運用系統模擬與基因演算法於解決非等效平行機台之人力分配排程

排程(Scheduling)是生產管理中十分重要的領域，它的意義為在某一段時間內，分配有限的資源及決定一群工作進行的順序，以達成一定的目標或效率。因此排程往往被視為一個有限資源最佳化的分配工具。處理非等效平行機台之排程問題明顯地比相同機台及等效機台之排程困難許多，更何況增加考慮工作安排中人力分配的情形，則問題變更加複雜化。為了在短時間內有效找出非等效平行機台最佳的工作安排與人力指派的方式，本研究應用系統模擬與基因演算法來尋找可使總完工時間最小的人力指派及排程。本研究中，首先利用系統模擬軟體來建構非等效平行機台排程模式，接著使用軟體之GA模組來尋找最佳排程及人力配置的問題。模擬之結果會與窮舉法之結果作比較，期望找出符合實務排程需求之最佳建議。模擬結果顯示，窮舉法雖然能保證求得最佳解，卻非常消耗時間；至於使用系統模擬與基因演算法是可行的並且可在很短時間內找到相似最佳解。因此，本研究建議可針對不同的需求選擇適當的解決方法。

關鍵字

排程；系統模擬；基因演算法；非等效平行機台；總完工時間

並列摘要

Production scheduling is a very important area of management, and its significance for a certain period of time, the allocation of limited resources and decided to order a group of work in order to reach certain goals or efficiency. So scheduling is often seen as a limited allocation of resources to the best tools. The present study, the first to use simulation software to construct non-equivalent parallel machine scheduling model, and then use software to find the best GA module scheduling and staffing issues. Simulation results of the exhaustive method and the results compared to expectations in line with the practice schedule to identify the needs of the best available. Simulation results show that although the exhaustive method can guarantee the optimal solution obtained, it is very time-consuming; As for the use of system simulation and genetic algorithm is feasible and can be found in a very short period of time similar to optimal solution. Therefore, this study suggests that different needs can be to choose the appropriate solution.