Shortest Characteristic Factors of a Deterministic Finite Automaton and Computing Its Positive Position Run by Pattern Set Matching

Abstract

Given a deterministic finite automaton (DFA) A, we present a simple algorithm for constructing deterministic finite automata that accept the shortest forbidden factors, the shortest forbidden prefixes, the shortest forbidden suffixes, the shortest forbidden words, the shortest allowed suffixes, and the shortest allowed words of the automaton A. We refer to these sets as the shortest characteristic factors of the automaton A. If the given automaton is local, and therefore the language it accepts is strictly locally testable, the sets of its shortest characteristic factors are finite, and these automata are acyclic. Otherwise, they accept infinite languages. This approach simplifies existing methods for the extraction of forbidden factors, allows the extraction of more types of characteristic factors, and also generalizes the extraction for all classes of DFAs. Furthermore, we demonstrate that this type of extraction can be used for a sublinear run of an automaton for certain inputs. We define a positive position run of a deterministic finite automaton, representing all positions in an input string where the automaton reaches a final state. Finally, we present an algorithm for computing the positive position run of the automaton, which utilizes pattern set matching of its shortest forbidden factors and its shortest forbidden or allowed suffixes, provided that the sets are finite. We showcase the computation of the positive position run of a local automaton using backward pattern set matching, which can achieve sublinear time.