Research Of Granular Computing And Correlative Models Based On Pansystems Theory

This paper summarized the developed motivations, the research status of granular computing and research meaning and introduced the main method of granular computing. Based these summaries we studied the resource of knowledge, scale of granulation and the relativity of granular from the philosophy view of Pansystems theory and mathematics facet This paper made an extension to current studies from Pansystem theory.At first, we described the knowledge of rough sets from the generalized systems of Pansystems. Based on this description we studied the resource of relations and the connections between levels of relations and generalized weights and analyzed the cause of uncertainty of knowledge. For the knowledge based on Panweighted nets, we studied the relativity and Pansystems methods to construct a new model about rough sets, which provide the new groundwork for rough set theory.Following, from the combining study about existing specific method, in this paper we studied the classical set theory, fuzzy set and rough set theory from the primary relation -membership relation under Pansystems theory guidance. Comparing the developments of fuzzy set and rough set from classical set, we analyzed the difference and identity of fuzzy sets and rough sets dealing with uncertainty problems. This work established the theory base for further combination of fuzzy set theory and rough set theory.At last, on the basement of philosophy understanding and thinking about knowledge of rough sets, we studied the relativity of scales of granulation using Pansystems relativity. On the other hand we researched the method of granulation from mathematics facet based on collection-scatter view of Pansystems. At the end of this paper, we analyzed the hierarchy of granular and the operators on the hierarchy from the generalized systems view. All the work not only deepened.our understanding about basic issues of granular computing and extended the content of granular computing but also provided a new kind of study thinking and methods for constructing a uniform abstract general model of granular computing.