Reset sequences for finite automata with application to design of parts orienters

Abstract

Natarajan reduced the problem of designing a certain type of mechanical parts orienter to that of finding reset sequences for monotonic deterministic finite automata. He gave algorithms that in polynomial time either find such sequences or prove that no such sequence exists. In this paper we present a new algorithm based on breadth first search that runs in faster asymptotic time than Natarajan's algorithms, and in addition finds the shortest possible reset sequence if such a sequence exists. We give tight bounds on the length of the minimum reset sequence. We further improve the time and space bounds of another algorithm given by Natarajan, which finds reset sequences for arbitrary deterministic finite automata when all states are initially possible.