Sunday, October 20, 2013

Putting the person in to the particle


Over the last decade physicists have developed “social force” models to describe the movement of individuals in crowds and gatherings. These models assume rather simple repulsion/attraction interaction forces to explain how people behave in these situations. They may be useful in, for example, designing escape exits at football stadiums and planning for mass events.

As useful as it can be to assume that people are particles, most of know that we are a bit more complicated than that. Nanda Wijermans’ presentation on Friday the 18th of October at the Futures Institute, addressed this issue head on. She has set up a general framework for modeling crowds at, for example, concerts and bars. In her model, each individual has a set of rules governing how they respond to their memory and their physiological state. For example, if you have drunk a lot you might need to go to the toilet, or if you have lost your friends you might want to find them. These are all rules we can relate to from our own lives. These types of behavioural rules can reproduce some collective outcomes that we wouldn’t see in particles, such as people wandering off towards the toilet not because they need to wee but because they aimlessly follow a friend.

I can see how this could be a powerful tool for pub or bar designers who could run simulations to decide how to arrange the social space. Nanda herself is more interested in general questions about spontaneous emergence of social groupings. Much of the lively discussion during and after the seminar was about exactly what one can and cannot expect to achieve with more detailed models. In my view, if you are going to make a model more complicated then you need to address questions that can't be tackled with simpler models. There are certainly plenty such problems!

Nanda also presented her current research work at the Stockholm Resilience Centre, where she is now a Postdoc. Here she is applying agent-based models to questions of resource use and co-operation in irrigation systems in Bali. Her overall goal is to include realistic behaviour in modelling of our interactions with our environment.

No comments:

Post a Comment