Procedural Content Generation

Martin Smed Christensen, Rune Bang Frederiksen, Michelle Dam Kristensen & Rasmus Wilhjelm Parelius

Studenteropgave: Fagmodulprojekt

Abstrakt

Procedural content generation has been a staple of video game development since the very beginning, with the game Rogue from 1980 (Lee 2014) being one of the best known and early games to use the technology. Procedural content generation has been used in video games for a range of purposes (Moss 2016) a significant one of these being terrain generation.
One of the game developer’s important tools for procedural generation is the noise algorithm. Rather than outputting uniformly distributed random numbers, noise algorithms are made to create natural looking gradients with recognisable but varying features. In this project, we put forward a vast amount of success criteria for what constitutes good game terrain. We then build a framework which uses an interface for noise algorithms, to create a tile-based 2d-terrain in which a player can drive around a small wagon to explore.
The success criteria are mostly met, although the question of whether the terrain is sufficiently natural looking is left unclear. Realistic terrain should be neither completely uniform, nor completely random, and using this definition, the framework generates realistic looking terrain from many different sets of input parameters, and using different noise algorithms.

UddannelserDatalogi, (Bachelor/kandidatuddannelse) Bachelor
SprogEngelsk
Udgivelsesdato29 maj 2017
Antal sider72
VejledereMorten Rhiger

Emneord

  • PCG
  • Procedural Generation
  • Terrain
  • Minecraft
  • Diablo
  • Noise
  • Perlin Noise
  • Improved Perlin Noise
  • Simplex Noise
  • random
  • hashing