Connor McDowell

Student(s): Connor McDowell

Project: Resident Rotation Scheduling Optimization | View Poster (PDF)

Major(s): Industrial and Systems Engineering (Computer Science and Data Science minors)

Advisor(s): Karmel Shehadeh, Frank Curtis

Abstract

Medical Students must complete a residency program in order to become fully licensed doctors. These programs include residents at different years, each with unique requirements and rotations across hospital departments. Resident scheduling is a complex problem that requires balancing hospital needs, individual resident preferences, and legal and educational requirements. In optimization terms, a typical residents schedule may involve thousands of variables and constraints, even for a small number of residents. This makes it quite difficult to obtain an optimal solution. In many hospitals, the program director manually constructs the annual rotation schedule. They must balance educational, legal, and hospital requirements, accommodate residents’ vacation preferences, as well as adhere to the Accreditation Council for Graduate Medical Education requirements. Manual schedules often fail to satisfy these constraints, leading to unfair workloads and burnout.

To address this challenge, our team is developing a scheduling tool that optimizes resident schedules automatically. Our tool uses Python and Gurobi to read in data, implement an optimization model, solve it, and output a schedule to the user. Residents use an Excel template to input the necessary data and feed it to our tool, which can provide a viable solution within minutes. Since accurate data is essential for the model’s effectiveness, we have implemented a robust data validation framework to ensure data quality and provide clear error messages. This tool can produce schedules that ensure effective coverage of hospital departments and incorporate resident preferences. It also takes less time, releasing the program director for other tasks.