Below is a link to an interesting article in The Economist about about modeling crowd dynamics on the behavior of molecules (a common model) and on the behavior of people (the innovation addressed in the article). Turns out there is an advantage to the use of both models.
The article was sent to me by Kenneth Ingraham, the president of Interactor Simulation Systems. Ken also has an interesting idea to address the hazards of crowd behavior disasters by using physical connection. While his idea might at first seem counter-intuitive, it also might be a successful one. I wonder how one might test such an idea?