Dynamic 2D Dungeon Game Level: A Synergistic Approach with Binary Space Partitioning and Genetic Algorithm

The rise of the fast-paced gaming industry forces developers to produce high-quality content to satisfy players worldwide. Good games require dynamic content to secure the flow of players. Leveraging variational game content can be achieved by using the procedural content generation (PCG) technique to enable automatic creation without being manually crafted by the game level designer. A dungeon-based level is one of the strategic objects for PCG applications and has been popular in many adventure and role-playing games. Several studies only focus on developing 2D dungeon levels to produce dynamic game rooms and corridors. However, the generated room contents are frequently ignored when deploying dynamic game levels. Thus, this study aims to implement binary space partitioning (BSP) for level generation and consider the Genetic Algorithm as a content generator for a <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$2 D$</tex> dungeon game environment. The results of the test experiment demonstrate the effectiveness of this synergic approach, with BSP consistently achieving dynamic structural variation in all test scenarios and the genetic algorithm and Fisher-Yates shuffle showing optimal performance specifically in scenarios with a dungeon level size of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$75 \times 75$</tex> and <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$100 \times 100$</tex> with five, six, and seven content types, reaching 100 % variation in content distribution. The findings of this study can improve the procedural-level generator system that uses the BSP and genetic algorithm to develop dynamic content.