Parallel Reducts and Decision System Decomposition

Abstract

In this paper, we continue to investigate the properties of parallel reducts. We reveal the drawbacks in the method of decomposing a decision system into a family of decision sub-tables for dynamic reducts, and present a novel method of decomposing a decision system into a series of decision sub-tables for parallel reducts, which also can be applied to dynamic reducts. We prove in theory that the method is effective. Moreover, the method provides a way of calculating the reducts of an inconsistent decision from a family of consistent decision sub-tables, and vice versa.