1、 外文翻译 原文 A fuzzy ANP model for supplier selection as applied to IC packaging Material Source: Journal of Intelligent Manufacturing Author: Amy H. I. Lee Supplier selection is an important business decision since a competitive advantage may be gained by the firm with the right suppliers to provide pr
2、oducts/services more effectively and efficiently. If the selection process is carried out correctly, a higher quality, longer lasting buyersupplier relationship is more attainable. The literature on supplier selection has two directions: one is mainly qualitative and focused primarily on methodologi
3、cal aspects, and the other is the introduction of mathematical or quantitative decision-making approaches (Stefanovic and Stefanovic 2008). Mathematical programming (MP) models can be further divided into linear programming, mixed integer programming, goal programming, and multi-criteria decision ma
4、king (MCDM). Fuzzy set theory has also been applied to tackle the impreciseness and uncertainty that are normally found in supplier selection processes (Zadeh 1965). For example, Jain et al. (2004) proposed an evolutionary fuzzy-based approach to tackle the complications related to the qualitative a
5、ssessments of suppliers. Cabrera and Mayorga (2008) provided a fuzzy set application in supplier selection for new product development. The applications of analytic hierarchy process (AHP) and fuzzy AHP (FAHP) to supplier selection have been very popular for the past two decades. Chan and Kumar (200
6、7); Chan et al. (2008) and Lee (2009a) did a brief review of the works that applied these methodologies. For example, Chan (2003) proposed an interactive selection model (ISM) with AHP to handle the supplier selection process systematically and quantitatively. Chan and Chan (2004) developed a suppli
7、er selection model, which adopted the AHP and quality management system principles. Some most recent studies using the FAHP are reviewed here. Hafeez et al. (2007); Wang and Yang (2007) and Yang et al. (2007) developed outsourcing decision models using the integration of the AHP and other methodolog
8、ies. Chan and Kumar (2007) presented a fuzzy extended AHP (FEAHP) approach to select the best global supplier for a manufacturer to supply a critical part used in assembly process. Chan et al. (2007) designed a decision support system (DSS) based on the AHP to solve the supplier selection problem in
9、 the airline industry. Chan et al. (2008) further proposed a fuzzy modified AHP (fuzzy-AHP) approach to efficiently tackle both quantitative and qualitative decision factors involved in the selection of global supplier. Lee (2009a) proposed a FAHP model, which applied fuzzy set theory and the benefi
10、ts, opportunities, costs and risks (BOCR) concept, to evaluate the forms of buyersupplier relationship, such as contractual alliance and joint venture, between a manufacturer and its supplier. Dagdeviren (2008) proposed an integrated approach by applying the AHP and preference ranking organization m
11、ethod for enrichment evaluations (PROMETHEE) for the equipment selection problem. Lee (2009b), based on extent analysis method (EAM) and the BOCR concept, constructed a FAHP model to evaluate various aspects of suppliers. In this paper, we propose an efficient and objective evaluation method to sele
12、ct the best supplier from Taiwans top IC packaging companies. Although the AHP has been a popular method to solve MCDM problems, it can only solve problems with top-to-bottom hierarchies and it cannot accommodate the variety of interactions, dependencies and feedback among elements (Ayag and zdemIr
13、2007). In addition, due to vagueness and uncertainty in decision makers judgments, crisp pair wise comparison cannot accurately capture the judgments. Therefore, this research proposes a fuzzy ANP (FANP) model, which introduces the fuzzy logic into the pair wise comparison in the AHP and the super m
14、atrix concept, to solve the supplier selection problem. By adopting the model, the IC packaging companies can be evaluated, and the final ranking can be obtained. The rest of this paper is organized as follows. In section “Related work”, supplier selection works adopting the AHP and FANP are reviewe
15、d. A FANP model for supplier evaluation is proposed in section “The FANP model for supplier selection”. A case study for evaluating IC packaging companies is presented in section “Application of the model on IC packaging factory evaluation” to examine the practicality of the model. Some conclusion r
16、emarks are given in the last section. Due to the fact that the ANP can consider the interrelationships among elements in a problem setting, the use of the ANP in supplier selection has increased substantially in recent years. The works are briefly reviewed here. Meade et al. (1997) adopted the ANP a
17、nd the utility theory to justify strategic relationships, which are driven by corporate strategies. Sarkis and Talluri (2002) constructed an ANP supplier evaluation and selection model by considering multiple factors including strategic, operational, tangible and intangible measures. Bayazit (2006)
18、and Gencer and Grpinar (2007) also applied the ANP for supplier selection. Shyur and Shih (2006) established a five-step hybrid model, by combining the ANP and the modified technique for order performance by similarity to idea solution (TOPSIS) for vendor evaluation. Demirtas and stn (2008) integrat
19、ed the ANP and the multi-objective mixed integer linear programming (MOMILP) in supplier selection. Through the ANP, both tangible and intangible factors are considered in ranking suppliers, and MOMILP is applied to calculate the optimum quantities among selected suppliers to maximize the total valu
20、e of purchasing and minimize the budget and defect rate. stn and Demirtas (2008a) proposed another integrated model similar to that in Demirtas and stn (2008), but multi-period lot sizing is considered here and the total cost among periods is balanced. Demirtas and stn (2009), under a multi-period i
21、nventory lot sizing scenario and single product with multiple suppliers, further suggested an integrated approach of Archimedean goal programming (AGP) and the ANP. stn and Demirtas (2008b), differentiated from other works by the same authors, integrated the ANP with achievement secularizing functio
22、ns consisting of min max goal programming (MGP) and AGP regarding to decision makers preferences. Wu et al. (2009) proposed an integrated MCDM process by using the ANP and mixed integer programming (MIP) to optimize the selection of supplier and to allocate order quantities if the supplier uses bund
23、ling strategy. Although there were abundant publications about the selection of suitable supplier in the literature using the AHP, FAHP or ANP methods, to the authors knowledge, there were relatively few works that applied the FANP to tackle the supplier selection problem. nt (2009) developed a supp
24、lier evaluation approach based on the FANP and fuzzy TOPSIS methods to help a telecommunication company in Turkey in selecting the most appropriate supplier. However, this research only considered six criteria and no sub-criteria at all. In addition, a matrix manipulation approach, instead of the su
25、per matrix approach, was adopted. The matrix manipulation approach, proposed by Saaty and Takizawa (1986) and adopted by Lee and Kim (2000), Karsak et al. (2002), Chung et al. (2005b) and others, can solve a simple network which is very similar to a hierarchy. The only difference between a hierarchy
26、 and this type of network is that criteria are independent under a hierarchy but there is dependence among criteria under a network. The matrix manipulation approach has some limitations, including the adoption of only two clusters and the use of simplified matrix multiplications (Raharjo et al. 200
27、8). The computation which uses simple matrix multiplications implies that the approach may not provide a stabilized solution. Another type of ANP applies a general super matrix approach that can tackle more complicated problems and has been used in various fields (Saaty 1996; Chung et al. 2005a; Aya
28、g and zdemir 2009). Therefore, for a more complex network, the super matrix concept must be adopted. Lin (2009) proposed a method with two stages; first is to combine ANP with fuzzy preference programming (FPP) for the selection of top suppliers, and second is to apply multi-objective linear program
29、ming (MOLP) to facilitate optimal order allocation. Although the super matrix approach was used in the work, the network was very simplified with only four criteria and no sub-criteria at all. A recent study by Razmi (2009) designed a decision support system to evaluate and select suppliers using FA
30、NP with the super matrix approach. Wong (2010) proposed a decision support system (DSS), which incorporated FANP and preemptive fuzzy integer goal programming (PFIGP), to select logistics outsourcing. Again, both models were too simplified because only six criteria, with no sub-criteria at all, were
31、 considered. A complicated supplier selection problem has to consider various criteria and sub-criteria; otherwise, the model simply cannot lead to a correct decision making. 译文 模糊层次分析法在集成电路封装供应商选择中的应用 资料来源： 智能化制造业杂志 作者：艾米李 供应商的选择是一项重要的商业决定，适合企业发展的供应商提供的产品或服务更有效和有效率就更能使企业在竞争中取得优势。如果供应商选择过程被证明是合适的，那么
32、更高质量、更持久的采购商和供应商的关系就更容易实现。 关于供应商的选择的文献主要从两方面阐述：一种是定性的方法，另一种是数学或定量决策方法（ Stefanovic 和 Stefanovic 2008）。数学规划的模型可以 进一步分为线性规划，混合整数规划，目标规划，多准则决策。模糊集理论也被应用到处理不精确和不确定性的供应商选择流程中，通常在供应商选择流程（扎德 1965）中可以发现。例如，耆那教等（ 2004）提出了一个进化模糊的方法来解决供应商的质量评估。卡布雷拉和马约尔加（ 2008）利用模糊层次分析法对新产品开发的供应商选择进行具体应用。 层次分析法和模糊层次分析法对于供应商选择的应用
33、在过去的二十年里很受欢迎。陈和 Kumar（ 2007 年）； Chan 等（ 2008 年）和李（ 2009）对于这两种方法的应用做了一个简单的回顾。例如 ，陈（ 2003）提出用层次分析法建立了一个交互式的选择模型来进行系统的、定量的供应商选择。陈和生（ 2004）开发了一个供应商选择模型，主要采用层次分析法和质量管理体系的原则。这里还有一些最近关于模糊层次分析法的研究。哈菲兹等（ 2007），王和杨（ 2007）利用层次分析法和其他方法的整合，开发了外包决策模型。陈和 Kumar（ 2007）提出了一种扩展模糊层次分析法的方法来为制造商选择一个提供重要装配过程的最优的全球化供应商。陈等人
34、（ 2007）利用层次分析法设计了一个决策支持系统，以解决航空业供应商的选择问题。陈 等人（ 2008）进一步提出了模糊层次分析法可以有效地解决以定量和定性的决策因素进行全球供应商选择。李（ 2009）提出了模糊层次分析法模型，应用模糊集理论与福利，机会，成本和风险的概念来评价买方与卖方的关系，如合同联盟与合资企业之间，制造商和供应商之间。 Dagdeviren（ 2008）所提出的综合办法，通过利用层次分析法解决设备选型问题。李（ 2009），凭借程度分析法和福利，机会，成本和风险的概念，构建了一个模糊层次分析模型来评价多准则供应商。 在本文中，我们提出了一个有效和客观的评价方法为台湾最大的
35、集成电路封装企业选择最佳供应商。虽然层次分析法已成为流行的方法解决多目标决策问题，但它只能解决高端到低端层次的问题，不适合多种影响因素，附属因素和反馈要素的综合（绫 G 和 zdem 2007）。此外，由于模糊性和决策者的判断不明朗，清晰的两两比较不能准确地被判断。因此，本研究提出一个模糊层次分析法模型，它主要介绍模糊逻辑和超级矩阵的概念，解决供应商的选择问题。通过采用该模型，集成电路封装公司可以被进行评估，并通过最终排名可以得到。 由于网络分析法可以考 虑各问题设置之间的相互关系，近几年，网络分析法在供应商选择中已经快速成长。具体的研究主要有米德等（ 1997 年）通过网络分析法和战略关系的
36、理论证明企业战略驱动。 Sarkis 和 Talluri（ 2002 年）建立了一个基于网络分析法的供应商评估和选择模型，通过考虑多种因素包括战略，实际操作，有形和无形的措施，多因素分析和选择模型。 Bayazit（ 2006）和 Gencer 和 Grpinar（ 2007）也利用网络分析法进行供应商选择。 Shyur 和Shih（ 2006）建立了一种包含五个步骤的混合模式，结合网络分析法，相似 性评价理念的解决方案（ TOPSIS 法）进行供应商选择。 Demirtas 和 stn（ 2008）将网络分析法和多目标混合整数线性规划结合在一起进行供应商选择。通过网络分析法，所有有形和无形因
37、素被考虑在一起进行供应商优先选择排名。多目标混合整数线性规划被应用于计算选定的供应商的最佳数量，以最大限度地提高采购的总价值，并减少预算和缺陷率。 stn 和 Demirtas（ 2008a）提出了另一种集成模型，该模型类似于 Demirtas 和 stn（ 2008 年）提出的，但这里将周期批量考虑在内，各时期的总成本是平衡的。 Demirtas 和 stn（ 2009）根据周期批量库存情况和一种产品的多个供应商，进一步提出了阿基米德目标规划（ AGP）和网络分析法的综合办法。 stn 和 Demirtas（ 2008b），将同一个作者关于网络分析法的不同研究整合起来，利用极大极小目标规划和
38、阿基米德目标规划法分析有关决策者的偏好。吴等人（ 2009）提出了利用网络分析法进行多目标决策过程和混合整数规划的供应商选择优化和如果供应商使用捆绑战略，就分配订单的数量。 虽然已有大量的文献表明关于供应商的选择通过使用层次分析法，模糊层次分析法或网络分析法 的有很多，但应用 FANP 解决供应商选择的问题相对较少。 nt（ 2009）以模糊层次分析法和 TOPSIS 方法为基础，开发了一种供应商评价方法，以帮助土耳其一家电讯公司选择最合适的供应商。然而，本研究只考虑六个准则，没有细化准则。此外，采取一个矩阵操作方法代替超级矩阵法。矩阵操作方法是由 Saaty 和 Takizawa（ 1986
39、）提出，并被李和金（ 2000）， Karsak（ 2002 年），钟等（ 2005 年 b）和其他人所采用，可以解决一个简单的网络，和层次结构非常相似。唯一的区别是这种网络准则在每个等级下是独立的， 但又依赖于整个网络。矩阵操作方法有一些限制，只有两个组比较，通过矩阵乘法的简化使用（ Raharjo 等 2008 年）。它使用简单的矩阵乘法意味着该方法可能不能提供一个稳定的解决方案。另一种类型的网络分析法应用超级矩阵的方法，可以解决更复杂的问题，并已在许多领域中使用（ Saaty 1996; Chung 等2005 年 a; Ayag 和 zdemir 2009 年）。因此，对于一个更复杂的
40、网络，必须采用超级矩阵。林（ 2009）提出了一种两个阶段的方法，第一是结合模糊层次分析法和优先规划程序选择最佳供应商数量，第二是应用 多目标线性规划（ MOLP），以促进最佳的订单分配。虽然超级矩阵法已在选择中使用，但该网络非常简单，只有四个准则，没有细化准则。 Razmi（ 2009）在最近的研究中，设计了一个决策支持系统，使用超级矩阵法进行评价和选择供应商。黄（ 2010）提出了一个决策支持系统，其中模糊层次分析法和模糊整数目标规划，选择物流外包。同样，这两种模式过于简化，因为只有六个准则，没有细化准则。一个复杂的供应商选择的问题要考虑各种准则和细化准则，否则，该模型根本不能做出正确的决策。
