A Column Generation Approach for Personnel Scheduling with Discrete Uncertain Requirements

In this work, we model a personnel scheduling problem with uncertain demand as a two-stage stochastic integer program. The model is a large integer program with a large number of columns and constraints which creates difficulty for optimization process. We apply column generation method and Benders' decomposition technique to solve the problem. We test our proposed algorithm on some generated instances and obtain satisfying results showing improvement in obtaining good solutions quickly over solving MIP on GAMS with CPLEX solver.