Ants the key to modelling crowd behaviour

Innovative research at the Institute of Transport Studies at Monash University has provided new insights into the behaviour of crowds in emergency situations. The research team, led by Dr. Majid Sarvi, adopted a unique approach, using Argentine ants to model the behaviour of human crowds.

In the past, there have been a number of incidents where overcrowding and panicking patrons has led to injuries and deaths at public events. These events, and the need to better manage crowds during emergencies, provided an impetus for Dr. Sarvi’s research.

Due to the lack of real-life, unbiased data to test crowd behaviour models, Dr. Sarvi’s team, in collaboration with Dr. Martin Burd from the School of Biological Sciences, was required to find an alternate method to test their computerised models. 

The solution came in the form of Argentine ants. It has been observed in the past that the behaviour of ants when escaping under panic is comparable to the behaviour of human crowds in similar situations. Further observations had found that ants on a trail can be compared to pedestrian movement.

Dr. Sarvi’s team identified through their experiments that the key factors that affected movement were the structural elements present within an area, and the socio-psychological factors associated with the group and each individual.

In each experiment, a structural factor was changed. When a column was placed at an exit, a reduction in evacuation time of 41%, on average, was observed. When extra exits were introduced, Dr. Sarvi and his colleagues found that certain exits were disproportionately favoured. For instance, with four exits in place, it could be seen that the majority of the ants only utilised two of those exits.

Further observations found that the tendency to follow others and the interactions between individual ants influenced the manner and speed of evacuation.

The results of this preliminary investigation give rise to a number of possibilities. It has demonstrated that the effectiveness of introducing columns, partitions and walls into public areas can first be tested using computerised models, and then validated by conducting experiments using ants. Dr. Sarvi’s research has also prompted the need for a greater emphasis on socio-psychological factors when considering crowd behaviour.

It is certain that Dr. Sarvi’s and his team’s research will play a key role in helping infrastructure planners establish measures to ensure the safety of crowds in future.