隨著全球化的影響,供應鏈體系日趨龐大,商業環境的變化迅速及企業競爭也日漸激烈,在這種情況下,使得供應鏈管理變得更加複雜。供應鏈的作業管理主要區分為四大領域,包括規劃、生產,採購及配銷,對採購管理而言,供應商的選擇及訂單分配問題是供應鏈管理成功的重要關鍵因素。從過去相關的研究中可以得知,以往訂單分配主要在探討二階層集中式的供應鏈而較少討論多階層分散式的供應鏈。由於分散式供應鏈內的組織都為分散式的獨立個體且根據自身的已知資訊自主決策權。因此,對於多階層分散式供應鏈而言,挑選合適的供應商和最佳化其訂單分配以提升供應鏈的績效相對困難許多。為解決上述問題,本文採用最常使用的供應商選擇指標並以該指標為基礎發展多階層分散式供應鏈的訂單分配決策模型。此外,我們修改了frontier portfolios search演算法並且應用修改後的演算法來幫助企業最佳化其供應鏈的訂單分配。為了評估該演算法之效益,本研究進行了三種實驗,實驗的結果證實利用本研究所建立的演算法,不僅可改善整體供應鏈在獲取訂單之能力並能最佳化訂單分配以最小化採購成本。本研究所提出之演算法能作為未來在分散式供應鏈架構下,實作代理人系統以協助廠商進行最佳訂單分配之決策。
With the impact of globalization, supply chain systems are growing in size, business environment is changing rapidly, and rivalry among established companies is intensifying. Under such circumstances, it has become a rather complicated task to manage supply chains. The management of a supply chain covers four major activities, including planning, production, procurement, and distribution. As far as procurement management is concerned, the most critical factors are supplier selection as well as order allocation. After a review of related literatures, we found that prior studies focused their attention solely on order allocation problems encountered by 2-teir centralized supply chains rather than those encountered by multi-tier decentralized ones. For the latter, how to select qualified suppliers, optimize their allocation of order quantities and enhance supply chain performance is much more difficult in that companies within the decentralized supply chains are all independent organizations which make their own decisions without referring to any others. To address the aforementioned problem, this study first adopts the most popular criteria for supplier selection, and based on the criteria, a decision model for companies' order allocation in multi-tier distributed supply chains is further developed. Moreover, to help supply chain companies optimally allocate their orders, an algorithm, modified from the frontier portfolios search, is proposed here. To evaluate the performance of the algorithm, this study carries out three experiments. The experimental results show that the proposed algorithm can enhance supply chain companies' abilities to not only secure orders but also minimize procurement costs in fulfilling orders. The proposed algorithm can be used in the future to implement a multi-agent system for supply chain companies in allocating their orders.