運用類神經於台灣製造業供應鏈分層與分群之研究

The Stratification and Cluster of Supply Chain by Neural Network for the Manufacturers in Taiwan

潘青佑(Ching-Yu Pan)

指導教授 ： 顏憶茹

國立臺中科技大學/資訊與流通學院/流通管理系碩士班/碩士(2012年)

摘要

現今台灣產業環境競爭逐漸增加，技術快速進步與變化，使得企業主必須面對快速變化與競爭激烈的市場環境。企業主期盼提升公司在同業間的競爭優勢，與供應鏈上下游進行協商合作，達到有效的供應鏈管理。隨著台灣製造業發展與業態的差異，政府較少針對個別產業提出強而有力的長期規劃與策略執行，無法有效針對個別產業進行有效率的資源分配與管理。
本研究主要透過模型建立，進行台灣製造業供應鏈分群與分層。歸納並發展台灣製造業供應鏈分群與分層策略及營運改善方針。首先使用因素分析縮減變數，再經由自組織映射網路建立台灣製造業分群模型，並使用變數的平均值標準差進行命名，分群結果命名分為「一般型」、「快速回應型」、「價格主導型」三個群體。
接著透過倒傳遞網路建立分層KPI模型，產生各群體的敏感度分析。最後以存貨精確度建立分層規則，結合敏感度分析歸納與發展各群體與各層次的策略與營運改善方針。

關鍵字

供應鏈管理 ； 自組織映射網路 ； 倒傳遞網路 ； 類神經網路

並列摘要

Today, Taiwan's industrial competition is gradually increasing due to technology advances and rapid changes. Business owners have to face a rapidly changing and highly competitive market. In order to have a more competitive edge in the industry, firms need to consult and cooperate more effectively with the upstream and downstream supply chain. In the development of manufacturing in Taiwan, many different industrial patterns have arisen. However, the government has rarely provided a strong and powerful long-term planning and implementation of strategies as per the needs of individual industry. Hence, there has been lacking an efficient resource allocation and management within the industries.
In this study, we propose a model to improve the stratification and clustering of Taiwan's manufacturers. First, factor analysis is used to reduce the variables. Second, a self-organizing map is employed to establish the clustering model for Taiwan's manufacturers. Based on the mean and standard deviation of the variables, the three groups are named as the general type, the quick response type, and the price-led type, respectively. Third, through the back-propagation network, a KPI model was developed to perform the sensitivity analysis of these groups. Finally, the hierarchy rules were established according to inventory accuracy. The strategies and improvement initiatives for each group of manufacturers were provided.

並列關鍵字

Supply Chain Management ； Self-Organizing Map ； Back Propagation Network ； Artificial Neural network

參考文獻

李品輝(民98)。以類神經網路探討全台蒸發量區域性分類與推估之成效(未出版之碩士論文)。國立台灣大學，台北市。

張丁財與陳家玲(民94)。應用資料探勘於壽險業之客戶分群研究。中華管理學報電子商務專刊，67-74。

林東慶(民96)。以灰色理論和類神經網路預測航空客、貨運量之變化(未出版之碩士論文)。國立成功大學，台南市。

黃弘毅(民91)。智慧型供應商選擇支援系統(未出版之碩士論文)。國立成功大學，台南市。

蔣明晃與毛靖宇(民97)。運用融通倉概念整合金物流模式之研究(未出版之碩士論文)。國立台灣大學，台北市。