A Belief-Desire-Intention Model for Narrative Generation

Abstract

Narrative AI needs to model more than just action. In this paper, we investigate the Belief-Desire-Intention (BDI) agent architecture to allow plots to be modelled in terms of character motivation. This allows authors to focus on elements of the character model which are highly relevant to plot. We describe an extended implementation of the ConGolog agent programming language which includes BDI syntax and semantics. Using this language, we provide an example of how plot could be advantageously modelled in terms of character motivation.