Illuminating super mario bros
Oliver Withington · 2020
We aim to build on the current understanding of how Quality-Diversity (QD) algorithms can be applied to Procedural Content Generation (PCG) for video games with our platform, Illuminated-Mario. The work presented here has two core aims: to explore the viability of our approach to PCG and to compare the performances of MAP-Elites and SHINE, two state of the art QD algorithms. Our experimental results suggest both that our approach to level generation is a viable foundation for future PCG systems, and that while the SHINE variants implemented do not significantly outperform MAP-Elites, they warrant further investigation.