Motivation Discrepancies for Rebel Agents: Towards a Framework for Case-based Goal-Driven Autonomy for Character Believability

Character believability is one of the key requirements of a successful narrative. In order to be believable, characters must appear to act in accordance with personal motivation, which is shaped through events occurring throughout the narrative. Goal-Driven Autonomy (GDA), is a model of goal reasoning in which agents identify, reason about, and, if necessary, modify goals that they will pursue. Our proposal for using GDA in character-centric interactive narratives is based on the new concept of “rebel agents” and on a set of new discrepancies, which we call “motivation discrepancies”. A rebel agent is an agent that may “refuse” a goal, plan, or plan component that it assesses to be in a conflict with its own motivation. Previously, in systems performing planning and execution, including GDA agents, a plan assigned to an agent was executed unless exterior conditions made it unfeasible or no longer useful to do so. In the case of rebel agents, plans may fail to be executed and new goals may be chosen due to the agent’s own “refusal”, based on the agent’s internal motivation. Motivation discrepancies cover a wider range of situations than the traditional GDA discrepancies, and are suited specifically to character-believability purposes.