Many approaches are used to handle uncertainty in stochastic combinatorial optimization problems. In this talk, we describe the application of a tabu search approach in a stochastic environment together with a real application in physician scheduling in a radiotherapy center. The goal is to determine a weekly cyclic schedule that improves the patient flow and shortens the pretreatment duration. High uncertainty is associated with the arrival day, profile and type of cancer of each patient.
Additionally, two approaches to improve the efficiency of the method are introduced, both are based on leveraging methods that originate outside the field of metaheuristics. The first one discusses hyperparameters tuning. Research shows that it is a nontrivial task and efficient methods are required to obtain the best possible results. We present how blackbox optimization can help choose the tabu search parameters efficiently. We are solving this problem through a Mesh Adaptive Direct Search (MADS) algorithm with no derivative information. The second one presents a learning algorithm for improving tabu search by reducing its search space and evaluation effort. The learning tabu search algorithm uses classification methods in order to better motivate moves through the search space.