Schools Use Higher Math for Better Bus Routing

Before and after route maps for Slippery Rock School District in Pennsylvania after optimization by local university researchers. Before and after route maps for Slippery Rock School District in Pennsylvania after optimization by local university researchers.

Expanding on a smart routing article in the September 2017 issue of School Transportation News magazine, this online exclusive shares how technology and higher math cooperated to create improved routing experiences for two school districts. 

School districts on a tight budget are turning to local universities for AI projects to optimize bus routing. Slippery Rock School District in Pennsylvania and Howard County Public School System in Maryland both realized savings when academic teams stepped in to evaluate their bus operations.

“With millions to billions of variables in routing, there’s always room for improvement. People may think they have something good, but familiarity isn’t necessarily the best solution,” said Professor Sam Thangiah of Slippery Rock University.

Through a grant, Slippery Rock University’s researchers wrote a program using heuristic algorithms and real-time data to configure routes for the nearby rural school district. The team considered the needs of five schools, 600 students, a mixed-fleet of 50 buses, 13 depots, and 71 pick-up points within 127 square miles. After integrating information with the Intelli-SIIMS school bus routing system, the district enjoyed a 15 percent operational efficiency, resulting in over $70,000 in annual savings.

For a nominal fee, the University of Maryland’s QUEST Program recommend ways to improve bus routing for Howard County Public Schools. As previously reported, engineering researchers used data analysis and mathematical modeling to find an optimal solution without modifying the bus trips. A computer algorithm assigned new routes to reduce deadhead time, save money and improve air quality. While individual routes become longer, the solution allowed buses to serve more trips in a single route.

“We developed models and algorithms to solve routing problems, and they can use any off-the-shelf optimization software. And there are several of those available,” said Professor Ali Haghani, project lead in the University of Maryland’s Department of Civil and Environmental Engineering, who presented the work this summer at the STN EXPO.

 

Last modified onThursday, 07 September 2017 12:33