Solutions of Maximal Compatible Granules and Approximations in Rough Set Models

This paper emphasizes studying on a new method to rough set theory by obtaining granules with maximal compatible classes as primitive ones in which any two objects are mutually compatible, proposes the upper and lower approximation computations to extend rough set models for further building multi-granulation rough set theory in incomplete information systems, discusses the properties and relationships of granules and approximations, designs algorithms to solve maximal compatible classes, the lower and upper approximations. It verifies the correctness of algorithms by an example. Keywordsincomplete information system; rough set model; maximal compatible class; algorithm; multi-granulation