台灣現在的產業環境代工生產佔了相當大的一個比例,而IC載板產業更是在全球占有一席之地。但隨著全球化的競爭勢必要提升競爭力來跟其他國家競爭,提升競爭力的方法有提升產品品質以及快速交貨這兩個大方向,品質方面比較仰賴技術的提升所以從提升生產速度方面下手。在IC載板的生產流程之中鑽孔製程往往是耗費最多時間的一項製程,一片銅箔基板可以產出約1000顆載板,每顆載板上的孔數雖然不到100,但一片銅箔基板上卻有數萬顆鑽孔,形成一個大型的TSP問題。在此種情況下路徑的安排將會影響到鑽孔的時間,在大量生產下更會放大路徑之間好壞的差距,所以本研究將針對鑽孔路徑問題發展求解方法,達到對IC載板產業的貢獻。 本研究針對IC載板擁有的多孔數特性以及陣列重複排列的特性,使用劉志宏(2009)以及吳世雄(2010)所提出的單顆優化法、單顆順序優化法、轉彎連結優化法以及二顆漸進優化法,改變核心演算法為變動鄰域搜尋法,與使用基因演算法以及螞蟻演算法為核心之情況以及與案例公司案例優化值作分析比較。從研究結果可以發現,在使用相同的優化方法下搭配變動鄰域搜尋法為核心演算法是可以得到較好的效果,而且轉彎連結優化法搭配變動鄰域搜尋法更是能夠得到比案例公司優化值更好的解,在求解速度上也十分優異,因此能夠改善鑽孔路徑達到提升IC載板產業競爭力。
A considerable amount of companies in Taiwan are original equipment manufacturing (OEM) companies. IC substrate industry is among the top in the world. However, in a highly competitive global market, competitiveness is vital to the survival of companies. Competitiveness can be improved through improvement of product quality and rapid delivery. In the manufacturing process of IC substrates, drilling operation is usually to most time-consuming step. Around 1,000 IC substrates can be produced from a single piece of copper clad laminate (CCL). There are less 100 holes that needed to be drilled on each IC substrates. The number of holes to be drilled on a piece of CCL can easily be over ten thousands. Planning a drilling path to cover such great amount of holes significantly affects operation time. Under mass production, its impact is further magnified. Therefore this is the goal of this study to develop more effective methods to solve drilling path optimization problem. Incorporating massive number of holes and the repeating design pattern on a piece of CCL, the variable neighborhood search (VNS) method is applied to combine with the two search algorithms from Liu (2009), single-substrate and two-substrate, and the two algorithms from Wu (2010), single-substrate-sequential and various-link. With the same test instances, it is shown that developed VNS-based algorithms outperform the others. Overall, the various-link method with variable neighborhood search provides solutions with the best quality within reasonable run time.