A Co-Design Framework for School Bus Network Redesign: Applying Category Theory to Routing, Resources, and Equity
Name
temkin-dtempkin-sb11-dusp-2026-thesis.pdf
Size
5.02 MB
Format
Adobe PDF
Checksum (MD5)
cd15f38b3e64ed229854b12e3207841c
Author(s)
Temkin, Diego
Advisor(s)
Aloisi, Jim
Date Issued
May 2026
Publisher
Massachusetts Institute of Technology
Abstract
Cities and school districts face persistent challenges in designing bus systems that are efficient, reliable, and equitable under limited fleet, staffing, budget, and accessibility constraints. Existing optimization approaches can improve routing performance, but they often treat equity, stakeholder priorities, and data integration as secondary evaluation steps rather than as part of a shared design representation. This thesis develops a co-design framework for bus network redesign using the monotone theory of co-design from applied category theory. The framework represents routing service, fleet capacity, drivers, monitors, fuel, maintenance, policy requirements, and algorithmic choices as interconnected components whose resource and functionality relationships can be compared through Pareto-front analysis.
The framework is demonstrated through a case study of Framingham Public Schools morning bus service. Using district-provided operational data, OpenStreetMap-derived travel arcs, school locations, student assignment information, fleet characteristics, and Census/ACS-derived demographic indicators, the study adapts a Biobjective Routing Decomposition backend into a Bus Routing Algorithm for evaluating existing routes and generating alternative designs. Under simplified travel-time assumptions, BRA serves the same 4,780 currently assigned riders with 48 buses rather than 51 and reduces estimated mean ride time from 43.08 to 30.52 minutes. Expanded-service experiments show that students living at least 1.5 miles from school can be served within the available fleet under the modeled constraints, while universal default service leaves a small number of students unassigned. These results are planning estimates rather than deployment-ready schedules because they rely on constant-speed travel times and incomplete student-level linkage data. Even so, the case study shows how compositional modeling and Pareto-front analysis can make cost, coverage, accessibility, and equity-related tradeoffs more transparent for transportation planning.
MIT Department
Massachusetts Institute of Technology. Department of Urban Studies and Planning
Terms of Use
Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0)
Copyright retained by author(s)
Persistent DSpace Link
