Using metaheuristics and queueing models to optimize schedules in the academic enterprise

We develop a prototype system for enterprise-wide scheduling optimization in an academic institution. The system is built around a rule-based optimization engine, the Optessa MLS™, that has proved successful in complex scheduling applications in manufacturing. The engine utilizes metaheuristics with sophisticated accelerators for local search. A novel contribution is a consistent queueing model characterization, of key aspects of course schedules and commuter parking lot requirements, that can be used to quantify optimization benefits and predict how course schedule changes impact commuter parking lot efficiency. In a case study at Monmouth University, our course schedule optimization “frees-up” 20-23% of schedulable rooms for other uses and increases parking rate capacities by 25-35%, with a potential net annual revenue increase of about $12-17M.