A simulation-based framework to schedule surgeries in an eye hospital

Abstract

In this article, we propose heuristics to schedule surgeries in a German Eye hospital on a daily basis. We are interested in reducing the waiting time of the patients and in increasing the utilization of the operating rooms (OR). The sum of the overtime of the staff and the time where the ORs are not utilized for surgeries are used as a surrogate measure for OR utilization. A Non-Dominated Sorting Genetic Algorithm II (NSGA-II) that uses a random key representation for the schedules is proposed to solve heuristically this scheduling problem. The NSGA-II approach is hybridized with a local search approach. Discrete-event simulation is used to assess the fitness of the chromosomes within each generation of the NSGA-II approach taking into account the expected availability of surgeons, anesthesia doctors, nurses, stochastic surgery durations, and preferences of the patients with respect to the point of time of the surgery. The simulation model is described in detail and used to study the impact of stochastic arrival patterns of the patients and stochastic surgery durations on the performance of the executed schedules. Rescheduling strategies are investigated in different situations. We present the results of computational experiments using data from an actual eye hospital. These results clearly demonstrate that schedules obtained by dispatching rules and manually by the medical staff are outperformed by the proposed NSGA-II approach. Finally, a successful real-world implementation of the scheduling heuristic is briefly discussed.