Automating Dungeon Level Design using Search-based Procedural Generation

The adoption of procedural content generation in electronic games is increasing, allowing for the quick and cost-effective creation of game artifacts like music, scenarios, and art, which aids or replaces manual content creation. Metaheuristics, such as Genetic Algorithms are effective in solving complex problems by exploring the solution space. This study implemented and analyzed this algorithm for generating procedural content in dungeon games, focusing on scenarios and enemies. Experimental evaluations demonstrated efficient optimization of room layouts and enemy distributions, with smaller room sizes producing better fitness results.