An Emotion Affected Action Selection Mechanism for Multiple Virtual Agents

This paper presents an action selection mechanism, for virtual flocking animals, inspired by ethology and theories of emotion. We investigate how emotions, in this case fear, affect behaviour, considering deer and sheep as exemplary flocking mammals. The paper also describes and applies a mechanism to measure different types of flocking, using singular values and entropy to compute complexity. Results show that emotion can be used to regulate two competing and vital behaviours like those seen in herding animals: the group behaviour (herding) and the individual behaviour (grazing), thus demonstrating that emotion is a functional organiser of group behaviour