Appointment scheduling with unscheduled arrivals and reprioritization

Inspired by the real life problem of a radiology department in a Dutch hospital, we study the problem of scheduling appointments, taking into account unscheduled arrivals and reprioritization. The radiology department offers CT diagnostics to both scheduled and unscheduled patients. Of these unscheduled patients, some must be seen immediately, while others may wait for some time. Herein a trade-off is sought between acceptable waiting times for appointment patients and unscheduled patients’ lateness. In this paper we use a discrete event simulation model to determine the performance of a given appointment schedule in terms of waiting time and lateness. Also we propose a constructive and local search heuristic that embeds this model and optimizes the schedule. For smaller instances, we verify the simulation model as well as compare our search heuristics’ performance with optimal schedules obtained using a Markov reward process. In addition we present computational results from the case study in the Dutch hospital. These results show that a considerable decrease of waiting time is possible for scheduled patients, while still treating unscheduled patients on time.