Multi-Level Evolution of Shooter Levels

This paper introduces a search-based generative pro-cess for first person shooter levels. Genetic algorithms evolve the level’s architecture and the placement of powerups and player spawnpoints, generating levels with one floor or two floors. The evaluation of gen-erated levels combines metrics collected from simu-lations of artificial agents competing in the level and theory-based heuristics targeting general level design patterns. Both simulation-based and theory-driven eval-uations target player balance and exploration, while re-sulting levels emergently exhibit several popular design patters of shooter levels. Procedural content generation (PCG) has often been used in the game industry to create single-player experience