摘要 在現今競爭激烈的全球市場中,由於顧客需求改變及產品生命週期縮短,使得企業必須把自我重心放在供應鏈管理上,特別是供應商選擇問題。供應商選擇問題在供應鏈管理上是非常重要的一環,因為現今與企業合作的供應商至少也有數十間,但是,因為各間供應商的產品價格、品質、交貨時間等因素不盡相同,挑選合適的供應商作為企業的合作夥伴,已經成為企業想要降低生產成本增加利潤並提升自我在市場上的競爭力,所必須要考慮的問題。故如何挑選供應商已經成為各企業的管理階層所關注的課題。而可靠度理論現今也普遍應用於各個領域,但是,回顧過去文獻中發現,供應商選擇問題尚未有探討關於可靠度方向的議題,由於供應商選擇問題屬於NP-hard問題,故需要從演算法方面著手,本篇提出以可靠度結合遺傳演算法應用於供應商問題。本研究將可靠度定義為供應商零組件供貨正常的機率,例如:供應商零組件可靠度為70%,代表其零組件供貨正常的機率為70%。在不同情境下,探討不只單純以各間供應商零組件的成本作為挑選合作的供應商的依據,也需要考慮供應商零組件有可靠度問題時,供應商選擇前置時間和供應商選擇的成本與可靠度之間有何影響,並對供應商選擇的前置時間和成本增加比例的數據歸納成兩個迴歸公式,比較其預測值與實際數據結果的差異比例。
Now in the competitive global market, customer needs change dramatically and life cycles of the product reduce rapidly so that enterprises pay a lot of attentions on supply chain management issue, for example of supplier selection problem especially. Supplier selection problem is a very important issue on supply chain management. An enterprise at least cooperates with dozens of suppliers. A close and good cooperative relation between the enterprise and its supplier is beneficial to its company long-term profit and make it more competitive in the intense market. The most used criterion of selection is“cost”. However, cost factor is not totally determined by itself. This thesis considers another factor, i.e. reliability of a supplier, into account to examine while choosing a supplier. Since reliability theory is seldom applied in the supply chain management issue. We tried to adopt genetic algorithm with considering reliability on supplier selection problem. Our research investigates the impact on lead time and selected cost of supplier. Finally, we generalize the regression formulation of lead time and cost by reliability of the supplier, numbers of suppliers and key product components for practical purposes.