5.1 研究的結論
對半導體晶圓廠來說，傳輸整合步進機是單價最昂貴的機台，因此成為工廠之瓶 頸機台，故傳輸整合步進機之排程之議題相當重要，然而在過去的研究上，多在滿批 量之情境作探討，因此往往將其視為一個機台，而未了解該機台之內部結構，可視為 一種流程式生產型態；隨著新製程和新產品之導入，晶圓廠出現小批量之情境，使得 傳輸整合步進機會因為機台結構之設計，造成產能之損失，若一間晶圓廠綜合了新製 程和新產品之情況，小批量之情境將漸趨增加，也就意謂著產能損失之發生頻率將隨 之增加，而過去對於小批量情境之傳輸整合步進機的排程研究僅有Wu & Chiou（2009）
和Wu , Lu , & Chiou (2009)作探討。
Wu & Chiou（2009）所探討的情境為單機和多機，皆未考慮 Job-family 之情況，
即未將光罩之限制納入考量，其透過七種演算法搭配單獨式派工法則進行求解； Wu , Lu , & Chiou (2009)為研究單機且考慮 Job-family 之情境，運用基因演算法搭配家族式 派工法則進行求解。然而實際的半導體晶圓廠存在著多機和 Job-family 之情境，故本 研究所提出之啟發式演算法搭配家族式派工法則可以在短時間內，得到一組近似最佳 解（工件加工順序），使得平均產出最大。而在經過大量的實驗後，可以將本研究的結 論歸納如下：
z 當欲排序的工件數目越多及排序工件的良率越低，應使用演算法結合家族式 派工法則。
z 在處理多個步進機的排序問題，本研究提出新的染色體設計，可在不增加染 色體的大小情況下，減少解題空間，使得解題品質及求解的速度獲得改善。
z mGA-Tabu-F 演算法的運算品質績效在大部份的情境下，比其他演算法結合 家族式派工法則有更優異的績效表現。
5.2 未來研究方向
本研究的未來研究方向，整理如下列數項：
1. 本研究以啟發示演算法 (meta-heuristic algorithm)找出近似解，因此未來可以 嘗試運用其他的方法求解，如：混合整數規劃、動態規劃或分枝界限法等。
2. 本研究主要考慮以最大產出為主，但對於部分代工晶圓廠之生產目標則主要 以客戶達交率為目標，故可以遲交為目標式作為探討。
3. 由於在小批量情境下，發生產能閒置問題的主要原因為 Port 數量不足，故未 來可針對Port 數量作研究，如：探討 Port 數量應設定多少，方可使傳輸整合 步進機不會發生產能損失的問題。
4. 發展新的演算法。
參考文獻
Cheng TCE, Gupta NDJ and Wang G (2000). A review of flowshop scheduling research with setup times. Production and Operations Management 9: 262-282.
Chern CC and Liu YL (2003). Family-based scheduling rules of a sequence-dependant wafer fabrication system. IEEE ctions on semiconductor manufacturing 16: 15-25.
Croce FD, Tadei R and Volta G (1995). A genetic algorithm for the job shop problem.
Computer Operation Research 22: 15-24.
Dorigo M (1992). Optimization, Learning and Natural Algorithms. Ph.D. Thesis, Dipartimento di Elettronica, Politecnico di Milano, Italy
Franca PM, Gupta JND, Mendes AS, Moscato P and Veltink KJ (2005). Evolutionary algorithms for scheduling a flowshop manufacturing cell with sequence dependent family setups. Computers & Industrial Engineering 48: 491-506.
Glover F (1989). Tabu Search - Part 1, ORSA Journal on Computing 1: 190-206.
Goldberg DE (1989). Genetic algorithms in search, optimization and machine learning.
Addison-Wesley, Boston.
Gonzalez T and Sahni S (1978). Flowshop and jobshop schedules: complexity and
International Conference on Neural Networks-IV:1942-1948.
Holland JH (1975). Adaptation in Neural and Artificial Systems. University of Michigan Press, Ann Arbor, Michigan.
Kirkpatrick S, GELATT JR, CD and VECCHI MP (1983). Optimization by simulated annealing. Science 220: 671-680.
Lin SW, Gupta JND, Ying KC and Lee ZJ (2009). Using simulated annealing to schedule a flowshop manufacturing cell with sequence-dependent family setup times. International Journal of Production Research 47: 3205-3217.
Mun KJ, Lee HS and Park JH(2005). Parallel Genetic Algorithm-Tabu Search Using PC Cluster System for Optimal Reconfiguration of Distribution Systems. KIEE International Transactions on Power Engineering 5:116-124
Potts CN and Kovalyov MY (2000). Scheduling with batching: A review. European Journal of Operational Research 120: 228-149.
Quirk M and Serda J (2001). Semiconductor manufacturing technology. Prentice Hall, New Jersey.
Rajendran RC and Hans Z (2004). Ant-colony algorithms for permutation flowshop scheduling to minimize makespan/total flowtime of jobs. European Journal of Operational Research 155: 426-438.
Reeves CR (1995). A genetic algorithm for flowshop sequencing. Computer Operation
Research 22: 5-13.
Sundarambal M, Chithara BJ and Anbalagan P (2008). Meta-Heuristic Techniques for Bandwidth Optimization through Dynamic Routing in ATM networks. International Journal of Soft Computing 3: 189-194.
Wang CS and Uzsoy R (2002). A genetic algorithm to minimize maximum lateness on a batch processing machine. Computers & Operations Research 29: 1621-1640.
Wilson AD, King RE and Hodgson TJ (2004). Scheduling non-similar groups on a flow line:
multiple group setups. Robotics and Computer-Integrated Manufacturing 20: 505-515.
Wu MC and Chiou CW (2009). Scheduling Semiconductor In-line Steppers in New Product/Process Introduction Scenarios. International Journal of Production Research (accepted).
Wu MC, Chiou SJ and Chen CF (2008). Dispatching for make-to-order wafer fabs with machine-dedication and mask set-up characteristics. International Journal of Production Research 46: 3993 - 4009.
Wu MC, Huang YL, Chang YC and Yang KF (2006). Dispatching in semiconductor fabs with machine-dedication features. International Journal of Advanced Manufacturing Technology 28: 978-984.
Wu MC, Lu CW and Chiou CW (2009). A Family-based GA Algorithm for Scheduling an In-line Stepper in a Low-Yield Scenario. Submitted to IEEE Transactions on Semiconductor Manufacturing (under review).
Ying KC and Lin SW (2009). Raising the hit rate for wafer fabrication by a simple constructive heuristic. Expert Systems with Applications 36: 2894-2900.