An Interactive Multi-Agent System for Game Design

Abstract

This paper presents a novel approach to procedural generation of game maps for multi-player, competitive video games. A multi-agent evolutionary system is employed to place streets, buildings and other items, resulting in a playable video game map. The system utilises computational agents that act in conjunction with the human designer to produce maps that exhibit desirable characteristics. This paper compares the impact that the additional agents have in terms of the quality of candidate solutions. The results indicate that the use of the agents produces higher quality solutions in comparison to a traditional interactive genetic algorithm.