跟診人員排班必須滿足一天班表限制，包含人員專長與診間的媒合度、不連續三連班，以及一週班表限制，包含不連續五天早班、不連續三天晚班和延診跟診人員不排班。目前跟診人員班表皆由專員手動排班來完成，是一件十分費時費力的工作。 本研究提出一個結合基因演算法與啟發式規則的跟診人員排班方法。首先利用基因演算法為一週中的每一天排班，以快速求得滿足一天班表限制的最佳班表。各天最佳班表所組成的一週班表，會存在違反一週班表限制的情況。因此，再利用啟發式規則針對違反一週班表限制之跟診人員進行班表調整，進而使得一週班表可以滿足醫院所提出之限制條件。利用個案醫院所提供的64位跟診人員、105早診、82午診、和53晚診為實驗資料進行測試，實驗結果顯示本方法可有效輔助排班專員完成滿足各種限制的跟診人員一週排表。 The nurse scheduling problem must satisfy many soft and hard constraints, therefore to complete the manual scheduling with the Commissioner is a very laborious work. This study proposes a combined method of genetic algorithms and heuristic rules for the nurse scheduling problem. At first, for every day of a week we employ genetic algorithms to quickly obtain the best schedule that meets the one-day scheduling constraints. The weekly schedule composed of the best schedule of each day will violate the constraints of weekly schedules in some cases. Therefore, we utilize heuristic rules to adjust the schedules of nurses that violate the weekly schedule constraints, so that the final schedule can meet the restrictions proposed by the hospital. We make use of the data provided by a case hospital to evaluate the proposed method. The data consist of the 64 nurses, 105 early clinics, 82 afternoon clinics, and 53 night clinics. Experimental results show that this method can effectively assist commissioner to complete the nurse scheduling that meets a variety of constraints.
