Nurse Scheduling With Multi-levels And Individual Preferences

The nurse scheduling problem is to allocate shift to every nurse , knowing the nurses'number of the department , the demand of every shift and the level differences between nurses ,considering nurses'personal preference ,making sure that the daily care work of the hospital can be done effectively, so as to help promote the work effiency and achieve the optimal allocation of human resources. The scheduling process needs to consider a variety of constraints, leading to a result that it is difficult to develop a reasonable frequency nurse shift table, let alone develop a good frequency table to be optimized, so how fast to develop an effective frequency table of nurses has become a very difficult problem in many hospitals. This article is aimed at this issue for study.Recently,a current number of scheduling algorithm used in the nurse scheduling have been proposed ,however ,some shortages of these algorithm have been noticed in practice. The randomness of these algorithm prevent improving the solution quality, this paper made some improvents, so that the solution process can be more efficient. In this paper, the solution process which is closely around the demand constraints of nurses, can be divided into two major stages to be carried out respectively. The first stage whose target is to ensure that the care can be accomplished by the nurses in the department ,the second stage which is aimed to ensure that there have enough nurses in each shift , to meet the nurses'preference as far as possible, and to ensure the fairness while discovering that there is not enough nurse in the first phase, then there is no need to waste time,one can report to the leaders directly,suggesting to recruiting nurses to complete the nursing work, as a result that the unnecessary scheduling process Time can be saved. This process is the core of nurse scheduling problem, which will directly affect the solution quality in the next stage. Solving the second phase is based on the feasible solution obtained in the first phase, using some rules to generate a new solution space to search a better solution in the solution space through heuristics. Finally, this proposed method has been validated through a number of instances, showing that the proposed algorithm for nurse scheduling is reasonable.