Abstract: To solve the construction problem of Sinarchy structure supermatrix in Analytic Network Process（ANP）, an improved method based on interval-valued intuitionistic fuzzy set is put forward. From the angle of the inherent attributes and characteristics of the last two layer which has the feedback relation in the Sinarchy, the traditional method has limitations for the construction of influence matrix of alternatives of criterions. So the new thought to solve the above problem has been proposed that should persist the combination of holism and reductionism, also should follow the ideology from the qualitative to the quantitative analysis of the comprehensive integration which advocated by Qian Xuesen. Then this paper introduces interval-valued intuitionistic fuzzy set to reflect expects’ preference in IMAC, instead of adopting point estimates. Applied of a real example, the improved method is validated to be feasible and applicable.

Key words: analytic network process, Sinarchy, interval-valued intuitionistic fuzzy set, supermatrix

摘要： 针对网络分析法（ANP）中Sinarchy结构的超矩阵构造问题，提出了一种基于区间直觉模糊集的超矩阵构造改进方法。从Sinarchy结构中具有反馈关系的最后两层元素集的固有属性及特征出发，分析了采用传统方法构造方案对准则的影响矩阵（IMAC）的局限性，提出了应结合整体论与还原论、并遵循钱学森倡导的从定性到定量的综合集成法作为ANP分析解决问题的思路。通过引入区间直觉模糊集代替传统点估计值反映专家偏好进而重新构造IMAC。具体实例应用表明，该方法针对Sinarchy结构的排序问题具有较好的实际应用可行性。

关键词: 网络分析法, Sinarchy, 区间直觉模糊集, 超矩阵