#338 Joint optimization of school bus routes and last mile services

Principal Investigator
Peter Zhang
Overdue Project
Start Date
July 1, 2020
End Date
June 30, 2021
Research Type
Grant Type
Grant Program
FAST Act - Mobility National (2016 - 2022)
Grant Cycle
2020 Mobility21 UTC


School bus routing across multiple districts requires the coordination of many resources. In this project, we plan to understand the challenges in current school bus routing through a partnership with Allies for Children, and improve the mobility of students via decision analytical models and deployment of information technology tools. Our main focus is to find efficient, safe, and implementable multi-modal transportation methods to help children move to and from schools. Last-mile and first-mile transportation is an existing problem in this context, due to the 1-2 mile walking distance between home and bus stops for some students, which poses safety and accessibility concerns. Our goal is to design an analytical model to support the joint optimization of bus routes and last-mile and first-mile mobility services for students, and provide a deployment technology that can help connect students, parents, buses, and ridesharing services to achieve reliable transportation. 
According to the Shared Transportation Guide of Allegheny County, per state law, elementary school students in Pennsylvania may need to walk up to a mile and a half to a school bus stop, and secondary school students may need to walk up to two miles. Such first- and last-segments of students' daily commute create a ``first and last-mile'' problem (or last-mile for short) for students' daily travel. The unavailability of first-mile and last-mile transportation service is one of the main deterrents to the use of school buses for certain families, and also a primary cause for safety and security concerns to parents, especially during bad weather, early morning, and evening time. Although the selection of bus lines and stops take into account the pick-up and drop-off locations for each student, the walking distance to the bus stop for some students will be still long under cost consideration given a limited fleet of school buses. In other words, the ``last-mile" problem is always serious for some students. 

With the rapid development and popularization of mobile and wireless communication technologies, ride-sourcing companies, such as Uber, Lyft, and Z-Trip, have been able to leverage internet-based platforms to operate transportation services with flexible routes and schedules. These companies can connect passengers and vehicles in real-time. We propose to explore the feasibility and actionable operational plans to use ride-sourcing vehicles as feeding last-mile services to the school bus. Specifically, a vehicle from Z-Trip with a pre-arranged route is able to bring several students from their home to the bus stop in one shared trip. Therefore, with integrated optimization models to consider both school bus routing and stops selection, as well as ride-sourcing feeding shared routes and schedules, we may provide more efficient and safer transportation technologies to students.    
Month 1 to Month 12:

M1-M2 Problem definition and literature survey;
M3-M4: Data collection and data mining;
M5-M8: Optimization model formulation and analysis;
M9-M10: Algorithms development and numerical experiments;
M11-M12: Deployment plan and pilot testing.    
Deployment Plan
Deployment starts with simulation in Month 9, after having an optimization model that prescribes the bus routes, bus stops, and ridesharing planning.

Phase 1 (M9-M10): testing
1. Setup of simulation systems for multiple districts and bus routing tools.
2. Implementation of optimal solutions in simulation systems, and stress testing with scenarios that include traffic uncertainties, weather uncertainties, and seasonality.

Phase 2 (M11-M12): pilot
1. Adding ridesharing and last-mile service to one existing bus route
2. Monitoring of student and bus times and rideshare timings
3. Quantifying times of various transportation segments (objective measure)
4. Service quality feedback from students, parents, bus driver, and rideshare services (subjective measure)
5. Constructing more detailed metrics for the optimization model, based on objective and subjective measurements of ridesharing performance
6. Developing mobile apps for connecting the service providers (buses and ridesharing) if time allows, and provide real-time monitoring functionality for parents.
Expected Accomplishments and Metrics
1. Data collection for the various components required for the design of efficient and equitable multi-modal transportation systems for multiple districts (details of data can be found in the data management plan). Data management and storage in a safe and re-usable format.

2. Understanding of key technical, strategic, and political issues underlying the problem.

3. Identification of timely and useful technical and strategic angles to improve transportation service to students. For example, a communication protocol that could connect buses, ridesharing and last-mile service vehicles, school districts, and student families.

4. Understanding the feasibility to collaborate with Allegheny County Port Authority (e.g., with the utilization of Global Transit Feed Specifications (GTFS)).

5. Pilot to test the feasibility of multi-modal transportation for students: joint optimization of bus routes, stops, and last-mile services.    

Individuals Involved

Email Name Affiliation Role Position
haiwang@cmu.edu Wang, Hai CMU Other Other
pyzhang@cmu.edu Zhang, Peter CMU PI Faculty - Untenured, Tenure Track


Amount of UTC Funds Awarded
Total Project Budget (from all funding sources)


Type Name Uploaded
Data Management Plan Data_management_plan.pdf Jan. 3, 2020, 8:22 p.m.
Project Brief project_description_2020_03_09.pptx March 17, 2020, 7:33 a.m.
Presentation SBRS_Poster_One_Page.pdf Sept. 28, 2020, 7:43 p.m.
Progress Report 338_Progress_Report_2020-09-30 Sept. 28, 2020, 7:45 p.m.
Presentation Fleet Allocation for a Last-Mile Transportation System under Demand Uncertainty March 29, 2021, 8:18 p.m.
Presentation Urban TSP estimation: a dataset and software tool for operations and transportation researchers March 29, 2021, 8:18 p.m.
Presentation Gaps and Opportunities in School Bus Transportation March 29, 2021, 8:18 p.m.
Presentation School Bus Sharing March 29, 2021, 8:18 p.m.
Progress Report 338_Progress_Report_2021-03-31 March 29, 2021, 8:19 p.m.

Match Sources

No match sources!


Name Type
Allies for Children Deployment Partner Deployment Partner