- 學位論文
應用於精密模具製造之動態排程系統開發
The Development of Dynamic Scheduling System for Precision Mold Manufacturing
摘要
現今工業漸漸走向智慧製造的時代，產生快速變動的需求與環境，尤其精密模具是由複雜零件所組成，多數零件都由不同的工序所構成，目前多以人力和經驗去規劃與排程，利用電腦的運算能力加入一些限制條件達到有效率的靜態全域排程，在先前研究中所使用的靜態排程方法如FIFO(First In , First Out)、EDD(Earliest Due Date)，動態排程則使用本研究所開發的Modified ATC(Apparent Tardiness Cost)排程演算法，其中以SPT(Shortest Processing Time)、MS(Minimum Slack)這兩項派工法則為綜合考量，來達到多目標最優排程之效果，也是本研究所採用的。 本研究所開發的MAS(Multi-Agent System)是由不同的智能體所組成的計算系統，因應未來工業將進行智慧製造，在大規模的工序排程運算下，必須減少電腦運算時間以及擁有立刻回應製造現場狀況的能力，才可能符合未來工廠的實務需求。本研究將動態排程系統與靜態排程系統做結合，再透過多個Agent之間有效率的溝通以及分工到達最優的排程結果，最後本論文也會套用實際的18個零件、92筆製程作為案例探討來驗證其效能。
並列摘要
Nowadays the industry is moving towards smart manufacturing, which produces rapidly changing demands and environments. Precision molds are composed of complex parts and most of the parts are composed of different processes. At present scheduling is depends on experienced staffs. And using computer's computing plus some restrictions to achieve efficient static scheduling. This study combines the dynamic scheduling with the static scheduling. In previous studies the static scheduling uses the methods such as First In First Out (FIFO) and Earliest Due Date (EDD). Dynamic scheduling uses the modified Apparent Tardiness Cost (ATC) scheduling algorithm developed by this study, in which shortest processing time (SPT) and minimum slack (MS) are two comprehensive rules to achieve multi-objective optimal scheduling. The Multi-Agent System (MAS) in this study is a computing system composed of different agents. For smart manufacturing large-scale computing of schedule planning is required, the system is necessary to reduce the computing time and has the ability to respond manufacturing site conditions immediately. This study combines the dynamic scheduling with the static scheduling, and by the efficient communication and computing division between multiple agents to achieve the optimal scheduling results. Finally, this study applied the actual 18 parts, 92 process to verify its effectiveness.