Producing planning of doctors’ roles, including night shifts, on-call shifts (doctor can be called back in case of need), and regular working day in a hospital is complex and it is often difficult to effectively address all real world constraints. Furthermore, the produced solution requires a very good user acceptance to be effectively deployed in production. This paper reports on the application of metaheuristics to solve a planning problem in an hospital and focuses on user acceptance aspects of an algorithm. One of the key aspects to ensure user acceptance is to ensure the comprehensibility of the delivered solution. This includes the understanding of both the algorithm itself and its executions. The algorithm is a composition of local search, greedy algorithms, tabu search, and a few additional metaheuristic principles, and proved both good and fast enough.
Voir en ligne : http://www.icores.org