Towards Non-Technical Designer Control over PCG Systems: Investigating an Example-Based Mechanism for Controlling Graph Grammars

Casper Harteveld, Abdelrahman Madkour, Stacy Marsella

Abstract: Increasingly, PCG systems are developed to help game designers create content for their games. However, game designers have limited control over the content current PCG systems generate. We investigate an interaction mechanism non-technical users can use to control generative grammars without the need for understanding the grammar’s rules. To demonstrate this control mechanism, we present a system, built using a probabilistic graph grammar, that allows designers to specify their desired generative space by defining a region on an expressive range plot. We ran a user study with game design students to assess its viability. Our findings suggest that designers have an easier time controlling the grammar using this mechanism over manually interacting with grammars rules.