Abstract
In order to divide a decision system into parts, and to simplify decisions of the decision system in sub-systems, formulas are defined in the decision system. By the formulas, granules and granular computing are introduced. This brings about a connection between granular computing and decision systems. Therefore, the condition partition corresponding to a decision system is obtained by granular computing and condition attributes. Also, on the basis of granular computing, the decision system is divided into sub-systems, each of the subsystems is based on a block in the condition partition. So, a decision can be transferred from the decision system to a subsystem, which has simplified the decision process. Furthermore, to deal with the incomplete decision, a subsystem can be subdivided into smaller systems, so that the incomplete decision is decidable in a smaller system. Making a general survey of this paper, the data processing method is connected with granular computing.