Optimal estimation on the order of local testability of finite automata

A. N. Trahtman

Research output: Contribution to journalArticlepeer-review

3 Scopus citations

Abstract

A locally testable language L is a language with the property that for some nonnegative integer k, called the order of local testability, whether or not a word u is in the language L depends on (1) the prefix and suffix of the word u of length k - 1 and (2) the set of subwords of length k of the word u. For given k the language is called k-testable. We improve the upper bound on the order of local testability of a locally testable deterministic finite automaton with n states to 1/2(n2 - n) + 1. This bound is the best possible. We give an answer to the following conjecture of Kim, McNaughton and McCloskey for deterministic finite locally testable automata with n states: "Is the order of local testability no greater than O(n1.5) when the alphabet size is two?" Our answer is negative. In the case of size two the situation is the same as in general case: the order of local testability is Ω(n2). The necessary and sufficient conditions for the language of an automaton to be k-testable are given in terms of the length of paths of a related graph. Some estimates of the bounds on the order of local testability follow from these results.

Original languageEnglish
Pages (from-to)59-74
Number of pages16
JournalTheoretical Computer Science
Volume231
Issue number1
DOIs
StatePublished - 17 Jan 2000

Keywords

  • Finite automaton
  • Language
  • Locally testable
  • Order of local testability
  • Semigroup

Fingerprint

Dive into the research topics of 'Optimal estimation on the order of local testability of finite automata'. Together they form a unique fingerprint.

Cite this