Optimizing supply chain management in hospitals can lead to a significant reduction in costs. One of the supply chain processes that can be optimized is the sterilization of reusable instruments used during surgical procedures. This process takes place at the Central Sterile Supply Department (CSSD) and consists of two main steps, `washing and disinfection' and `sterilization'. These steps are executed by batch processing machines, with a setup time prior to each step. The setup at each stage is executed manually. The machines of a CSSD are renewed every 10 years, which provides a opportunity to assess all capacity planning decisions. In this thesis, the aim is to determine the required resources of the CSSD, while minimizing the total costs and guarantying the availability of instruments for scheduled surgeries. Costs include: acquiring machines, machine batch costs, machine maintenance costs, and staff costs depending on the number of opening hours. This thesis contributes to current research by proposing a framework for the capacity planning decisions at a CSSD, by extending existing models by taking specific characteristics of the CSSD into account, and considering a new objective function, namely minimizing the total costs. Capacity planning decisions are considered on three hierarchical levels. First, the strategic level involves long-term decisions, where the number and type of required machines have to be determined. Second, on a tactical level, the amount of opening time has to be determined. Third, on an operational level, the instrument sets have to be scheduled within batches and machines. The sterilization process is formulated as a mixed integer linear problem (MILP). The problem is described as a multi-stage hybrid flow shop with additional constraints to represent the specific characteristics of the CSSD. This model takes into account capacity planning decisions on all three levels. The MILP formulation is proven to be NP-hard. The demand for sterilized instruments is determined from historical data from the Leiden University Medical Center. To evaluate the model, three instances with a timespan of a week are created. Preliminary results showed that the multi-stage flow shop is difficult to solve for real-life instances. Hence, a decomposition, based on the three hierarchical levels, is proposed. The decomposition leads to three levels of optimization models. The strategic model takes into account the scheduling of instruments per day, while the tactical model schedules per day parts, and the operational models determine a specific point in time. The proposed models are individually tested on their performance. The results from the strategic model are used as input for the tactical model. Results show that the strategic model underestimated the required amount of opening time, as it does not take into account the spread of release times. Furthermore, the computational results show that the tactical and operational models are difficult to solve for real-life instances. Hence, a heuristic approach is proposed by forming a chain of the models. The best results were obtained by setting a minimum amount of opening time for the strategic model and using the resulting machines as input for the tactical model. The results show that the instrument sets are equally spread over the week. To conclude, the results of this thesis contribute to quantifying the required resources for a sterilization process within a hospital. However, to obtain more practical results, future research is required. Suggestions for future research include: taking uncertainties within the process into account, the application of metaheuristics, and the required number of employees.