AHP中群组评判的可信度法(Ⅰ)

通过构造各判断矩阵的可信度，将群组评判中的评分算术平均法、评分几何平均法和权重算术平均法推广到更一般的形式     

AHP中群组决策的综合判断

本文从理论上对群体决策过程程的判断矩阵几何平均综合法进行了分析。结论表明, 几何平均综合矩阵的一致性指标要小于各决策矩阵一致性指标的算术平均数。如果判断矩阵的干扰因子是随机的, 文章证明了几何平均综合矩阵将依概率收敛于一个一致性的正互反矩阵, 其排序向量刚好是各判断矩阵的几何平均综合排序向量。无疑这对于在群组决策时几何平均综合判断矩阵法的使用是重要的。     

Two Methods for Deriving Members' Weights in Group Decision Making

The Analytic Hierarchy Process is a powerful technique for group decision making. Both theWeighted Arithmetic Mean Method (WAMM) and Weighted Geometric Mean Method (WGMM) are themost common group preference aggregation methods in AHP. In order to use the WAMM and WGMM, onehas to find the weights to be assigned to the members of the group. This is often a difficult task, especiallyso if the group is large as in the case of public policy decisions. These situations need an objective method toderive members'weights. But a few studies are available in the literature. Based on judgement matrices anderror analyses, this paper presents two practical and efficient methods for addressing such situations. Somenumerical examples are also given.     

On consistency of the weighted arithmetical mean complex judgement matrix

The weighted arithmetical mean complex judgement matrix(WAMCJM)is the most common method for aggregating group opinions,but it has a shortcoming,namely the WAMCJM of the perfectly consistent judgement matrices given by experts canot guarantee its perfect consistency.An upper bound of the WAMCJM's consistency is presented.Simultaneously,a compatibility index of judging the aggregating extent of group opinions is also introduced.The WAMCJM is of acceptable consistency and is proved provided the compatibilities of all judgement matrices given by experts are smaller than the threshold value of acceptable consistency.These conclusions are important to group decision making.     

综合判断矩阵的几个性质

指出了一些文章关于判断矩阵凸组合和Hadamard凸组合一致性性质叙述中的几个小问题,在 已有证明方法的基础上给出了综合判断矩阵的几个性质,对群组决策具有较大的实用性.     

On Consistency in AHP and Fuzzy AHP

The analytic hierarchy process(AHP) is used widely for analyzing decisions made in various real-world applications. Its basic idea is to construct a hierarchy of concepts encountered in a given decision problem and to choose the best alternative according to pairwise comparison matrices given by the decision maker. Under the assumption of fully rational economics, a reasonable decision should be consistent. It becomes an important issue on how to analyze and ensure the consistency of comparison matrices together with the judgments of the decision maker. The main objectives of the present paper are threefold. First, we review the basic idea and methods used to define the consistency and the transitivity of multiplicative reciprocal matrices, additive reciprocal matrices and comparison matrices with fuzzy interval and triangular fuzzy numbers. The existing controversy behind the applications of fuzzy set theory to the AHP in the literature is presented. Second, the consistency of the collective comparison matrices in group decision making based on AHP and fuzzy AHP is further analyzed. We point out that the weak consistency of preference relations with fuzzy numbers in fuzzy AHP and group decision making should be investigated comprehensively. Third, under the consideration of the vagueness in the process of evaluating the judgements, a new concept of fuzzy consistency of comparison matrices in the AHP is given.     

一种基于可能满意度与加权几何平均的一致性改进方法

对层次分析法中判断矩阵的一致性改进方法进行了研究。针对现有调整方法多重视收敛速度而忽略相对原始判断信息偏离的不足,将可能满意度的概念引入改进过程。利用判断矩阵的最大特征值及其F roben ius范数,给出判断矩阵可能度和满意度的定义与计算公式,分别考察一致性改进程度和相对原始判断矩阵的偏离程度,并将两者合并为一个衡量一致性改善效果的综合指标:判断矩阵的可能满意度。利用该指标,并结合常用的加权几何平均改进算法,可以有效地控制不一致判断矩阵的调整力度,在对决策者原始判断信息偏离最小条件下,逐步达到可接受的一致性。最后通过算例对比说明了新算法的有效性。     

A Reexamination of Methods of Hierarchic Composition in the AHP

1　 IntroductionHierarchic composition orpriority synthesis is an importantstep of the Analytic HierarchyProcess ( AHP) .The original method used in the AHP( Saaty,1977) to conducthierarchic composition is the weighted arithmetic mean.Since the weighted a…     

组合判断矩阵的相容性与一致性关系

在相容性概念的基础上,研究了加权几何平均组合判断矩阵的相容性以及相容性和一致性的关系。在相对较弱的条件下,获得了组合判断矩阵不仅与其自身的特征矩阵具有满意的相容性,而且它还与加权几何平均组合排序向量所构成的特征矩阵具有满意的相容性。这为在群组决策中使用加权几何平均向量排序法提供理论依据。最后进行了实例分析,验证了该结论。     

AHP中群组评判的可信度法(Ⅱ)

给出了AHP中群组评判可信度法的理论分析，揭示了该方法的实质，同时基于可信度给出权重几何加权平均法，且通过构造判断矩阵的差异度和相似度，给出了另外一种确定判断矩阵可信度的方法．     

基于区间数群决策矩阵的专家权重确定方法及其算法实现

针对专家偏好信息为区间数群决策矩阵的多属性群决策问题,提出了一种专家权重确定方法并利用自适应迭代算法实现。首先,给出了区间数和专家群决策矩阵的定义。然后,使用加权几何系数法计算专家综合权重,并通过比较专家个体与专家群体决策矩阵的偏差距离计算出专家的客观权重,经过多次迭代后得到稳定的专家客观权重与专家综合权重。最后,实例验证了该算法的可行性与有效性。     

An analytic hierarchy process model of group consensus

In group decision making, a certain degree of consensus is necessary to derive a meaningful and valid outcome. This paper proposes a consensus reaching model for a group by using the Analytic Hierarchy Process （AHP）. It supports people to improve their group consensus level through an updating of their judgments. In this model, a moderator suggests the most discoraant aeclslon manet to update his judgment in each step. The proposed consensus reaching model allows decision makers to accept or reject the suggestion from the moderator. This model ensures that the judgment updating is effective and the final solution will be of acceptable consistency. Finally, a numerical example is given to illustrate the validity of the proposed consensus reaching model.     

Approaches to Improving Consistency of Interval Fuzzy Preference Relations

This article introduces a consistency index for measuring the consistency level of an interval fuzzy preference relation（IFPR）.An approach is then proposed to construct an additive consistent IFPR from a given inconsistent IFPR.By using a weighted averaging method combining the original IFPR and the constructed consistent IFPR,a formula is put forward to repair an inconsistent IFPR to generate an IFPR with acceptable consistency.An iterative algorithm is subsequently developed to rectify an inconsistent IFPR and derive one with acceptable consistency and weak transitivity.The proposed approaches can not only improve consistency of IFPRs but also preserve the initial interval uncertainty information as much as possible.Numerical examples are presented to illustrate how to apply the proposed approaches.     

AHP中群决策判断矩阵的构造

由于在群决策矩阵的构造时,必须考虑每位专家的意见,又要保持决策矩阵具有AHP中判断矩阵的一般特性。为此,针对层次分析法中群决策的不同专家所建立的判断矩阵,利用不同判断矩阵所对应元素的几何平均法,构造出平均判断矩阵,即群决策判断矩阵。并利用严密的数学方法证明了“所构造的矩阵能保持一致性和满意性”的重要结论。该方法运算简单,不仅为AHP群决策判断矩阵的建立提供了一种有效的方法,而且为群决策的研究提供了一定的理论依据。     

群体思维收敛性定量验证

本文证明当专家数足够多的时候,加权几何平均综合判断矩阵与加权算术平均综合判断矩阵都依概率收敛到客观排序向量,从而从数学上解释和验证了群体思维具有收敛性,这对钱学森先生提出的综合集成研讨厅理论具有一定的理论价值.     

一种新的用于群组排序的最小平方逼近法

本文将文献[1,2]所给的最小平方逼近法推广应用于层次分析法中的群决策排序。文中对由多个判断决策者给出的多个不同的判断矩阵,提出通过求解群组判断矩阵最小平方偏差的方法得出判断矩阵的最佳排序和理想综合判断矩阵;通过对群组判断矩阵进行一致性讨论,据此又给出了一种用于群组判断矩阵排序的最小加权平方逼近法。理论分析和应用实例表明,应用最小平方逼近法和最小加权平方逼近法对群组判断矩阵进行排序是可行的和有效的。     

基于纯语言混合几何平均算子的多属性群决策方法

在纯语言加权几何平均(PLWGA)算子和推广的有序加权平均(EOWA)算子基础上给出纯语言混合几何平均(PLHGA)算子,研究了专家权重、属性权重及属性值均以语言形式给出的纯语言多属性群决策问题,提出了一种纯语言多属性群决策方法。最后将该方法应用于解决虚拟企业中的战略合作伙伴选择问题。     

一种不确定型OWGA算子及其在决策中的应用

把有序加权几何平均(OWGA)算子推广到所给定的数据信息均为区间数形式的不确定环境之中,基于区间数两两比较的可能度公式和模糊互补判断矩阵公式,提出了一种不确定有序加权几何平均(UOWEGA)算子,给出了其在应用过程中的具体步骤,并提出了一种相应的集结决策信息的方法。最后通过算例说明了方法的可行性和有效性。     

基于ET-WG和ET-OWG算子的二元语义群决策法

针对一类属性值和权重值均为语言评价信息的多属性群决策问题,提出了一种群决策分析方法.首先,为了便于语言评价信息的集结运算,提出了一些新的集结算子:扩展的二元语义加权几何(ET-WG)算子和扩展的二元语义有序加权几何(ET-OWG)算子,并分析了这些算子的性质.然后提出了一种基于ET-WG算子和ET-OWG算子的群决策方法,其核心是通过语言信息的集结运算,得到各方案的群体综合评价信息,从而根据二元语义信息的比较原则,得到所有方案的排序结果.最后给出了一个实例分析,说明了本文提出方法的可行性和实用性.     

区间直觉模糊幂加权算子的动态多属性决策

为了解决集结算子处理动态多属性决策问题时,现有的区间直觉模糊(interval valued intuitionistic fuzzy, IVIF)加权平均算子未考虑集结数据之间的相互关系、决策结果精度不高的不足,利用幂加权几何平均(power weighted geometric average,PWGA)算子的非线性特性将集结数据之间相互关系联系起来,提出了IVIF PWGA算子的动态多属性决策方法。首先,将实数形式的PWGA算子扩展到区间直觉模糊集(IVIF set,IVIFS),利用数学归纳法证明了数据融合后的综合集结值是区间直觉模糊数(interval valued intuitionistic fuzzy number, IVIFN)的结论。然后,定义了IVIF条件下,处理动态多属性决策问题的PWGA算子。通过动态PWGA算子集结多个时间点的单一集结值得到综合集结值,根据综合集结值的得分函数和精确函数,对各方案排序。最后,通过实例说明了该算法的有效性。