Abstract: There are some problems of choosing evaluation experts in the process of intensive land use evaluation, and the constraint of relative matrix in AHP scoring process. In order to solve these problems, MDS and [K] means clustering are used to select different tendency of the experts; then the improved generation algorithm of judgment matrix is used to make consistency test of experts’ judgment matrix. It can solve the problem that the expert’s judgment matrix may exist inconsistent issues of pairwise comparison item. At last provide effective support construction of evaluation system.

Key words: Analytic Hierarchy Process（AHP）, Multidimensional Scaling（MDS）, sum-product method, judgment matrix, questionnaire survey

摘要： 在对土地集约利用评价过程中，存在着评价专家选择的问题以及AHP打分过程中相对矩阵的约束问题。针对这些问题，应用MDS和[K]均值聚类，筛选不同倾向的专家；利用改进的判断矩阵生成算法，对专家的判断矩阵进行一致性检验，解决可能存在的成对比较项目不一致的问题，为构建评价指标体系提供有效的支持。

关键词: 层次分析法（AHP）, 多维标度分析, 和积法, 判断矩阵, 问卷调查