Procedural Generation of Quests for Games Using Genetic Algorithms and Automated Planning

Abstract

The production of high-quality commercial games requires the work of a few hundred individuals, including designers, artists, and story writers, to produce game content, such as 3D models, textures, and narratives. Over the last decade, the production of game content has grown to the point of becoming a bottleneck in companies' schedules and budgets. In this context, procedural content generation techniques are increasingly being applied to reduce the work overload of the development teams. Although game developers and academic researchers have extensively explored procedural content generation, there is a lack of techniques to handle procedural generation of quests. In this paper, we present a new quest generation method based on genetic algorithms and automated planning. By combining planning with an evolutionary search strategy guided by story arcs, the proposed method can generate coherent quests based on a specific narrative structure. Preliminary results show that quests created with our method are nearly at par with those created by game design professionals.