Abstract: Granular Computing （GrC） is theory which solves problems with imprecise，partial true，fuzzy and unnumbered information. It is conformity and transcendency of this kind of theories such as rough set，fuzzy set，computing with words，quotient space and interval computing etc. It covers all theories，methodologies，techniques，and tools which make use of granules. It takes proper approximate answer instead of optimal answer and it is supplement and development of the precise math. This article introduces the content of Grnular Computing，explains the feasibility and necessity of studying Grnular Computing，explores basic problems of Grnular Computing，shows some main theories of Grnular Computing and their comparison. Also it proposes Grnular Computing’s further investigation objects，for example，using defined mode for references in Granular Computing.

Key words: Granular Computing（GrC）, granule, computing with words, Rough Set, quotient space

摘要： 粒计算是解决不精确、不完整、模糊以及海量信息问题的理论，是粗糙集、模糊集、词计算、商空间及区间计算等理论的整合和超越，覆盖了所有有关粒的理论、方法、技术和工具的研究。它应用合适的近似解代替最佳解，是对精确求解问题的数学的补充和发展。阐述了粒计算的内涵，研究了粒计算的可行性和必要性以及其基本问题;综述了粒计算的研究现状、主要理论及其比较，并对粒计算进一步发展的方向提出了诸如借鉴可定义方式等设想。

关键词: 粒计算, 粒, 词计算, 粗糙集, 商空间