Skip to Content

A Model for Hurricane Evacuation

Software developed at MIT could save lives and money by improving hurricane planning.
September 10, 2008

As Hurricane Ike approaches the Gulf Coast, local authorities and emergency managers face critical decisions: when and how to evacuate residents, and when to draft in supplies and other aid. Proper planning could help save thousands of lives and hundreds of millions of dollars, but the unpredictable nature of any storm also makes effective decision making difficult. A researcher at MIT is now testing computer software that combines an unprecedented amount of data to help emergency managers make faster, more informed decisions.

Chaotic congestion: On August 30, traffic on Interstate 10 in Louisiana backed up as residents sought to escape Hurricane Gustav. Approximately two million people fled in two and a half days.

“Evacuation planning is very complicated,” says Ozlem Ergun, an associate professor and codirector of the Research Center for Humanitarian Logistics, at the Georgia Institute of Technology. “Given how bad the [2005 hurricane] Katrina process was, it is evident that there is a big need for this to be done in a systematic way.”

To coordinate a hurricane response, emergency managers have to rely on locally-drafted evacuation plans and guidelines provided by the Federal Emergency Management Agency (FEMA). After the chaotic and poorly planned evacuations carried out during the 2005 hurricane season, new planning procedures were introduced. FEMA also now uses a computer model to estimate how long it will take to clear a city based on its population, although this model can’t advise on when and how best to evacuate it.

These new measures appear to have had at least some impact. Last week, Hurricane Gustav forced most of south Louisiana to be evacuated–some two million people–in just two and a half days. The process went smoothly, but Ergun believes that this may be partly because the memory of Katrina is so fresh in so many people’s minds. “In five years, people won’t be so conscious about it and again may wait until the last minute to take action,” she says.

The new software might help the authorities avoid future catastrophes by alerting them to unforeseen problems. The software combines historical hurricane data, current weather conditions, and projected hurricane paths to help authorities work out where a hurricane will most likely strike and how intense it may be. “The model is more efficient in clearing people over time because it eliminates the confusion that leads to gridlock,” says Michael Metzger, who developed the model and is a PhD student in the Operations Research Center (ORC), at MIT.

The software is also flexible, allowing emergency managers to input their city’s demographics and geography. And it adopts a novel approach, categorizing a city’s population into different demographics, such as the elderly, tourists, hospital patients, and families with children, for separate evacuation. In addition, says Metzger, the model considers details of available evacuation routes. For example, as there is only one highway that leads out of Key West, FL, emergency managers would need to phase out evacuations to avoid congestion. This allows the software to provide more finely tuned recommendations, he says. The software can even advise emergency managers when to start bringing in supplies, where to set up shelter locations, and when to call in the National Guard.

Better methods: The new computer model requires a population to be grouped into different key demographics. It estimates the cost of evacuating each group and the amount of lead time required.

The new software was built using dynamic programming–a technique often employed to solve problems involving uncertainty and parameters that change over time. “It evaluates all the different, possible future outcomes of a decision and then works backwards to make a decision at current time,” says Metzger. Richard Larson, a professor of engineering systems and civil and environmental engineering at MIT and Metzger’s advisor, says that the technique is well proven. For example, similar approaches are used by American Airlines to determine seat pricing and by football coaches to make real-time strategy decisions.

As the system has not yet been tested on a real evacuation scenario, Metzger says that it is difficult to estimate how much time or money it could save. However, Ergun describes the technology as “very impressive.”

Once the data is available, Metzger’s model will be tested against the decisions made by officials during the 2008 hurricane season. Larson says that this will be completed by May 2009, and he believes that the model could be used initially as a training tool. “Just as pilots use flight simulators, if decision makers can go through different scenarios, they are going to be better at it when the real thing happens.”

Ike is expected to make landfall this weekend near Corpus Christi, TX. Texas governor Rick Perry has put 7,500 National Guard members on standby, and Louisiana governor Bobby Jindal has told residents to start stocking up on supplies as shelters and evacuation transportation is readied.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.