RASON: A new approach to the scheduling radiotherapy problem that considers the current waiting times
Résumé
Scheduling Radiotherapy treatments for cancer patient is a major concern for hospital and clinics. The main problem consists in minimizing the patient waiting time in order to maximize the treatment effectiveness. Most of the modern scheduling approaches use expert systems based on scheduling heuristics and algorithms to develop detailed schedules, in order to efficiently map the patients requirements to the treatment capacity of the health center. In this paper, we propose RASON, a new heuristic based scheduling algorithm for radiotherapy treatments, which main objective is to minimize the average waiting time for each patient. In contrast to well-known existing approaches, our solution manages a priority list that can be dynamically updated according to both the patient category and his/her current waiting time. The generated schedule also impacts the minimization of the average tardiness of the first treatment sessions for each patient. We have evaluated our algorithm using both real data from the Institute of Radiotherapy in Santiago, Chilean and artificial cases generated with a self-developed generator called GeneRa. GeneRa is able to generate cases according to particular constraints inherent to several countries like UK, France and Italy. We show in our proposal evaluation that an on-the-fly scheduling has a great positive impact, allowing to reduce the average waiting time and tardiness for all patients categories. Our algorithm outperforms the JIT and ASAP well-known approaches, with a 95% statistical significance. Our scheduling algorithm allows to significantly reduce the treatment waiting time for different categories of patients. This is a major improvement for the patients as time and delays are crucial parameters to achieve the best effectiveness in cancer treatments.