional) ] 

[ Slides Kickoff/Final (optional)]