A sufficient condition to polynomially compute a minimum separating DFA

Abstract

The computation of a minimal separating automaton (MSA) for regular languages has been studied from many different points of view, from synthesis of automata or Grammatical Inference to the minimization of incompletely specified machines or Compositional Verification. In the general case, the problem is NP-complete, but this drawback does not prevent the problem from having a real application in the above-mentioned fields. In this paper, we propose a sufficient condition that guarantees that the computation of the MSA can be carried out with polynomial time complexity.

An Experiment in Learning the Language of Sequence Motifs: Sequence Logos vs. Finite-State Machines

“…We therefore turned to algorithms for inferring FSMs from sets of positive and negative examples. This is a well-studied NP-hard problem [3,1] (see, e.g., [4,11] for recent discussions) for which there are several practical heuristics. We implemented a version of Evidence-Driven State Merging (EDSM) [10], modified such that it tries to merge only states at the same distance from the initial state, so the resulting FSM should be easier to understand.…”