Abstract: Because of the complexity and uncertainty of information systems, it??s hard to use accurate value to represent the object??s attribute value. The interval-value which blurred is used to deal with the issue. This paper introduces dominance relations to establish inconsistent interval-valued fuzzy ordered decision information system. Then, this paper considers distribution reduction and maximum distribution reduction to simplify the expression of knowledge, at the same time, finds out the relationship between them. Moreover, this paper obtains judgment theorem of distribution reduction and maximum distribution reduction and discernibility matrix in the system. It also provides specific method of distribution reduction and maximum distribution reduction in interval-valued fuzzy ordered decision information system. Furthermore, this paper analyzes a specific case about the venture investment and discusses the significance of study on distribution reduction. Finally, this experiment enriches rough set method for interval-valued fuzzy ordered decision information system.

Key words: rough set, interval-value, ordered information system, distribution reduction

摘要： 因信息系统的复杂性和不确定性，对象的属性值难以用精确的数值来表达，而是采用区间形式表示。针对这一问题，对区间值进一步模糊化，并引进优势关系，建立了不协调区间值模糊序决策信息系统。通过分布约简和最大分布约简来简化知识的表达，找出二者之间的关系，得到了分布约简和最大分布约简的判定定理以及可辨识属性集和可辨识矩阵；提供了不协调的区间值模糊序信息系统的分布约简和最大分布约简的具体方法；结合投资风险这一具体案例的求解分析，进一步阐述了对分布约简研究的意义，丰富了区间值模糊序决策信息系统中的粗糙集方法。

关键词: 粗糙集, 区间值, 序信息系统, 分布约简