The students in the USA are eligible for school bus transportation if they live beyond a particular radius of the school. The yellow school buses are largely unchanged since their debut in 1939 even though their routes are getting more and more complex in the past 80 years. This is because of the number of students, schools, road systems and the system of rules.
If you are a Boston citizen you already know that elementary students are assigned to a bus stop up to ½ mile away from their home and when they grow to middle school, the radius of the bus stop is up to 1 mile away. Boston Public Schools allows parents to select their child’s school from 10 possible options to reduce inequalities that might occur because of isolating kids to their neighborhoods. This is an excellent opportunity for parents and students but when you have to map out the school bus routes, it might be an enormous task.
On top of this, Boston Public Schools had the highest transportation costs in the USA. The local authorities are spending about $2000 per student per year. The total amount is about 10% of the district budget. The problem is worsening since more and more parents are driving their kids to school while the school bus system is getting more and more expensive every single year.
This is what has pushed the authorities to look for a solution to solve such a complex and costly problem. The solution should account for a lot more variables than the mentioned above. Road width might be an issue in Boston, a city established about 400 years ago. The presence of wheelchair lifts is a factor too, as well as students with monitors, just like students who have been in a fight which is resulting in their move on a different bus. About 5000 students with special needs in Boston require a door-to-door pickup and drop off is another factor that should be evaluated.
Boston Public Schools system adopted an interesting approach. They hosted a competition for a solution and gave out anonymized data to sets make things possible. The best of all applications came out of the MIT Operations Research Center. In 30 minutes, their algorithm created a system-level route map that was 20% more efficient than the human-made routes. Previously, this task was performed by 10 highly skilled and experienced Boston Public School Expert. Well, they need thousands of hours to create something not that good.
We`d like to open a bracket and to inform you that the PBS experts are still with the organization but currently they are tracking routes that struggle with on-time performance. Now, the MIT algorithm is creating the entire routing system at once, providing a base for human routers to tweak and improve. It is not just this. Among other benefits was the elimination of 50 buses, an 8% drop in the fleet that was the largest Boston had seen in a particular year. Busses covered a million miles less resulting in 20,000 pounds less carbon dioxide emissions per day. It allowed a lot of schools to start classes later, after 8 AM which is crucial for the pupils’ wellbeing and development.
The monetary effect of this AI-powered exercise was about $5 million. The saved amount was invested back into classroom initiatives.
Within logistic business, the ability of big data to create savings and efficiency is widely known. AI-based algorithms are saving a staggering amount of money and time. This is resulting in lower delivery costs as well as minimization of late shipments.
UPS, for instance, is a real-world example of data-driven logistics leading to better efficiency. After examination of the data generated by UPS trucks, the predictive analytics experts found out that vehicles turning left are costing a lot of money. This is because oncoming traffic, causing delays, wasted fuel and last but not least – safety risk.
Because of this, UPS redesigned its routing software to drop the shortest route in favor of a route eliminating as many left-hand turns as possible. As a result, the parcel service company managed to save 10m gallons of fuel, 20,000 tonnes of carbon dioxide and 350,000 more deliveries a year. Currently, just 10% of the turns of the UPS trucks are left and they occur only when they are completely necessary.