Over two years ago, when Boston Public Schools decided it was time to completely redo its bus routing and timing operations, it also realized this overhaul was something staff members couldn’t figure out on their own, at least not optimally. That’s when the district decided to open the challenge to technology and academic leaders in what they called “a three-month hackathon-style initiative.”
The winner, a team from the Massachusetts Institute of Technology, was able to produce an algorithm that would do just what BPS wanted: improve the routing system from a cost, efficiency and community engagement perspective. The algorithm was devised by doctoral students Arthur Delarue and Sébastien Martin as part of a team with Professor Dimitris Bertsimas at MIT’s Operations Research Center.
BPS supplied a lot of optimized data, which MIT used to understand what areas of the busing system needed retooling. The school district then used that analysis to determine what kind of solution the hackathon’s participants should aim for.
How an MIT Team Helped Boston’s Schools Improve Its Bus System
The winning team ran the data they received from the school district along with data from Google Maps, students’ addresses, school start times and Boston traffic patterns, and produced an algorithm that would spit out the best routes, including pickup points. The algorithm strategically reconfigured bus stops, maximizing the number of students riding each bus and reducing the amount of time buses travel when no students are on board.
As a result, students’ walk-to-stop times were reduced, more students were able to take advantage of the bus system, some 50 underutilized routes were cut and alternate, more optimal routes introduced in their place. Plus, it saved BPS a cool $5 million. For a city whose school districts have the highest per-pupil school transportation costs in the country, this was nothing short of a miracle.
Two years later, the number of trips district buses have to make are down by almost 400 and the system is 20 percent more efficient, according to a report from nonprofit educational news site The 74.
“This is a huge win not only for the MIT team, but also for the City of Boston and our students,” said former BPS Superintendent Tommy Chang in a press statement. “This new model of routing buses will free up millions of dollars for reinvestment back into our schools. It will also reduce traffic and carbon dioxide emissions by decreasing the number of buses needed, helping improve our environment and ease congestion on our roadways,” he said.
Before the MIT team’s algorithm was devised, BPS transportation staff manually built school bus routes using what they considered an industry leader in pupil transportation software. Yet, it took several weeks to complete the routing exercise.
Georgia Tech Comes to Denver Public Schools’ Aid
A couple of years ago, Denver Public Schools found itself fielding requests from individual schools in the district to change their start schedule, according to a report from Education Dive.
Nicole Portee, executive director of transportation for DPS, wanted to see if she could accommodate those requests and at the same time maximize the utilization of the district’s school buses, which she felt was inefficient. Monica Villarreal, a DPS senior planning analyst, discussed the challenge with Pinar Keskinocak, her former professor at Georgia Tech’s School of Industrial and Systems Engineering.
Keskinocak connected Villarreal with doctoral student Amanda Chu, who began to work on redoing the district’s bus schedules. Chu’s goal was to make the schedules more adaptable to changing conditions and find a way to do this without having to manually adjust route assignments.
After months of work, Chu developed a Microsoft Excel–based decision-support tool that gives the school system a new, holistic picture of the bus fleet’s scheduling.
“It’s designed to present alternatives. It actually generates new bus route assignments, based on entered data, reflecting objectives and constraints. It also helps the school system quickly analyze bus route assignments,” Chu noted in an interview.
In the 2018-2019 school year, the use of the tool helped cut 34 routes and contributed to a 12 percent decrease in the number of running school buses, according to Chu’s data, as reported in Education Dive.
For their work, Chu, Keskinocak and DPS were named finalists for the 2019 Daniel H. Wagner Prize for Excellence in the Practice of Advanced Analytics and Operations Research.
A New Era in University-K–12 Partnerships?
The BPS and DPS examples show that technology solutions can be found for seemingly intractable problems — busing is just one example — at overburdened and underfunded public K–12 institutions. But to find answers, academia and K–12 administrators need to talk to each other.
“I didn’t know about the MIT Research and Operations Center, and I’m excited to think about what other problems we can work with them on for huge impacts,” Will Eger, BPS’ strategic projects manager, told The 74.
Such collaborations can also have positive spillover effects.
For instance, Dimitris Bertsimas, who led the MIT team that devised the algorithm for BPS, is now looking to expand the team’s work. He has launched Dynamic Ideas, a company that markets the algorithm to other school districts.
Chu’s tool as well is seeing signs of interest from other districts, including Atlanta Public Schools and the Tulsa (Okla.) Public Schools.