在缺乏實際生產的情況下，對一個委外製造商而言，如何從供應商(fabs)所獲得的在製品資料以評估外包生產的晶圓產出時間為一大挑戰。本研究提供委外製造商以改善其產出預測模組來評估外包生產晶圓的產出時程，可藉由規劃模型對缺料部分提出下單建議或拉貨協調。本研究所改善的預測模組為整合五個不同參考歷史線上資訊或預設時間的預測方法，經由移動視窗的方式來進行在製品的產出預測，可適用於各種不同生命週期階段的IC產品。此模組為從兩種選擇機制單選法與決策樹所建議的方法中，擇一合適的預測方法，以對不同產品進行產出時間預測。在研究成果上，對三家大型供應商於六個月內的全部產品而言，本研究可達78%的平均預測產出時間精度。因此，本研究的貢獻即在於可藉由在製品資料，提供委外製造商可有效地預測在不同生命周期階段之IC產出時間。

Without actual production, a manufacturer is challenged how to estimate the wafer-output times of outsourcing products from WIP data which are collected from its contract manufacturers (fabs).
The work assists the manufacturer to enhance a forecast module for estimating output schedules of outsourced wafers. Integrating five different forecasting methods relying on historical production data and default cycle times, this module with the moving window can adapt to various phases of and IC product life cycle. In addition, this module suggests an appropriate method to forecast wafer-output time from two alternatives, i.e., multiple choice and decision tree.
As results, the proposed module can forecast wafer-output times of whole products of three fabs from snapshot based WIP data with average 78% accuracy while comparing to actual wafer-output times, where the products were released during six months. Hence, the contribution of this work is that the wafer-output times of various products in different phases of life cycle can be effectively forecasted.

[1] Voluntary Inter-industry Commerce Standards, Collaborative Planning, Forecasting and Replenishment Version 2.0, Global Commerce Initiative Recommended Guidelines, pp. 10, June 2002.
[2] H.-C. Yang, Y.-L. Chen, M.-H. Hung, and F.-T. Cheng, “Virtual Production Control System,” in the proc. 6th annual IEEE Conference on Automation Science and Engineering, pp. 984–989, 2010.
[3] C. Hill, International Business Competing in the Global Marketplace, 6th ed. McGraw-Hill, pp. 168, 2007.
[4] Chien-Yi Chao, A Grey-Prediction-Based Production Output Estimation Method with Reliance Index, master thesis, 2011.
[5] T. M. Mitchell, Machine Learning, McGraw-Hill, 1997.
[6] K. Muata, and O. Bryson, “Evaluation of decision trees: a multi-criteria approach,” Computer and Operation Research, vol. 31, 2004.
[7] S. Kotsiantis, “Supervised Machine Learning: A Review of Classification Techniques”, Journal of Informatica, vol. 31 pp. 249-268, 2007.
[8] J. R. Quinlan, “Induction of Decision Trees,” Mach. Learn., vol.1, pp.81-106, 1986.
[9] J. R. Quinlan, C4.5: Programs for Machine Learning. Morgan Kaufmann Publishers, 1993.
[10] H. Zimmermann, Fuzzy set theory and its applications, Boston: Kluwer Academic Publishers, pp. 317-319, 2001.
[11] X.-Z. Wang, L.-C. Dong, and J.-H. Yan, "Maximum Ambiguity-Based Sample Selection in Fuzzy Decision Tree Induction," IEEE Transactions on Knowledge and Data Engineering, vol.24, no.8, pp.149-1505, Aug. 2012.
[12] L. A. Zadeh, Fuzzy sets. Information and Control, vol. 8, pp. 338–353, 1965.