A column generation mathematical model for a teaching assistant workload assignment problem

This paper presents a column generation-based modelling and solution approach for a teaching assistant workload scheduling problem that arises at academic institutions. A typical weekly workload schedule involves teaching deficiency classes, instructing problem-solving tuto- rial sessions, and allocating help-hours for students. For this purpose, a mixed-integer programming model that selects valid combinations of weekly schedules from the set of all feasible schedules is formulated. Due to the overwhelming number of variables in thismodel, an effective column genera- tion procedure is developed. To illustrate the proof-of-concept along withmodelling and algorithmic constructs, a case study related to the Department ofMathematics atKuwait University is addressed. Computational results based on real data indicate that the generated schedules using the proposed model and solution procedure yield improved weekly workloads for teaching assistants in terms of fairness, and achieve enhanced satisfaction levels among assistants, as compared to schedules obtained using ad-hoc manual approaches.