Medical doctor rostering problem in a hospital emergency department by means of genetic algorithms

• Published in: Computers & industrial engineering Vol. 56; no. 4; pp. 1232 - 1242
• Main Authors: Puente, Javier, Gómez, Alberto, Fernández, Isabel, Priore, Paolo
• Format: Journal Article
• Language: English
• Published: New York Elsevier Ltd 01.05.2009; Pergamon Press Inc
• Subjects: Emergency medical care; Genetic algorithms; Health services; Heuristic; Medical staff; Rostering; Scheduling; Shift work; Studies; Timetabling; Spain; Timetabling; Health services; Medical staff; Rostering; Genetic algorithms
• DOI: 10.1016/j.cie.2008.07.016
• ISSN: 0360-8352

Organising shifts, or work rosters, is a problem that affects a large number of businesses where employees are subject to some kind of work rotation. Researchers in the fields of Operations Research and Artificial Intelligence have resorted to several different optimisation systems to solve the problem. The motivation for the medical-staff shift-rotation research presented in this paper stems from the needs of an actual hospital emergency department (HED) and from the observed growing staff of these services in Spain. The problem approach, which has been hardly dealt with in the literature, intends to automate the creation of time-tables by applying genetic algorithms (GAs) in an actual HED. HEDs work organisation becomes different because of the combination of shifts and 24-h duties. After knowing the HED workers’ requirements (which will allow to identify the hard and soft constraints imposed to the problem) and after defining the adequate encoding to be used in the solutions, a heuristic-schedule builder –designed ad hoc to satisfy the hard constraints – produces an initial population of feasible solutions. Afterwards, iteratively, GA obtains new generations of feasible individuals, thanks to the use of a specific crossover operator, based in the exchange of whole work weeks, that operates together with a repair function. Once the optimum is reached, the results obtained are discussed as a function of the degree of satisfaction of the constraints under which the system operates and of the adaptability of the system as the constraints vary.