基于奇异值分解和一致-可信度的判断矩阵一致性改进

Consistency Improving of Comparison Matrix Based on Singular Value Decomposition and Consistency-Credibility Degree

判断矩阵的一致性改进是层次分析法中的难点。针对该领域研究中存在的不足之处,介绍基于奇异值分解的权重确定方法,提出一个衡量判断矩阵一致性和可信度的综合标准,即一致-可信度,在奇异值分解和一致-可信度的基础上给出一种获得最佳改进判断矩阵的一致性改进方法。

The improvement of consistency of comparison matrix is a difficult problem in Analytic Hierarchy Process. (According) to the deficiencies of the research in this field, a new approach for determining the associated weights using the method of Singular Value Decomposition is introduced, and Consistency-Credibility Degree, an index which is used to (measure) the consistency and credibility of comparison matrix, is proposed. Based on Singular Value Decomposition and (Consistency-Credibility) Degree, a new approach is given to obtain the optimal improved comparison matrix.