New Scientist: Virtual rioters predict crowd movement
November 16, 2011
University of Maryland geographer Dr. Paul Torrens is featured in New Scientist for his work predicting how crowds move in a riot:
Based on how people interact with their environment, the model incorporates about 30 different behaviours, from processes like path planning and collision avoidance to group dynamics such as herding and following a leader. Concepts are translated into mathematical models by applying statistics or using methods like game theory. Different algorithms can be applied to each agent, or to the group as a whole, to see how they affect overall behaviour.
Note that he doesn’t think he can predict a riot based on this approach – he’s using a virtual riot to study how crowds move in a certain geographical space. He thinks this work could help in guiding disaster relief efforts and emergency planning, for example.
There’s a point in the not-so-distant future where I can see computational and methodological advances allowing agent-based modeling to feed off of survey-based microsimulation projections to create ‘deep’ agents in a model that can behave even more realistically.
I suspect this type of simulation would also be useful to savvy law enforcement types designing more efficient crowd control measures that prevent crowds from gathering in ways that would fuel a riot or violent behavior. And savvy organizers of marches, demonstrations and yes, even riots, could one use this work to maximize a crowd’s appearance and effect. For those non-violent types trying to maximize media coverage, induce social epidemics, cause tipping points and so on, this kind of work could greatly increase their return on investment. Think of the organization of Susan G. Komen’s Race for the Cure combined with a political demonstration and using these simulations to plan how to manufacture effect of people on the ground. It will make today’s political street theater seem amateur by comparison. The nonviolent terrorists in my novel would exploit these techniques to win over the media and political class.
On a side note, this work is just another example of how social science benefits from other fields’ tools and expertise. For another example of cool work by Dr. Torrens, see his Geosimulation website, where he is working on, among many other things, accelerating agent-based modeling to better study complex systems, if I am understanding this correctly. Pencil me in as a fan of Dr. Torrens (and not only because he uses Lego mini-figures in some of his computer animations).