• robust modeling
• MSS problem
• OR scheduling

This thesis work is based on the creation of optimization models for the Master Surgical Scheduling (MSS) problem, i.e., given a scheduling horizon, a list of patients, a number of available beds, and a grid in which, for each day of the time horizon, the total number of hours available for each operating room, we must determine a schedule in which, for each day and for each operating room, a family of surgeries and the number of surgeries to be performed are assigned such that the number of patients does not exceed the number of available beds and that there is an appropriate balance between long and short scheduled surgeries.
The work is concerned with creating a deterministic model with dynamic session lengths, delegating to the model the division of days into blocks of time. It uses a pattern-based approach divided into two phases: the first phase deals with generating the patterns using three different objective functions, while the second phase deals with associating a pattern for each day of the planning horizon and for each operating room using two different objective functions. For the model experimentation we used the Branch&Bound technique. The results obtained show that the new model, being more flexible, returns better results than the model with static sessions defined a priori.
In the second phase of the thesis work, we formulated a robust model of the previously proposed model to handle the uncertainty due to the duration of the surgical operations.