Design and Evaluation of a Procedurally Generated Dungeon Game
Arthur R. Pinheiro So, Alinne C. Corrêa Souza, Lincoln Magalhães Costa, Rafael Gomes Mantovani, Francisco Carlos M. Souza · 2024
Game development involves creating various content, from artistic tasks to technical activities. Procedural Content Generation (PCG) and metaheuristics like Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) automate and optimize this process. This paper presents Dungeoncide, a game using GA and PSO for procedural content generation. Dungeoncide features a medieval knight battling monsters in procedural generated dungeons. The primary contributions include the development process and player feedback on the generated content’s quality and feasibility. Experiments showed that while generated levels were well-received, they were less challenging and immersive than manual ones. Future work will refine the algorithms to improve balance, aesthetics, and engagement.