Research On Multi-Attribute Decision Making Of Interval-Valued Informa-tion System Based On Dominance Relation

Because of existing noise, information defect or preference and so on in the actual decision problem, so the classical rough set theory based on equivalence relation is not practical. At present, some scholars have proposed the dominance relation instead of the equivalence relation, and the classical rough set theory is promoted to the interval-valued information system.This dissertation is researched the interval-valued information system with domination relation. Main tasks as follows:Firstly, it puts forward probability dominance relation under the interval-valued information system and emphasizes its advantages. Based on these, the interval-valued rough set is established, the upper and low approximation sets of this system are constructed, attribute reduction is given by discernibility matrix.Secondly, the rough set theory of the interval-valued target rough set is researched under the probability dominance relation which is raised on previous chapter, accumulate upper and low approximation sets are raised. Accumulate upper and low discernibility matrix are given by identify function. So attribute reduction and useful simplified rules are provided. The research of attribute reduction under the interval-valued information system based on the dominance relation is added. Finally, the rough set theory and rough entropy are mixed, condition rough entropy is raised under the interval-valued rough set, which is to divide the advantage class under the interval-valued rough set, and give out the conditions rough entropy of condition class under on decision class. At last, it is proofed that the conditions rough entropy can accurately reflect the roughness of the information system accurately.