Optimizing school bus routes with an algorithm reduces costs by over 500 million yen
by
You may have seen yellow school buses in movies and TV shows set in the United States. This school bus first appeared in 1939, but since then, it has become increasingly complex as students, schools, and road systems have changed, leading to problems such as excessive costs and poor performance. This problem has been solved by an algorithm, and it is now a hot topic.
How One City Saved $5 Million by Routing School Buses with an Algorithm - Route Fifty
https://www.routefifty.com/tech-data/2019/08/boston-school-bus-routes/159113/
Parents of children attending Boston public schools have several options for which school their children will attend. This is intended to eliminate inequality, but because Boston has more schools than other areas, parents have more choices, which inevitably leads to more complicated bus routes.
Another factor that makes the problem more difficult is that Boston has a wider range of school bus service than other cities. Some schools have students in as many as 20 different zip codes. Furthermore, schools have different start times, with some starting as early as 7:15 a.m. and others as early as 9:30 a.m., further complicating bus route planning.
In 2017, transportation costs per student in Boston public schools reached $2,000 per year, accounting for 10% of the school district's budget.
by
Unable to find a solution, Boston Public Schools decided to hold a competition for researchers to determine whether anonymized data could be used to determine optimal routes and start times for students in the Boston school district.
According to Will Eger, project manager for Boston Public Schools, 'the transportation landscape in our district is full of quirks.' These quirks include road width, bus infrastructure issues such as whether or not the bus has a wheelchair lift or safety seat, whether or not the bus has a chaperone, and whether or not the bus needs the same driver every year. There are also approximately 5,000 children who require door-to-door transportation, and this request needed to be met.
Arthur Delarue of the MIT Operations Research Center , winner of the competition, explained that there were countless ways to solve the problem, and that solving it took hundreds of hours. Previously, bus routing required thousands of hours of work by 10 route planners. However, MIT's algorithm creates a complete route map in one go, providing a foundation for human workers to refine. 'The work of route managers, who interact with drivers, principals, parents, and students, is important and irreplaceable,' Delarue said. 'But the order of stops and route design are not something humans can solve efficiently. That's the added value we provide.'
When Boston first implemented the algorithm between 2017 and 2018, the results were dramatic. The algorithm generated route maps that were 20% more efficient than manual maps in 30 minutes, and the longer the algorithm was run, the more efficient the maps became. Using the algorithm in the summer of 2017 reduced the number of buses used by 50, resulting in an annual reduction of 1 million miles (approximately 1.6 million kilometers) and a reduction in carbon dioxide emissions by 20,000 pounds (approximately 9,000 kilograms) per day. As a result, the school district was able to use the savings of $5 million (approximately 525 million yen) to invest in classroom equipment.
by
Eger explained that the new routes didn't require students to take a different bus or walk any longer, and expressed surprise at the algorithm's solution: 'We were able to reduce the total number of stops while shortening the walking time for students living in risky areas.'
Then, in December 2017, based on the algorithm's recommendations, the Boston School Committee approved the first school start time change in 30 years, affecting 85% of Boston's schools.
The approval of the start time change follows research showing that setting school start times earlier than 8 a.m. has negative effects on teenage students. An MIT study also found that while many parents in Boston prefer later start times, it's students with wealthier parents who actually attend schools with later start times.
However, the change in school start times did not go ahead. The proposed change would have been to make the start times later for older students and earlier for younger students, but this was opposed by parents who felt that a major change in school start times would have an impact on family schedules and increase the burden of childcare.
by stem.T4L
While no drastic changes to school start times have been made, Eger emphasized that the project has had a positive impact on the city as a whole. 'This is one example of how we've taken advantage of the great research potential that Boston offers,' he said. 'Many school districts have similar issues, but these challenges are outside our scope of expertise. We should rely on the experts and focus on doing our own work best.'
Related Posts:
