Hybrid Approaches to Procedural Content Generation for Game Design, Production, and Security

Babin, Mathias

Hybrid Approaches to Procedural Content Generation for Game Design, Production, and Security

Files

babin_mathias_phd_2025_thesis.pdf (10.02 MB)

Date

2025-08-07

Authors

Babin, Mathias

Abstract

Procedural content generation (PCG) is the process of automatically generating game assets through algorithmic means. PCG enables developers to produce scalable and diverse content while significantly reducing their manual workload. Beyond its efficiency, PCG offers additional benefits, such as data compression, or playing a vital role in certain game genres, such as Rogue-likes, where new environments are required for every play-through.

This dissertation explores hybrid approaches to PCG as a means to improve its applicability and effectiveness. We demonstrate how methodologies from various domains of Computer Science can contribute to the procedural generation of video game content. By hybrid, we refer to advancing existing PCG techniques through the integration of AI models and extending PCG's applications into previously unexplored domains, particularly in the areas of game security and game balance.

We present solutions to several gaps in current research, along with the foundational background and related studies relevant to each. Specifically, we discuss four published works contributing to game design (automated game balancing), production (level generation), and game security (anti-cheat for competitive online games).

Summary for Lay Audience

Video games often require vast amounts of content, such as textures, music, levels, and characters, which traditionally take a lot of time and effort for developers to create. Procedural Content Generation (PCG) is a technique that automates this process using algorithms to allow games to generate content on their own. This not only saves developers time but also makes games more dynamic and unpredictable. This research explores ways to improve PCG by combining it with artificial intelligence (AI) and applying it to new areas beyond just content creation. Our goal is to expand PCG beyond level generation into new fields like game security and balancing. The dissertation covers three key contributions: 1) Level Generation: Developing better ways to create engaging and varied game levels. 2) Game Security: Using PCG techniques to help prevent cheating in competitive online games. 3) Game Balance: Generate new game elements that help ensure fair and enjoyable gameplay. By addressing these areas, this research aims to make PCG more practical for developers while simultaneously improving the player experience. It highlights how AI can be combined with traditional PCG methods to push the boundaries of game production, design and security.

Keywords

Procedural Content Generation, Artificial Intelligence, Deep Learning, Reinforcement Learning, Level Generation, Game Security, Automated Game Balance