Dynamic-Stochastic Radiotherapy Appointment Scheduling

Petra Vogl, Roland Braune, Walter Gutjahr, Karl Franz Dörner

When scheduling the start times for treatment appointments of patients in hospitals or outpatient clinics such as radiotherapy centers, minimizing patient waiting time and simultaneously maximizing resource usage is crucial. Significant uncertainty in the appointment durations makes scheduling those activities particularly challenging. To address and analyze this uncertainty, the current study uses real-world data on appointment durations gathered from an ion beam therapy facility and fits a distribution function to appointment durations. The authors introduce a novel buffer concept, based on percentiles of the fitted distribution, and thereby calculate a planned activity duration that deviates from deterministically assumed durations. The goal is to find a robust
baseline schedule which minimizes a weighted sum of patient waiting time, machine uptime, and penalties due to time window violations by determination of an optimal buffer parameter. To derive a dynamic scheduling strategy from the baseline schedule, the authors introduce a reactive procedure that adapts the schedule to the actual patient flow. The quality of the actual schedule derived this way from a candidate baseline schedule is evaluated by different sampling approaches. For different buffer parameter sizes, different solution evaluation strategies turn out to be superior.

Department of Business Decisions and Analytics, Department of Statistics and Operations Research, Research Network Data Science
Computers and Industrial Engineering
Publication date
Peer reviewed
Austrian Fields of Science 2012
502052 Business administration
Portal url