Extending Recursive Backtracking for Procedural Generation of Interconnected Rooms and Staircases in Multi-Level 3D Dungeon Layouts

Main Article Content

Zhen Shern Soh
Kah Chun Chong
Bee Sian Tan
Jia Hui Ong
Chin Hui Ooi
Kim Soon Chong

Abstract

Procedural Content Generation (PCG) is a powerful tool in video game development, enabling the automated creation of diverse and replayable game environments. This paper presents a novel extension of the recursive backtracking algorithm, which adapted to procedurally generate interconnected rooms and staircases across multiple vertical floors in 3D dungeon layouts. By using parameters such and maximum room chain lengths, staircase probability and directional constraints, the system generates coherent, non-repetitive and fully traversable environments. Through 13 test samples, the API demonstrates its ability to produce varied and scalable dungeons, including multi-level environments, showcasing its versatility and adaptability for diverse game design needs. Performance testing further reveals the API’s efficiency, with low CPU and GPU demands, optimized memory usage, and high frame rates ensuring smooth and visually responsive gameplay. These results highlight the API's ability to deliver high-quality content generation while maintaining system stability, making it suitable for use on mid-range hardware. The implementation offers game developers a robust and flexible tool for creating dynamic and engaging game experiences with minimal resource overhead.

Article Details

How to Cite
Soh, Z. S. ., Chong, K. C. ., Tan, B. S. ., Ong, J. H. ., Ooi, C. H. ., & Chong, K. S. . (2025). Extending Recursive Backtracking for Procedural Generation of Interconnected Rooms and Staircases in Multi-Level 3D Dungeon Layouts. International Journal of Creative Multimedia, 6(2), 23–47. https://doi.org/10.33093/ijcm.2025.6.2.2
Section
Articles