Analysis of Procedural Content Generation and Stylisation Techniques in Developing Video Games

Abstract

This paper explores the evolving role of Procedural Content Generation (PCG) technology in developing video games, highlighting its importance in enabling automatic generation of scalable, dynamic, and visually distinct gameplay experiences. This research aims to provide an overview of the current methods of procedural terrain generation and assess how emerging generative artificial intelligence (AI) technologies can address their limitations. Specifically, the study reviews methods, including Perlin noise, fractal noise, voxel systems, physical simulation, and generative adversarial networks (GANs). Exemplary applications of traditional and modern PCG systems, including game applications, are analyzed. Experimental observations reveal that while PCG effectively generates structural diversity, generative AI models significantly improve context awareness and content richness, offering potential solutions to historical limitations of PCG. These conclusions indicate future developments where intelligent PCG frameworks enhance artistic control and environmental narrative depth. Integrating generative AI into PCG processes elevates asset creation and introduces a new text-driven and context-aware content generation paradigm.