連續作業等候時間限制下之多工作站生產系統控制

Production Control in Multi-stage Tandems Systems under Sequential Process Queue Time Constraints

曾俊翔(Chun-Hsiang Tseng)

指導教授 ： 吳政鴻

國立臺灣大學/工學院/工業工程學研究所/碩士(2012年)

https://doi.org/10.6342/NTU.2012.10359

摘要

本論文探討在連續等候時間限制下，多工作站序列式生產系統之最佳生產控管策略，所謂連續作業等候時間限制指在序列式製程中，任兩連續製程，都必須在限制的時間內完成，若任一製程未在時間限制內完成，將造成製品的品質下降，需要重新加工(Rework)或報廢，又因為此類生產系統皆具有隨機性，故生產控管相當困難，若無妥善的生產控管策略，將造成系統生產成本增加。
現今有許多產業面臨到此類的問題，其中對科技業之影響甚鉅，故本研究將以高科技業中之晶圓製程為例，利用正規化(Uniformization)方法，將連續的時間切割成每一事件發生之層次，例如：新訂單出現、設備完成服務、製品違反時間限制報廢、設備當機和維修完成…等，並利用馬可夫決策過程(Markov Decision Processes, MDP）以及動態規劃方法(Dynamic Programming)，目標為最小化總生產成本，其包含系統之存貨持有成本以及報廢成本；在利用動態規劃控制多工作站序列式系統時，需利用拆解(Decomposition)將多工作站系統拆解為數個小型子系統進行求解，最後利用啟發式演算法，以及求解子系統所得之最佳解，計算出系統之允入控制策略—啟發式動態允入控制（Pairing Heuristic Control, PHC），並利用此策略精確的控制多工作站序列式生產系統。
方法驗證部分將利用離散事件模擬（Discrete Event Simulation, DES），比較啟發式動態允入控制策略和其他文獻策略，在總生產成本、總產出量、總報廢量等績效指標中之優劣，結果顯示在20組實驗設計中，啟發式動態允入控制表現優於其他文獻中策略。

關鍵字

多工作站序列式生產系統 ； 連續作業等候時間限制 ； 動態規劃 ； 拆解 ； 啟發式演算法

並列摘要

Sequential process queue time constraints are important properties in semiconductor and TFT-LCD manufacturing. In these factories, any time violation will result in significant scrap or rework costs. Besides, most process in those industries are stochastic. Without accurate production control in such systems, production yield rate will decrease and production cost will increase.
This research considers sequential process queue time constraints in multi-stage tandems production systems. The objective is to minimize production costs which are include inventory holding cost and scrap cost. We use dynamic programming and uniformization method to model these systems. However, the model is too complex to solve in feasible time. Therefore, we use decomposition and pairing heuristic control (PHC) policy to fix complexity problem.
To verify the robustness and effectiveness of the proposed model, we use discrete event simulation (DES) to compare the performance of PHC with other policies’ in literatures.

並列關鍵字

decomposition ； dynamic programming ； heuristic control ； multiple tandems systems ； sequential process time constraints

參考文獻

[16] Diamantidis, A., & Papadopoulos, C. (2004). A dynamic programming algorithm for the buffer allocation problem in homogeneous asymptotically reliable serial production lines. Mathematical Problems in Engineering, 3(2004), 209-223.

[1] Akcalt, E., Nemoto, K., & Uzsoy, R. (2001). Cycle-time improvements for photolithography process in semiconductor manufacturing. IEEE Transactions on Semiconductor Manufacturing, 14(1), 48-56.

[2] Akkerman, R., Van Donk, D. P., & Gaalman, G. (2007). Influence of capacity-and time-constrained intermediate storage in two-stage food production systems. International Journal of Production Research, 45(13), 2955-2973.

[3] Andradottir, S., Ayhan, H., & Down, D. G. (2008). Maximizing the throughput of tandem lines with flexible failure-prone servers and finite buffers. Probability in the Engineering and Informational Sciences, 22(02), 191-211.

[4] Archibald, T., Black, D., & Glazebrook, K. (2009). An index heuristic for transshipment decisions in multi-location inventory systems based on a pairwise decomposition. European Journal of Operational Research, 192(1), 69-78.