以類神經網路研究半導體封裝廠銲線機台選擇問題

A Syudy of Wire Bond Machine Allocation in IC Packaging using Neural Net

劉近興(Jin-Sing Liu)

指導教授 ： 陳建良

中原大學/工學院/工業工程研究所/碩士(2004年)

https://doi.org/10.6840/cycu200400567

摘要

本研究嘗試建立一個類神經網路模型，以處理半導體封裝廠訂單的銲線機台選擇問題。銲線機台是半導體封裝廠中數量比例最高的設備。通常晶圓切割機台、黏晶機及銲線機台的數量比例為1：4：16。選擇銲線機台的績效會嚴重影響半導體封裝廠的生產力與競爭力。

不同的半導體產品因不同的製程與技術限制，會有不同的銲線機設備需求。而且因判斷的複雜性與時間限制，生管人員很難快速、正確的選擇出適合訂單的銲線機型。因此，本研究應用類神經網路處理訂單的銲線機台選擇問題。以專家調查法分析出影響銲線機台選擇的8種變數，並以類神經網路架構為基礎發展出因應訂單的機台選擇原則。

本研究的建議方法實際於個案公司驗證後，銲線機台的稼動率由56%提昇至78%，銲線機台選擇的正確率由74%提升至93%。這些顯著的成效顯示本研究能有效的處理半導體封裝廠焊線機台的選擇問題。

關鍵字

訂單指派 ； 類神經網路 ； 半導體封裝 ； 銲線

並列摘要

This research proposes a neural network (NN) model to assign orders to wire bond equipment of IC packaging plants. Wire bond equipment has the largest equipment percentage in IC packaging plants, as generally the distribution of equipment requirement of wafer cutting, die bond, and wire bond is 1: 4: 16. Therefore, the utilization of wire bond equipment significantly affects the productivity and competitiveness.

Different IC products require different wire bond equipment with different process and technical constraints. It is difficult for a production controller to dynamically allocate orders to equipment due to the decision complexity and time constraints. As a result, this research applies NN to solve the order assignment problem of wire bond equipment. Experts’ decision processes are extracted and converted into eight decision variables. Order assignment rules are developed accordingly based on NN structures.

The proposed approach is used in an IC packaging plant and the utilization of the wire bond equipment increases from 56% to 78%, as the accuracy of order assignment to equipment increases from 74% to 93%. This significant result shows the effectiveness and efficiency of this research.

並列關鍵字

IC Packaging ； Wire Bond ； Order Assignment ； Neural Network

參考文獻

1.Mitchell， D. and R. Pavur， “Using Modular Neural Networks for Business Decisions”， Management Decision， Vol. 40 No. 1 pp. 58-63， 2002.

3.Fernandez-Rodrigues， F.， C. Gonzalez-Martel and S. Sosvilla-Riveroon “On the Profitability of Technical Trading Rules Based on Artificial Neural Networks： Evidence from The Madrid Stock Market，” Economics Letters， 69， pp. 89-94，2000.

4.Hu， M. Y.， Zhang， G. P.， Jiang， C. X.， and Patuwo， B. E.， “A Cross-Validation Analysis of Neural Network Out-of-Sample Performance in Exchange Rate Forecasting，” Decision Sciences， Vol. 30， No. 1， pp. 197-216， 2000.

5.Markham， I. S. and Rakes， T， R. . “The Effect of Sample Size and Variability of Data on The Comparative Performance of Artificial Neural Networks and Regression” ， Computer Operations Research， vol.25(4)，pp.251-263.1998.

7.Kim， S. H.， Park， T. S.， Yoo， J. Y. Park， and G. T.， “Speed-Sensorless Vector Control of an Induction Motor Using Network Speed Estimation”， IEEE Transactions on Industrial Electronics， Vol. 48， No. 3， pp. 609-614， June 2001.