Abstract: In practice, most of information systems are not only fuzzfied internationally or not, but also based on dominance relations because of various factors. To acquire brief fuzzy decision rules in fuzzy objective information systems based on dominance relations, attribute reductions are needed. The main objective of this paper is to study this subject. The concept of generalized decision is proposed. Three types of reductions are minimal subsets of the condition attributes which preserve the 1ower/upper of generalized decision and generalized decision for all objects in the universe respectively. Practical approaches to compute all three kinds of reductions are then presented by introducing of discernibility matrix and discernibility function. Moreover, a numerical example is employed to substantiate the conceptual arguments.

Key words: attribute reductions, fuzzy objective information systems, dominance relations, generalized decision

摘要： 在现实生活中，许多信息系统不仅是模糊的还是基于优势关系的。在基于优势关系下模糊目标信息系统中引入了广义决策的概念，提出了分别保持下广义决策、上广义决策、广义决策不变的三种不同约简，进一步给出了各属性约简的判定定理和辨识矩阵，提供了在优势关系下模糊目标信息系统属性约简的具体方法。通过实例验证了该方法的有效性，进一步丰富了粗糙集理论。

关键词: 属性约简, 模糊目标信息系统, 优势关系, 广义决策