Development of Virtual Agents with a Theory of Emotion Regulation

In order to endow virtual agents with more realistic affective behavior, it is important to provide them not only the capability to generate and regulate emotions, but also the ability to reason about the emotion regulation processes of other agents. To this end, this paper introduces a computational model for a theory of emotion regulation (ToER). The model has been implemented and tested using the modeling language LEADSTO. In addition, a virtual environment application has been developed, which is inhabited by agents that are equipped with the model for ToER. A first evaluation indicates that the model indeed enables the agents to show more realistic affective behavior.