Using BDI to Model Players Behaviour in an Interactive Fiction Game

摘要

Player Modelling is one of the challenges in Interactive Narratives (INs), where a precise representation of the players mental state is needed to provide a personalised experience. However, how to represent the interaction of the player with the game to make the appropriate decision in the story is still an open question. In this paper, we aim to bridge this gap identifying the information needed to capture the behaviour of players in an IN using the Belief-Desire-Intention (BDI) model of agency. We present a BDI design method to mimic the interaction of players with a simplified version of the Interactive Fiction Anchor-head. Our preliminary results show that a BDI agent based on our player model is able to generate game traces more similar to humans than optimal traces.