Research Of Granular Computing And Extension Of Variable Precision Rough Set Theory Based On Pansystems Theory

Rough sets theory was first introduced by Pawlak in 1982, and it is a new method to research unintegrated, uncertain knowledge as well as the expression, learning, induce of data. Granular computing (GrC) is a new hot issue rising in artificial intelligence. The foundation and development of rough sets greatly influence and impulse the foundation and development of GrC. The development of rough sets provides the research of GrC with basic theory and application model. This paper studies rough sets theory and GrC under the direction of pansystems theory, and makes an extension to current studies from pansystem theory:Firstly, we discuss equivalence relation and tolerance relation from two relations (body-shadow relation & whole-part relation) of pansystems theory under the direction of the pansystems view of generalization of pansystems theory, and extend more, we put forward maximum-tolerance relation. Similarly, we extend the membership relation and inclusion relation of standard set theory to the generalized membership relation and generalized inclusion relation, and then we put forward the majority membership relation and majority inclusion relation, therefore, we import generalized variable precision rough set model based on pansystems theory, which extends the conventional variable precision rough set model, and we also discuss the contacts between it and other rough set models, in a mostly certain extent, it perfects the concept of approximate space and extends the applications of rough sets.Secondly, we educe difference-identity principle by studying the correlative pansystems theory under the pansystems view of system and the pansystems view of difference and identity, and then base on that, we analyze the correlative GrC by thechoice of rough set model------one of the GrC models. We give a qualitative andquantitative analysis of granule from a new angle and put forward that granule is based on class, granule is determined by difference-identity principle and a rough set model of GrC based on difference-identity principle, moreover, we classify granule as two classes, absolute-granule and relative-granule, from the pansystems view of difference and identity, and then we import a new concept, called granular-capacity, to quantify granularity, at the same time, we elementarily extend the research of granule to the covering of universe. Therefore we understand the essence of granule qualitatively and quantitatively from the view of system to achieve more excellentclassification.Finally, we adopt a mass of simple examples to adjectively explain the correlative theory, and combine theory with practice to hope more excellent academic value and applied value.