Drawing Between the Lines: Uncovering School Attendance Area Boundaries Through Simulation Algorithms with Context-Specific Constraints

Abstract

Simulation algorithms have become increasingly popular in solving apportionment problems in political redistricting, but they are largely untapped in the practice of designing school boundary maps. Many school districts have initiated processes to modernize attendance area boundaries, which determine school assignments based on residential addresses; however, most education boundaries today continue to be drawn manually. Fairfax County Public Schools (FCPS) in Virginia, the ninth largest school district in the United States, recently conducted a comprehensive review of its attendance area boundaries. Using FCPS as a timely case study, this thesis shows that statistical sampling algorithms can uncover new boundary configurations that improve proximity, connectivity, resource balance, and continuity of learning. I illustrate the need for problem-specific constraints by demonstrating that maximizing compactness is not representative of minimizing commute times, despite past works conflating the two metrics. I show that simulations can identify geographic focus areas that contribute the greatest reduction in split feeders, which are unrealizable from local adjustments. Drawing school boundaries is a complex problem, and simulation algorithms can reveal possibilities that are invisible to manual methods, which can alleviate financial, health, and time burdens for students and families.