Towards adaptive calculus of granules

An importance of the idea of granularity of knowledge for approximate reasoning has been stressed in Pawlak (1997) and Zadeh (1966, 1997). We address here the problem of synthesis of adaptive decision algorithms and we propose an approach to this problem based on the notion of a granule which we develop in the framework of rough mereology. This framework does encompass both rough and fuzzy set theories. Our approach may be applied in the problems of approximate synthesis of complex objects (solutions) in distributed systems of intelligent agents.