Smart cities researchers simulate human behavior during floods

Mon, 2013-06-03 15:35 -- Jesse Berst

Many of the most exciting smart city possibilities relate to predictive analytics -- crunching large amounts of data to forecast what will happen next. But there's always been a "blind spot" -- namely, how to predict and simulate human decisions, especially during unexpected, unusual events.

Now RMIT University, a member of the Smart Cities Council Advisory Board, has made real progress on this front. As you will read, they applied this breakthrough to understanding what might happen in emergency situations such as floods. That's an important area, since climate change means that many of the planet's cities will be facing extreme weather events more often.

But the core insights apply to almost any part of a smart city. You'll want to read how. - Jesse Berst, Smart Cities Council Chairman

The Global Cities Research Institute at Melbourne, Australia-based RMIT University has released a research report detailing how computer simulation and interactive games could help communities and government with climate change adaptation and emergency management.

The report - Exploring the Adaptive Capacity of Emergency Management Using Agent Based Modelling (pdf)  - follows a year-long research study to develop computer models to simulate how people behaved during flash flooding in the Melbourne suburb of Elwood, which is prone to floods.

"Elwood has had two one-in-a-hundred year flood events in seven years, causing millions of dollars in damage," explained Professor Lin Padgham, the project lead.

And while the focus was on Elwood residents and specifically the use of sandbagging to protect against flash flooding, Padgham believes the potential for a wider use of the technology is huge.

"Extreme weather events are increasing so there is a greater urgency for us to be prepared and to respond effectively," she said.

So how does it work?

Using behavioral data gleaned from interviews with local residents, the RMIT team developed agent-based modelling, described as technology that simulates the behavior of people (or agents) – and then replicates the behavior across a sample population to demonstrate responses to various scenarios, in the case of Elwood flash floods.

"Our long-term vision is to work closely with social scientists and end users to create support tools, methodology and frameworks to make it much easier to use simulation technology," Padgham said. "This will allow both governments and local communities to use simulations to analyse complex issues to inform planning, policy and local strategies."

Padgham added that this work is part of a long-term research goal to combine simulation technology and artificial intelligence to address major world issues.


Get Smart Cities Now, our weekly newsletter highlighting smart city trends, technologies and techniques.