Clients in the News: Schools pursue better directions
From District Administration
Written by Brian Nadel
April 13, 2018
Quick quiz: After the buildings, salaries and benefits, what is one of the biggest expenses for most school districts?
Chances are that it’s those ubiquitous yellow buses that pick up kids in the morning and return them home (or close to it) after school.
Close scrutiny, experience and a dose of cutting-edge technology can control this expense, allowing more money to go to instruction.
Administrators in the know streamline the process of getting students to and from school by taking into account traffic patterns and other potential delays.
They have also staggered opening times for their schools. And in the future, routes might even be reshuffled in real time to further improve each bus’ efficiency.
The techniques at the administrator’s disposal range from a variety of off-the shelf routing programs to relying more on an administrator’s experience with local conditions. The best approach might lie in a combination of the two.
Roll in the supercomputers
At one extreme is Boston Public Schools. On any given day, about 25,000 students are bused to 230 public, parochial and special ed schools. It costs about 10 percent of the district’s $1.1 billion annual budget.
The district recently optimized its routes using state-of-the-art technology and what mathematicians call the “classic salesman problem,” which attempts to find the most efficient way for a traveler to visit a number of cities.
“Boston Public Schools busing is much harder,” says Will Eger, strategic project manager in the district’s finance department. “Instead of a single traveler, we have 25,000 children to get to the right school on time every day. It’s a daunting challenge.”
Enter the Massachusetts Institute of Technology’s Operations Research Center, located just across the Charles River from district headquarters. To model the busing problem, MIT’s software mapped new bus routes by analyzing every rider’s location and destination, along with historic traffic patterns.
Along the way, the software looked at an astronomical number of route combinations—on the order of 300 followed by 60 zeros. (That’s more than the number of hydrogen atoms in the sun.)
Using MIT’s supercomputer network, the problem was solved in 30 minutes. The result is 650 discrete bus routes, and a reduction of Boston schools’ fleet by 50 vehicles. The savings is between $3 million and $5 million for the 2018 academic year.
The changes resulted in a more efficient transportation network as well, with an estimated 1 million fewer miles traveled and 20,000 fewer pounds of carbon dioxide emitted. “The next step is to rewrite the software to run on a laptop so every district can use it,” Eger says.
Not every school district, of course, has access to world-class engineering expertise and supercomputers to route their buses. But logical actions can still be determined.
Making speedier adjustments
Richard Gallagher, the transportation director at Bay Shore Union Free School District on New York’s Long Island, uses a combination of software and local knowledge.
He tracks routes with software from Transfinder, a company that provides districts with bus routing and related services, and he also relies on his experience to make tweaks for efficiency.
Each summer when Gallagher creates the district’s bus routes, he also accounts for expected changes during the year by leaving a few empty seats for new students who move into neighborhoods with more transient populations.
He even takes into account the Long Island Railroad commuter train schedules, since the town is bisected by grade-level train crossings.
“We do pretty well at reducing costs,” he says. “Any district not looking closely at its buses and routes is leaving money on the table that is better spent on teaching.”
To control costs, districts should try to pick up more students at each stop, adds Peter Lawrence, a busing consultant and transportation director of the Fairport Central School District in New York. This forces administrators to determine the maximum duration of a bus run.
Fairport, for example, caps routes at 45 minutes to an hour. Rural districts, however, may have runs that exceed two hours.
“A bus run’s efficiency is measured in time and distance,” says Lawrence, who is also his area’s representative for the National Association for Pupil Transportation. “If you’re making a lot of stops along the way, times costs you money.”
Districts should also ensure drivers follow policies (such as those that bar unscheduled “courtesy” stops); and try to reduce “deadhead miles” when buses travel empty, Lawrence says.
“Figure out, if you drop off students at a school, can you pick up students at a nearby school for the next run?” he says. “It can depend on geography. City and suburban districts have it easier—it’s harder for a rural district.”
Districts should always include civil transportation leaders when planning any new program that requires busing students. Otherwise, leaders may win community support without recognizing they also now face an increased transportation bill, Lawrence says.
“They may study the benefits of the program, and say, ‘Yeah, this is super, it’s phenomenal,’” he adds. “Then they ask transportation, and we say, ‘Yeah, we can do that, but we’ll probably need another dozen busses,’ and the district feels like it wasted its time on the idea.”
Boston can also optimize routes by rerunning MIT’s software if the student body or traffic patterns change. Eger looks forward to a time when the technology can adjust to daily incidents, such as an accident slowing traffic or a broken traffic light that might delay pickups and school arrivals.
“Our ultimate goal is adaptive routing, where we could alter the routes of nearby buses to compensate and still get everyone to school on time,” he says.
In the future, GPS sensors on a bus stuck in a traffic jam could alert headquarters. The bus would then be automatically rerouted via text messages to the driver or an interactive street map.
The result might be a more circuitous route, but one that gets students to school before the bell rings—because the hardest part of the day is not a late start, but the fourth-period algebra test.