Deep learning for procedural level generation has
been explored in many recent works, however, experimental
comparisons with previous works are either nonexistent or
limited to the works they extend upon. This paper’s goal is
to conduct an experimental study on four recent deep learning
procedural level generation methods for Sokoban (size = 7 × 7)
to explore their strengths and weaknesses and provide insights
for possible research directions. The methods will be bootstrapping conditional generative models, controllable & uncontrollable procedural content generation via reinforcement learning
(PCGRL) and generative playing networks. We will propose
some modifications to either adapt the methods to the task or
improve their efficiency and performance. For the bootstrapping
method, we propose using diversity sampling to improve the
solution diversity, auxiliary targets to enhance the models’ quality
and Gaussian mixture models to improve the sample quality.
The results show that diversity sampling at least doubles the
unique plan count in the generated levels. On average, auxiliary
targets increases the quality by 24% and sampling conditions
from Gaussian mixture models increases the sample quality
by 13%. Overall, PCGRL shows superior quality and diversity
while generative adversarial networks exhibit the least control
confusion when trained with diversity sampling and auxiliary
targets.