Smart-Cities-Artwork-824x549

Green light: Algorithms that use real-time data to change outcomes in complex systems

July 11, 2014
Physics & Mathematics

erything is getting smarter: phones, homes, meters, cars and soon perhaps, entire cities. It’s already begun: In Los Angeles, when traffic peaks and city roads become backlogged, the timing of the 4,100 synchronized traffic lights changes to help ease some of the congestion.

This iteration of smart technology interests USC Viterbi Professor Ketan Savla of the Sonny Astani Department of Civil and Environmental Engineering at the USC Viterbi School of Engineering. He develops algorithms that can use real-time data to change outcomes in complex systems.

“That’s where my research is: between information gathering and decision-making,” said Savla, who currently focuses on infrastructure networks: transportation, smart grids, water networks and the coupling between them.

What this means is that someday when you’re stuck on La Brea inching from red light to red light, pulling out your hair just trying to get to your destination, Savla’s algorithms could save the day. Stop lights could respond to the traffic, having green lights that last longer and letting more cars proceed through the intersection, while smaller streets would get less frequent and shorter green lights, giving priority to the busiest traffic flows.

Less time spent idling in traffic isn’t just good for your mood and car; it’s better for the air quality in Los Angeles and the environment as a whole.

Easing congestion around town

In 2013, Savla gave a presentation to city officials to let them know that he plans to optimize the way the city of Los Angeles’ Department of Transportation software works to ease traffic congestion on city streets. The officials gave him meaningful feedback about what they’re looking for and a taste of some of the on-the-ground realities he needs to incorporate. He’s been designing software and running simulations since then. They also gave him a letter of support that he submits along with proposals for funding, which, he said, has been very useful.

Fortunately, the city currently uses software designed and maintained in-house, which leaves the possibility of translating research from theory to practice. Other cities purchase off-the-shelf software packages that cannot be easily altered.

“Traffic signal control is emerging as an important industry,” Savla said.

Working on an algorithm that could be used to help the city in the immediate future is a rare and promising outcome of his research, but Savla balances the short-term and long-term benefits of his work.

“As an academic, yes, I want applications,” Savla said, “but I’m thinking more long-term by developing basic models and theory for doing this not only in Los Angeles, but anywhere in the world.”

Hoping to take a big step forward

Savla soon plans to give a presentation to the Department of Transportation on cases studies and data comparing the performance of his algorithms to existing ones, and identify opportunities to incorporate his software into the department’s existing system. This would be a huge step for the controls and optimization branch of engineering.

“If this becomes a successful project, I think this will be a good advertisement for the kind of work that not only I do, but that my community does,” he said. “It will be a good showcase of the fact that we’re not doing theory in isolation, but that there are practical effects.”

And, in the future, when Savla turns his optimization focus to other city infrastructure networks, he’ll have a track record of success that can sway decision-makers.

“If you have one success story,” he said, “then if I move these ideas into smart grids or water networks, I think it becomes easier to convince the appropriate people to give you a chance.”


Green light: Algorithms that use real time data to change outcomes in complex systems

No comments yet.

Leave a Reply

* Copy This Password *

* Type Or Paste Password Here *