隨著半導體產業逐漸成熟的今日,帶動了後段產業之晶圓針測與IC封測訂單量的成長,為了滿足客戶需求,企業的生產型態已由過去的單一廠區,已經演變為多廠區的生產模式,而廠區之間需要達成產能資源共享並支援,同時並藉由多座生產成本與製造能力的不同達到最小化的規劃成本。藉由適當的訂單分配來調配產能,避免因產能不足而喪失了企業本身獲利的先機。故本研究欲探討如何在多廠區多階製程的生產環境下,考量各廠區各製程段不同的生產限制與成本,進行多廠區多階製程的訂單分配。 本研究針對多廠區多階製程訂單分配以晶圓針測、IC封裝及IC最終測試段產業為例,利用二元整數規劃方式求解問題模式,考量各廠區各階製程段的產能、產品生產限制、製程能力、上下游製程間供給規劃、單位轉換限制及相關製造成本,將訂單合理分配至最適合的廠區進行生產,並達到多廠區多階製程製造成本、外包成本、延遲成本及閒置成本最小化。 透過實例驗證本研究所提出的多廠區多階製程訂單分配模式,在最小規劃成本之下,可得到最佳的訂單分配結果、各廠區晶圓針測、IC封裝及IC最終測試製程段訂單的產出量、外包量及延遲量,其結果廠區生管人員可進行現場作業排程與派工的依據,亦可提供半導體企業管理者進行決策與判斷及未來業務員接單的參考。
While the semiconductor industry is mature today, it also push forward the orders increasing in the Wafer Probing, IC Packaging and Testing industries. In order to satisfy customer’s demand, companies became multi-plant production from single-plant production and achieve the minimizing of the total cost by sharing the capacity resource. A company can deploy the capacity through applicable order assignment to avoid losing a chance to get great profit. Under these considerations of different production limit and cost, it is important to develop a model to assign orders properly based on the multi-plant and multi-stage manufacturing process. In this research, a multi-plant and multi-stage manufacturing process order assignment model is applied to wafer probing, IC packaging and IC final testing of industry by binary integer programming. According to the constraints which are capacity limitation, production limitation, process capability, the supply between up-stream and down-stream manufacturing process, unit change and manufacturing cost, the orders will distribute to produce in the suitable factories and achieve the objective that is to minimize the total production cost, outsourcing cost, lateness cost and idle cost. This research uses three real examples to verify the feasibility of the model. In the minimizing of the total production, we can find out the optimal order assignment, the Wafer Probing, IC Packaging and IC final testing stage volume of orders producing, outsourcing and order lateness. The result could help management arrange the detail production scheduling and make decisions.