Experiments with Emotion Contagion in Emergency Evacuation Simulation

Abstract

Multi-agent systems simulation is used to predict human behaviour in emergency evacuation cases. However, as human behaviour can change under the effect of emotions, it is essential to create models of artificial agents and simulations that mimic such behaviour in order to make prediction of the overall system performance. In emotional agents, the role of emotional contagion is important. Emotional contagion is a result of interaction between agents which could affect each others emotions. It is the case that in emergency situations, emotions (especially calmness, fear and panic) may propagate in various ways, depending on the agents personality type as well as other factors. In this paper, we review various methods of emotional contagion. In order to develop emotional agent simulation, we start from a formal state-based modelling method and devise a number of variations of known emotional contagion methods. NetLogo visual simulation is used, in which a number of experiments is conducted. The results are useful to demonstrate different behaviour of different emotional contagion models in the evacuation of an open square area.