Self-Validated Behaviour Trees through Reflective Components

Developing the AI for non-player characters in a video game is a collaborative task between programmers and designers. Most of the times, there is a tension between the freedom that designers require to include their narrative in the game and the effort required from programmers to debug faulty AI specified by good story tellers who are not programmers. In this paper is presented an architecture for building the AI of an NPC that extends the component-based approach, which represents the functionality of an entity as a collection of functionality-specific components. By associating an action in a behaviour tree with a collection of components, and equipping those components with some reflection capabilities, we are able to identify faulty behaviour trees at design