目前醫院的跟診人員班表，是由排班專員負責手動排班，排班時需要考量到人員專長和值班診間之媒合度、休假問題以及醫院的規範，因此手動排班十分費時費力，且其排班專員亦有其跟診之工作。 本研究參考以往文獻，使用基因演算法求解，並提出自行設計之交配方法。本研究方法欲求解排定一天班表之問題，並使用參考個案資料所模擬之資料進行實驗，其實驗結果證明，本研究提出之基因演算法，平均在33代可求得近似最佳解，平均在165代達到收斂之情況，且每次求得之近似最佳解，染色體適應值皆有一定之水準，求解的穩定性極高。 At presently, the shift in the nurse schedule in the hospital is arranged by human resource. While arranging, the matching degree between the duty clinical rooms and nurse’s professional specialty, leave timing, and the term of hospital are all considerate. According to this, it takes so much time and is so strenuous. Besides, the commissioner has his own work in the clinical room when the nurse schedule arranging. In this study, the past literature is referred and genetic algorithm (GA) is used to find the solution, and bring out a new crossover method which is designed by my selves. In order to solve the nurse schedule problem for one day, the simulated data of reference case is used in experiment. From the results of experiments, the GA which is proposed in this study is able to get the best solution after 33 generations and convergence after 165 generations averagely. Every time the fitness value of the best solution from GA maintains good quality and stability.
