半導體供應鏈路徑配置之多目標最佳化模式

A Goal Programming Model for Semiconductor Supply Chain Configuration

鄭孟哲(Meng-Tse Cheng)

指導教授 ： 郭瑞祥

國立臺灣大學/工學院/工業工程學研究所/碩士(2004年)

https://doi.org/10.6342/NTU.2004.00463

摘要

隨著外在環境變遷與技術快速進步，半導體供應鏈日趨複雜，競爭加劇。究其原因，半導體供應鏈生產系統遭受許多不確定性，即時的干擾和動態的變異對系統生產週期時間，在製品水準與產品產出率帶來嚴重的衝擊。
在過去的研究中，對於生產優先順序差異化、生產績效指標的衡量、供應鏈生產系統控制與監控、生產系統最佳化配置與半導體供應鏈長期規劃等議題鮮少有人提及。再則，過去簡單的供應鏈模型或是生產規劃軟體以無法有效的分析與配置複雜的半導體供應鏈。
有鑑於此，本研究探討在半導體供應鏈中生產服務優先順序化、生產路徑多樣性與生產路徑控制點差異化下，如何透過半導體供應鏈長期生產路徑設計與配置，有效的達到生產績效指標最佳化，並同時降低生產系統績效指標之變異。首先，藉由模擬資料建構反應曲面模型，來描述供應鏈各控制因子，包括生產服務優先順序、生產路徑與控制點與生產績效指標之間的關係。其次建構有優先權多目標生產規劃模型，透過生產服務優先順序組合、生產路徑組合與控制點組合優化各生產週期時間、在製品水準與顧客達交率，同時降低系統的變異。最後發展一套有效的啟發式演算法，透過降低多維多次規劃問題的複雜度，以片段二次半正定函數，有效而快速的找到近似解。
而本研究透過模擬系統加以驗證，顯示產品生產優先順序配置、產品生產路徑比例配置的不同水準設定，對系統各績效指標表現確實有顯著的影響，且經由本演算法所求得的近似解與直接原目標函數帶入套裝軟體所求的結果所求得的結果十分相似。

關鍵字

半導體供應鏈 ； 生產配置與規劃 ； 生產變異 ； 多項式目標規劃 ； 啟發式演算法

並列摘要

Semiconductor fabrication is itself a very complicated manufacturing process. Its global, cross-company supply chain operations are even more complicated and dynamic that usual planning and scheduling solution have become impossible to employ. As semiconductor supply chain become widespread and the competition pressure is very fierce, the detrimental effects of increasing varieties and variations are magnified in the supply chain. But many important issues, such as differentiability of quality of service, adaptability, controllability and scalability of performance metrics, have not been addressed in the literature. Conventional modeling techniques of supply chain operations are no longer effective for supply chain configuration.
Therefore, the first objective of this research is to use response surface method (RSM) to build up the empirical model to describe the relationship between supply-chain configuration and performance metrics under the influence of different variability sources, an optimal supply chain configuration model including mean and variance of performance measures is formulated as a polynomial goal programming model to accommodate different goal objective. Finally, an efficient solution methodology named as semi-definite quadratic approximation method is developed further to find the most optimal supply chain configuration.
Our result shows that our proposed model can easily be adapted the practices in semiconductor supply chain, and the solution methodology developed in this paper is truly effective in terms of quality of our solution comparing to other heuristic providing by come commercial soft wares.

並列關鍵字

Semiconductor Supply Chain Configurations ； Polynomial Goal Programming Model ； Heuristics

參考文獻

[1] Arjan, B. B., B. Jansen, T. Terlaky and K. Roos, “Sensitivity analysis in (degenerate) quadratic programming,” Econometric Institute Report 30, Econometric Institute, Erasmus University, Rotterdam, The Netherlands, 1996.

[2] Arjan, B. B., B. Jansen and T. Terlaky, “The optimal set and optimal partition approach to linear and quadratic programming,” Econometric Institute Report 30, Econometric Institute, Erasmus University, Rotterdam, The Netherlands, 1996

[3] Auslender, A. and P. Coutat, “Sensitivity analysis for generalized linear-quadratic problems,” Journal of Optimization Theory and Applications. 1983

[5] Boot, J.C.G. “On sensitivity analysis in convex quadratic programming Problems,” Econometric Institute, Rotterdam, Holland, 1963.

[6] Boot, J. C. G. “Binding constraint procedures of quadratic programming,” Econometric, Vol. 31, No. 3 464-498, 1963.