A Multi Agent Approach to Analyse Shift in People Behaviour Under Critical Conditions

The aim of this study is to illustrate the possibilities offered by using a multi agent approach – agent based modelling (ABM) – to deepen the understanding of behavioural phenomena possibly arising in a multitude of people exposed to a sudden variation in environmental conditions. The environment, people are in, is one typical exhibition hall that is part of a larger interconnected space, such as a mall or an art centre. The normal condition is characterised by the hall with the exhibition of artworks, with people inside to enjoy the performance. Suddenly an alarm is triggered, thus defining the onset of the critical condition, activating the sequence of emergency operations. Thus, there is a change in the behaviour of the people inside the hall: from enjoying the exhibition to finding their way out.

This study tries to figure out some behavioural patterns that may appear to be present among the people in real life situations similar to the one considered here. Once defined the basic quantitative assumptions in terms of spatial domain and number of people involved, a proper toolkit is used to manage the qualitative issues, such as environment modifications and people characteristics. The open source ABM platform NetLogo is adopted for modelling purposes and capturing the resulting behavioural patterns.