An Optimization Approach to the Make-Up Final Exam Scheduling Problem with Unique Constraints

Final exam scheduling is typically a simple task. Exam scheduling at the United States Air Force Academy, however, is unique in that it must abide by a number of institutional constraints. Some include: ensuring all 4,000 cadets complete their exams within one week, assigning a longer time block for final exams than regular class meetings, and limiting the number of exam periods per day. The Dean of Faculty Registrar’s office must also accommodate cadet absences during the scheduled final exam week on short notice. Addressing these issues is currently a reactive process—cadet absences are identified, and then the Registrar creates adjusted schedules by hand. In this study, we create a data-cleaning VBA module for input data and use Xpress software to execute one of three versions of a scheduling optimization model. The model reduces the scheduling task time from several days to less than five minutes.