We consider the daily appointment scheduling for a given set of outpatients needing chemotherapy at a facility on a given day. The chemotherapy of a patient includes three phases: consultation with the referee oncologist, drug preparation, injection on a bed. A patient may leave with a given deferral probability after consultation due to her inability to receive the chemotherapy. We pursue several contributions through this research : (1) characterize global sequences as reasonable solutions for this problem ; (2) introduce properties of sequences with high quality if the daily demand is fixed ; (3) provide a tool to evaluate overbooking strategies ; (4) develop a subroutine to evaluate daily scenarios of demand ; (5) investigate scenario reduction methods and their application in different objective functions. Following these goals we compare some list algorithms against an exact stochastic program and a GRASP heuristic. We investigate the minimization of both the closing time and the overworking time of the facility as objective function. Benchmark instances are derived from data of a chemotherapy center.
Mots clés :