Rough approximations of multi-fuzzy sets

Rough sets have been invented to cope with uncertainty and to deal with intelligent systems characterised by insufficient and incomplete information. This paper proposes a general framework for the study of relation-based multi-fuzzy rough approximation operators. We define rough multi-fuzzy sets and study their properties. An arbitrary multi-fuzzy relation is used to approximate multi-fuzzy sets. A pair of lower and upper rough approximation operators is hence obtained and basic properties of these approximation operators will be examined. By introducing the cut sets of multi-fuzzy sets, we study the connections between multi-fuzzy rough approximation operators and crisp approximation operators. Finally, the multi-fuzzy approximation operators will be defined by axioms.