关键词: 多粒度, 时间序列, 决策演化, 同源演化

Abstract: At present，most of the research applications of rough set theory remain in the field of static tables，but as the data in the real world is constantly increasing and updating，there are apparent limitation when dealing with the changing data.Multi-granularity time series is introduced in this paper based on the classical rough set theory.The relevance among decisions which are produced by each granule is studied and the relevant granular decision evolution model is built after the division of decision-making information system.Finally the feasibility of homologous evolution is formulated.

Key words: multiple granularity, time series, decision evolution, homologous evolution