| Abstract: | 由於在研究所階段有幸能夠參與實習計畫，因此有機會能實際的接觸到企業內部的產線及環境，而實習的產業別則是半導體產業，在實習期間觀察到了許多產線的可改善現象，因此便決定將所觀察現象與論文做結合，透過論文之觀察結果，提出對企業有實際幫助的相關建議
半導體產業一直以來都是世界上相當重要的產業，其產業鏈也非常完整，上游為IP設計及IC設計業，中游為IC製造、晶圓製造、相關生產製程檢測設備、光罩、化學品等業，下游為IC封裝測試、相關生產製程檢測設備、零組件（如基板、導線架）、IC模組、IC通路等業。而其中的特殊產品，陶瓷基板，在近年來有持續成長的趨勢，成為許多半導體產業所使用的基板材料之一，再加上自己曾經接觸過陶瓷基板相關產業，因此便想利用自身對此產業的了解，來解決在該產業中實際觀察過之情形。
而在半導體產業中，產品的生產絕大多數都是以一個批次(By lot)為生產單位，因此本篇論文也會使用一批來做為生產單位，進而去探討本篇論文之主題，而根據本人所觀察到的產業實際情形，產品生產與批次處理習習相關，由於基板經常有尺寸大小之分，如何安排好每個批次進入機台，將會很大的影響到生產時間，由於不同尺寸產品會有不同的容量，導致會有不同的生產數量，因此找出最佳的批次分配組合，以減少生產上的延遲時間，將是此篇論文之主軸，而由於考慮生產實際情形，機台環境方面選用非相關平行機台，將更加貼近實際情形。
而在論文的實驗部份，將會透過Java Gurobi程式的協助，找出具有最小延遲時間的批次，並在一個批次只能容納96片的限制下，查看此批次內的工件組合，得知何種尺寸的產品搭配在一起能使得延遲時間達到最小，進而使得生產成本降低。而在本論文的最後結論部份則會比較本論文模型運行結果及產業實際資料，比較兩者不同，並且透過其中的相異處制定未來研究方向。
;Because of the Internship experience in semiconductor industry during my graduate studies, having the chance to see the production line and environment, after the observation in the industry, found out some defects that can be improve, so it let me decided to wright out this master thesis, in order to give suggestion to the company by the modeling result of my thesis.
Semiconductor industry is one of the most important industry in the world, the industry chain of semiconductor industry is truly complete. For its upstream industries, contains IP and IC design, midstream contains IC and wafer manufacturing, downstream contains IC packaging and testing. One kind of substrate, called ceramic substrate, was growing up in recent years, became popular material for some semiconductor industries. Because of the experience to observe the product line of ceramic substrate, let me decide to enhance the defects in the production process for ceramic substrate.
In semiconductor industry, it usually use lot as the unit to produce the product, so this thesis will also use lot as the production unit. Batch processing is also highly related to semiconductor industry and the problem is, the substrates have different sizes, choose the way to arrange different size of substrates in the lot will influence the production time, so find out the best way to arrange the substrates in the lot in order to minimize the tardiness of the lot will be the goal of this thesis.
In the final part of this thesis, it will use java Gurobi to run the model and shows the best combination of substrates that will minimize the tardiness. After that, this thesis will compare the modeling result and the true data of semiconductor industry, find out the difference, in order to set up the future goal. |
