Skill-based difficulty adaptation for random game plot generation

Develop an algorithm that assesses a player's skill level and, based on this assessment, generates a random game plot (puzzle graph) with a suitable difficulty level using the procedural graph-grammar approach described in the paper.

Background

The paper introduces directed non-confluent graph grammars (Dnc-eNCE and Dnc-eNCE-JGG) and a procedural content generation algorithm to produce solvable puzzle graphs for 2D game levels. While the proposed method can generate diverse random plots via regular-controlled application of production rules, it does not yet adapt difficulty to the player's skill.

To sustain engagement and tailor challenge, the authors point out the need for an algorithm that estimates player skill and adjusts the random generation process accordingly, ensuring that generated plots remain solvable and appropriately challenging.