Towards a Theory of Complexity of Regular Languages

We survey recent results concerning the complexity of regular languages represented by their minimal deterministic finite automata. In addition to the quotient complexity of the language – which is the number of its (left) quotients, and is the same as its state complexity – we also consider the size of its syntactic semigroup and the quotient complexity of its atoms – basic components of every regular language. We then turn to the study of the quotient/state complexity of common operations on regular languages: reversal, (Kleene) star, product (concatenation) and boolean operations. We examine relations among these complexity measures. We discuss several subclasses of regular languages defined by convexity. In many, but not all, cases there exist “most complex” languages, languages satisfying all these complexity measures.