RESEARCH NEWS

Computer models help with riot control

25 May 2007



The randomness of riots is also what makes them so dangerous. Now, a geographer at Arizona State University is designing a computer model to help predict human behaviour at its most unpredictable.

The model will be used to simulate mobs, evacuations and natural disasters to assist the planning of cities and shopping centres. It should also indicate the best way to diffuse disturbances during a crisis, without the use of force.

The researcher, Paul M Torrens, says: ‘It is impractical to establish live experiments to reproduce mob behaviour during riots for the purposes of academic experimentation. You couldn’t stage a realistic rehearsal of an evacuation because people are not going to panic appropriately.’

Torrens believes current models do not fully account for the individuality of the crowd members. The new model will incorporate factors such as age, sex, size, and health, together with more emotional factors such as body language and the levels of panic people are feeling.

The model will not just be used to predict states of fear, however. It could also be used to predict the transmission of disease, and help governments find more persuasive ways of discouraging the use of cars.

It is very extensive, recording the state of every person 60 times a second. It has already been used to model a crowd’s reaction to fire in densely clustered buildings, with realistic results.

‘When trying to evacuate, people start to run and panic,’ says Torrens. ‘Jams will occur and the evacuation doesn’t proceed as efficiently as it might otherwise.’

His research relies heavily on the National Science Foundation Career Award. The $400,000 award is very prestigious, and rarely given to geographers.

Related internet links

Arizona State University