薄膜液晶顯示器組立廠主生產排程快速規劃系統之設計

Abstract

薄膜液晶顯示器既輕且薄的特性，近年來受到市場的青睞，廣泛的被應用於顯像產品上，因此各面板組立廠莫不以產出最大化為目標，以搶佔市場。然而，顧客需求隨時間不斷變動的特性，使得規劃幅度內之產品組合與各期之產出目標皆不相同。因此，在此非穩態的生產環境境下，本文提出「主生產排程快速規劃系統」，在因應顧客需求變動之情境，同時滿足市場對生產週期時間之要求，來制定主生產排程。 對於此一課題，本系統以達到各規劃時期之產出目標為原則，建構三個模組來加以因應：「產能推估模組」、「產線配置模組」與「生產週期時間估算模組」。首先，針對組立廠整備時間相當長的特性，在「產能推估模組」裡，本文以最大可換線次數作為定義系統瓶頸之判斷。在定義出系統瓶頸後，「產線配置模組」在滿足各期各產品別之最小需求的條件下，以達到產出最大化為規劃目標，建構線性規劃模式來求解，並據以安排瓶頸工作站的產線配置。其次，在非瓶頸工作站方面，則以減少換線次數與機台產能負荷平衡為原則進行配置，以適時完成各期瓶頸工作站所規劃之產出目標。最後，在「生產週期時間估算模組」裡，依瓶頸工作站的產出速率，估算批量工作站之理想載入批量，並以區段基礎式生產週期時間法（BBCT）之理念為基礎，及產線配置之結果，預估各產品在各期之生產週期時間，並藉此判斷是否超過等候時間限制。 實驗結果顯示，依據瓶頸資源產線配置模式，求解出瓶頸工作站各機台每期產品之加工數量，本文以此安排各產品在各期之投料順序，可達到產出最大化的目標。透過非瓶頸充分配合瓶頸工作站之原則及兼顧同一工作站內各機台負荷平衡之作法，來配置產線及設計投料時點，能使各產品別在各期之產出數量與規劃目標相當一致。在此產線配置模式下，各產品之生產週期時間估算結果亦有相當不錯的準確度，整體而言，本文提出之「主生產排程快速規劃系統」能有效率地安排顧客需求變動情境之主生產排程。

The thin film transistor-liquid crystal display (TFT-LCD) has the light and thin characteristics and is generally applied in display industry, in recent years. In order to seize the market, cell assembly factory sets “maximized throughput” as the target of production planning. However, due to customer demand changed by time, make product mix and throughput target in every planning period are all not the same. So under the non-steady state environment, this thesis proposes Rapid Master Production Scheduling System to deal with the situation of customer's demand variation, and to meet the request for product’s cycle time of the market at the same time. For this issue, this system takes reaching the throughput target each planning period as the principle, and contains three modules: capacity evaluation module, capacity allocation module and cycle time estimation module. First of all, due to the long setup time, in capacity evaluation module, this thesis uses number of setups available as the judgement for defining the system bottleneck. After defining system bottleneck, in capacity allocation module, on terms of meeting minimum demand of every product type, and of reaching maximized throughput as the objective, a linear programming model is proposed to solve the production quantity of each product type in each period. Then, the production line is allocated for the bottleneck workstation, based on the solution. Secondly, for each non-bottleneck workstation, machines are allocated to each product type based reducing setup times and balancing load of machines of the workstation, in order to finish the throughput target in time. Finally, in cycle time estimation module, matched with the production rate of bottleneck workstation, and the ideal loading batch size is determined. Then, the product’s cycle time is estimated so as to judge whether waiting time limit is exceeded or not. Experimental result shows, the quantity of each product type in each period derived by capacity allocation module can be produced through the principle of matching throughput rate of bottleneck and non-bottleneck workstation. Also, the cycle time estimation has good accurity. The target of maximizing throughput thus is achieved by controlling cycle time within the limit. Overall, the proposed system can rapid arrange the MPS efficiently and effectively for the demand variation environment.