供应商选择问题中的专家权重多属性决策方法-运筹学与控制论专业论文.docx

摘要 摘要 摘要 合理地选择供应商能使企业降低成本，提高竞争力，因此如何选择供应商是 供应链管理中的一个重要问题。对待选供应商的考虑因素即属性，如质量、价格、 交货时间、风险等是多属性决策问题。决策信息包括两个方面的内容：专家权重 和属性值，其中专家权重的确定是多属性决策中的一个重要研究内容。 本文围绕多属性决策问题中的逼近理想解法和投影法对供应商选择问题进行 研究。首先比较了可能度的两种定义，介绍了基于可能度的区间数排序方法等相 关知识。 其次针对区间型多属性决策问题，在传统 TOPSIS 方法的基础上，引入了矩阵 相似度的概念，通过比较加权规范化决策矩阵与区间型理想解的相似度值来计算 专家权重，提出了基于矩阵相似度的 TOPSIS 法。通过算例从多个角度将本文方法 与传统 TOPSIS 方法及改进的 TOPSIS 法进行比较，表明了本文方法的有效性。 最后通过对加权规范化决策矩阵进行两两比较得到相似度矩阵，再对相似度 矩阵每行或每列求和，计算出粗权重，粗权重代表的是专家权重的粗略值，给出 了区间型理想矩阵的概念。与区间型理想解相比较，区间型理想矩阵更加精确。 进 一步提出了基于矩阵相似度的投影法，通过算例显示了该方法的优越性。 关键词：供应商选择；多属性；矩阵相似度；TOPSIS 法；投影法 Ab Abstract Abstract Rational selection of suppliers can help the enterprise to reduce cost and improve competitiveness. Consequently, how to select the suppliers in a supply chain is significant. In the supplier selection problem, the factors such as quality, price, delivery time and risks etc. is a multiple attribute decision making problem. Decision making information includes the expert weights and the attribute values, which the expert weight is an important research content of multiple attribute decision making. In this paper, the traditional TOPSIS and projection method in multiple attribute decision making are discussed for supplier selection problem. Firstly, two kinds of definition of possibility degree are compared, and related knowledge such as ranking method of interval numbers based on possibility degree are also covered. Secondly, for the interval multi-attribute decision-making problem, the concept of matrix similarity is introduced based on the traditional TOPSIS method. By comparing the similarity values of weighted standardization decision matrix and the interval ideal solution, the weight of experts is obtained. The TOPSIS method based on the similarity of matrices is proposed. The multiple perspectives comparison of traditional TOPSIS method and the improved one are given. Examples show the validity of the proposed method. Finally, similarity matrices are obtained by comparing weighted s