運用多目標決策方法評選供應鏈組成夥伴

Abstract

供應鏈組成夥伴評選之議題廣受注目，然選擇對的供應鏈組成夥伴，對高階管理者言，是一項艱難的任務。因為供應鏈組成夥伴之選擇並不是獨立的，乃與其他成員間之互動息息相關，深受決策模式之影響。本論文探討兩個不同選擇供應鏈上供應商之議題。 第一個議題，運用層級分析法 (AHP) 進行供應商評選作業。我們以投票式排序評選模式，即所謂投票式層級分析法 (VAHP) 以取代既有AHP成對比較的方法。此投票式層級分析法區分三個步驟，首先，由每一位決策者針對受評估目標進行排序，以避免兩兩比較方法的不一致性問題；其次，運用線性規劃模式求出排序之權重值；再其次，計算出受評估目標的總得分數，以排列優先順序。 第二個議題，運用多目標二元整數規劃模式，以個別受評單元進行組合評估方式，評選不同組合之供應商。在假設有K個供應商時，則有2^K個不同之受評供應商組合，並應用成本、交期、彈性與品質等四項績效衡量指標，結合資料包絡法 (DEA)，進行多元組合供應商評選。最後，針對落在高效外廓之受評單元，實施敏感度分析。

The issue of supplier selection catches many attentions in supply chain management. The suppliers of the supply chain operate interactively rather than independently, as the output of one organization could be the input of another organization. In this dissertation, we are dealing with two issues of supplier selection in supply chain. The first issue is that a group of decision-makers to rank a set candidates of suppliers. We employ Analytic Hierarchy Process (AHP) for supplier selection. The pair-wise comparison method proposed by Saaty in AHP is substituted by a voting method. The voting method contains three steps. In the first step, each decision-maker ranks the alternatives to avoid the inconsistency that usually appeared in pair-wise comparison method. The second step is to summarize the votes each alternative earned in every rank. The third step is using a linear programming model to determine the weights assigned to every votes in those ranks. Then, the score of each alternative earned is the sum of weighted votes and gets priority of alternatives. The second issue is to select a set of multiple suppliers. There are 2K possible sets of multiple suppliers under selection if there are K supplier candidates. Cost, delivery, flexibility and quality are the four indices used to measure the suppliers’ performance. These four indices are also used to measure the performance of the possible sets under selection. The value of each index of a set is equal to the sum of values of the suppliers in it. We employ data envelopment analysis (DEA) to measure the relative performance of each set of multiple suppliers against the 2^K possible sets with the four indices. Then, we perform sensitivity analysis on each index of candidates and provide different strategies in order to select various suppliers of a supply chain for customers.