Evolutionary Behavioral Design of Non-Player Characters in a FPS Video Game Through Particle Swarm Optimization

Evolutionary computation covers the family of artificial intelligence techniques inspired by nature and biological evolution. These methods, such as swarm intelligence, may have a very positive impact on video games, for instance, for the design of Non-Player Characters (NPCs) to obtain a realistic intelligent behavior in a simple way. To this aim, we describe an evolutionary behavioral design of NPCs using particle swarm optimization in a first-person shooter video game. Several computer experiments have been carried out to analyze the feasibility and performance of this approach. Our experimental results show that the proposed method performs very well and can be successfully used in a fully automatic (i.e., without any human player) and efficient way.