Historical data-driven nurse flexible scheduling problem

The nurse scheduling problem (NSP) is a complex combinatorial optimization problem, we aim is to use the nurse resource reasonably. In this paper, firstly, with the method of time series analysis, an autoregressive integrated moving average (ARIMA) model is established to forecast the number of patients, which is used as input to calculate the volumes of nurse for scheduling by queuing theory. In the aspect of NSP, a comprehensive integer programming model considering nurse's levels and their preferences to different shifts is established, with a series of labor regulations. Finally, in order to get a near-optimal scheduling, a heuristic algorithm combined with a series of transformation rules is designed. The contribution in this paper is threefold. Firstly, it satisfies all the constraints and obtains a near-optimal scheduling. Secondly, it can control patient waiting time validly. Thirdly, it can adjust the number of nurses to the shift dynamically.