|Abstract: ||在全球化和資訊發達的現代，隨著產業競爭的日益劇烈，企業為了維持本身的競爭優勢，必須與供應鏈體系進行密切合作，故如何透過有效的供應鏈運作以便快速地接觸市場、降低庫存風險及成本壓力，快速回應已是企業生存之必要條件。供應商管理存貨(Vendor-Managed Inventory；VMI)就是一種最好的方法，其可透過共用買賣雙方之間的庫存資訊，由供應商負責控制買方的庫存水準，避免買方因庫存太高或不足而造成缺貨，以達到降低庫存與提高供應商供料的服務水準…等目的。
本計畫第一年擬從產業協同之觀點，以台灣TFT-LCD 廠商為例，探討該產業供應鏈中供應商管理存貨之問題。欲探討的問題計有：（1）瞭解台灣不同TFT-LCD 廠商對VMI 的規劃與運作模式有何異同？（2）以五個構面與三十個研究變項作為問卷設計的基礎，藉由對台灣三家主要TFT-LCD 廠商之管理階層進行問卷調查與訪談及資料收集，期待能進一步瞭解VMI 對廠商生產TFT-LCD之績效有何影響？ (3) 探討影響VMI 績效之重要因素及關鍵成功因素有那些？（4）再藉由對各廠商之間績效差異的比較，探討差異之原因，而差異之原因與公司之性質及零件特性有何關連？ (5) 探討各構面間的相互影響程度為何？故本計畫目的在於瞭解台灣TFT-LCD 廠商在實施VMI 前後的績效及存在的關鍵問題，研究結果可提供台灣生產TFT-LCD 之廠商或其它產業廠商導入VMI 之重要參考。
本計畫第二年乃針對第一年計畫的研究結論，找出影響VMI 績效之重要因素與問題，作為建構產業協同供應鏈整合性供應商管理存貨模式的重要基礎與內涵，考慮與實際狀況十分契合之一家核心廠商與多家供應商之二階層VMI 系統，建構數學規劃模式，發展有效的求解方法，並進行實驗設計及求解效能分析，再以台灣三家主要TFT-LCD 廠商從事實務應用研究。
5. 利用結構方程模式(Structural Equation Modeling；SEM)驗證理論模型及找出影響供應商管
理存貨績效的重要因素(預計使用SPSS-AMOS 7.0 軟體)。
6. TFT-LCD 廠商間VMI 績效差異分析。
9. TFT-LCD 廠商實務個案研究。
Businesses operated in a globalization and advanced information technology era and in a drastically competitive environment must closely collaborate with their supply chain partners to maintain and even improve their competitive advantages. Through effective supply chain operations, rapid market penetration, low inventory level, cost reduction, and quick response to customer requirements have become prerequisites for businesses to survive. Vendor-managed inventory (VMI) is one of the best approaches to achieve the goals of increasing customer service level, decreasing stock level and shortage risk, and reducing the impact of the bullwhip effect. VMI integrates operations between suppliers and buyers through inventory information sharing. Suppliers are responsible for tracking and autonomously replenishing their customers’inventories.
In the first year of this project, we will study the case of TFT-LCD companies in Taiwan from the viewpoint of industrial collaboration. The purpose is to explore the VMI problems in the TFT-LCD supply chain. These problems are: (1) to understand the differences among Taiwanese TFT-LCD companies in designing, planning, and operating VMI systems. (2) to prepare a questionnaire based on five research dimensions in which 30 research variables are included. The survey will be conducted by interviewing top managers and gathering questionnaire data from the other managers in three major Taiwanese TFT-LCD companies. It is expected that we can further understand the influence of VMI on the TFT-LCD company’s performance. (3) to find out important factors that significantly affect the VMI supply chain performance and to investigate the critical success factors (CSFs) for VMI. (4) to compare the performance among the three
TFT-LCD companies and discuss the causes that result in the different performance, and then to attribute these causes to the characteristics of the TFT-LCD company. (5) to measure the extent to which the interactions exist among the five research dimensions. In summary, the purpose of this project is to investigate the performance of the three Taiwanese TFT-LCD companies before VMI and after VMI and to reveal the key VMI problems. The results will be beneficial to other Taiwanese TFT-LCD companies or those companies in other industries for their VMI installation and implementation.
In the second year of this project, an integrated VMI model will be formulated for the industrial supply chain according to the research results (i.e., VMI performance influential factors and crucial problems) in the first year. We will consider a two-echelon VMI system consisting of a central TFT-LCD company (e.g., AU Optronics (AUO), Chunghwa Picture Tubes (CPT), or Chi Mei Optoelectronics (CMO)) and several suppliers that extremely meets the practical situation. The mathematical programming model will be solved by an optimization package (e.g., LINGO) for the small-scale problems since it may be NP-hard. For this reason, an effective heuristic method will be developed to solve the large-scale problems, and an experiment and a computational efficiency analysis will be also conducted. Finally, a case of the three well-known Taiwanese TFT-LCD companies will be studied in order to demonstrate the aptness of the proposed model.
This study can be divided into the following nine parts:
1. Set the research purpose and describe the problem.
2. Review the related literature.
3. Design research dimensions (or study’s constructs) that are directly related to the VMI performance.
4. Prepare a questionnaire based on the designed research dimensions and gather statistical data.
5. Use Structural Equation Modeling (SEM) to verify the theoretical model and find out important factors that significantly affect the VMI performance (SPSS-AMOS 7.0 may be used).
6. Analyze the performance differences among the three Taiwanese TFT-LCD companies.
7. Formulate a mathematical programming VMI model and develop an effective heuristic algorithm.
8. Conduct an experiment and a computational efficiency analysis for the proposed heuristic method.
9. Study a case of the three well-known Taiwanese TFT-LCD companies and demonstrate the aptness of the proposed model.