Formal Agent-Based Modelling and Simulation of Crowd Behaviour in Emergency Evacuation Plans

Crowd behaviour deviates from normal when an emergency evacuation is needed. Thus, simulation of evacuation situations has been identified as an important tool for assessing design choices of urban areas, such as buildings, stadiums, etc., and Agent Based Modelling has been employed to tackle such problems. In this paper, we propose that formal modelling can rigorously define but also naturally lead to realistic simulations of such cases. Our main contribution is presenting how formal state-based methods, namely X-machines, can be employed to model agents in emergency evacuation plans. We also discuss the role of emotions, model artificial emotions that change the behaviour of agents under emergency situations, and provide a formalism that models the role of emotions and personality traits in order to create a more realistic scenario. Finally, we demonstrate how the developed formal models can be refined to code, a combination of Netlogo and Prolog in this case, that is able to simulate crowd behaviour with and without artificial emotions.