應用基因演算法於半導體機台備用零件存貨分類模式建置

The Development of Classification Model for Spare Parts in Semiconductor Equipment Using Genetic Algorithm

許原嘉(Yuan-Chia Hsu)

指導教授 ： 林逾先

國立臺北科技大學/管理學院/商業自動化與管理研究所/碩士(2005年)

摘要

半導體機台設備投資成本極高，為了減少機台設備當機對企業所造成的影響，必須使機台設備的失效零件可以及時替換，因此，將半導體機台備用零件存貨實施針對性的分級管制，並正確有效分類將是相當重要的課題。一般的ABC存貨分類法常只考慮存貨的資金佔用比例做為分類準測。然而半導體機台備用零件存貨管理與一般生產物料之特性不同，因此，應採用多屬性的ABC存貨分類法。本研究著重在實作參數組合最佳化的基因演算法程式，並提出適當之編碼方式與交配函式，針對某些已分類的存貨資料檔案，搜尋其多屬性ABC存貨分類法中，各屬性(如存貨單價、年耗用量、關鍵性、前置時間等)的最佳權重參數組合，以學習其最接近的分類模式。
此外，資料包絡法(DEA)可以用來衡量存貨間的相對效率，是一種由實際資料來決定結果的非參數分析方法，本研究先以資料包絡法將不同存貨品項數量的存貨資料檔案分類，以獲得良好之分類結果，再使用基因演算法程式搜尋其若以多屬性存貨分類時，最接近資料包絡法分類模式的權重參數組合，以建立良好的半導體機台備用零件存貨分類模式。實驗資料將分為訓練資料與測試資料進行，實驗結果証明本研究之基因演算法程式，對於資料包絡法分類模式的學習，有良好的成效，相關的實驗結果與參數將一併分析。

關鍵字

多屬性ABC存貨分類法 ； 基因演算法 ； 資料包絡法

並列摘要

Since the investment of semiconductor equipment is large, the invalid parts must be replaced immediately to reduce the cost by machine interruption. The spare parts inventory management is an important issue, and companies must classify the items in inventory and develop effective inventory control policies for these classes. The criterion utilized in the classical ABC classification is the annual dollar usage. However, it has been generally recognized that the traditional ABC analysis may not be able to provide a good classification of spare parts inventory items in practice. The problem of multi-criteria inventory classification has been addressed in this thesis. This thesis designed a program of optimizing a set of parameters that represent the weight of criteria using genetic algorithm (GA), and a new crossover operation and encoding method is proposed. This program can find the weights of criteria and learn the classification rule by classified inventory items.
The data envelopment analysis (DEA) is a non-parameter analysis method that compute the relative efficiency by real inventory data. This thesis try to classify the inventory items by DEA, and then find the optimal parameters that represent the weight of criteria in multicriteria inventory classification. By way of learning the classification rule of DEA, this thesis developed the classification model for spare parts in semiconductor equipment using genetic algorithm. The inventory data in this experiment are divided into train data and test data. It is proved that the GA program in this thesis can learn the classification rule of DEA effectively. The related parameters and experimental results are also reported in this thesis.

並列關鍵字

Multiple Criteria ABC Classification ； Genetic Algorithm ； Data Envelopment Analysis

參考文獻

[3] Charnes, A., Cooper, W. W., Lewin A.Y., Seiford L.M., 1994, Data envelopment analysis: theory methodology and applications. Kluwer: Boston.

[4] Cook, W. D. and Kress, M., 1990, A data envelopment model for aggregating preference rankings, Management Science, 36(11), 1302-1310.

[5] E.A. Silver, R. Peterson, 1985, Decision Systems for Inventory Management and Production Planning, Wiley, New York.

[6] Gengui Zhou ,Hokey Min ,Mitsuo Gen., 2002, The balanced allocation of customers to multiple distribution centers in the supply chain network: a genetic algorithm approach, Computers & Industrial Engineering, 43, 251-261.

[7] Glodberg, D.E., 1989, Genetic Algorithms in Search, Optimization, and Machine Learning, Addison-Wesley, USA.