New Computer Model Predicts Crowd Behavior

Arizona State University assistant professor in the School of Geographical Sciences Paul M. Torrens is developing a computer model that will realistically simulate the reaction and behaviors of people in a crowd.

The computer model could be used by city planners, shopping center developers, and public safety and health officials to simulate situations that would be impossible to create in a live experiment, such as the evacuation of a city or large building.

Torrens' goal is to create a simulation program that accounts for the panicked and desperate state that people would feel under such situations. The current behavior modeling programs have not proven to have the veracity this model could have, according to Torrens.

In Torrens' model, each simulated person will behave independently and have different characteristics, such as age, sex, size, health, and body language. The program will also account for crowd and environmental features such as group panic and safety levels.

Torrens said the model will be used for realistic experiments exploring "what if" and unforeseen scenarios that could affect cities. Additionally, the model can be used to explore sustainability in downtown settings, such as how a city can promote walking instead of driving and how pedestrian flow can fit better with city traffic. The spread of a pathogen through a city could also be simulated.

The completed prototype model collects data from each element in the simulation every 60th of a second.

Torrens' research is funded by a $400,000, five-year National Science Foundation CAREER Award.
Abstracted by the National Law Enforcement and Corrections Technology Center (NLECTC) from Arizona State University (05/22/07).