Loughborough University
Browse

Complexities for jumps and sweeps

Download (560.89 kB)
journal contribution
posted on 2022-03-11, 09:39 authored by Szilárd Zsolt Fazekas, Robert MercasRobert Mercas, Olivia Wu
The recently introduced model of one-way jumping finite automata skips over letters for which it does not have defined transitions instead of halting and rejecting as classical machines do. It is known that as an acceptor it is strictly more powerful than classical finite automata. The extra power comes from the ability of temporarily jumping over parts of the input. Here we define classes of machines and their accepted languages when this resource is bounded asymptotically, similar to computational complexity classes. We initiate the study of the gap between the resulting constant and linear jumping complexity classes. We conjecture that there is no intermediate jumping complexity but show that for a restriction of the model where the length of the jumped-over factors is limited, machines with logarithmic jumping complexity exist. We also introduce a measure called sweep complexity in order to get closer to a characterization of the regular language class in terms of one-way jumping machines with limited resources.

Funding

JSPS KAKENHI Grant Number JP19K11815

History

School

  • Science

Department

  • Computer Science

Published in

Journal of Automata, Languages and Combinatorics

Volume

27

Issue

1-3

Pages

131 - 149

Publisher

Institut für Informatik, Justus-Liebig-Universität Giessen

Version

  • AM (Accepted Manuscript)

Rights holder

© Institut für Informatik, Justus-Liebig-Universität Giessen

Publisher statement

Accepted for publication in the Journal of Automata, Languages and Combinatorics. The definitive published version is available at https://doi.org/10.25596/jalc-2022-131. The article may be used for non-commercial purposes only.

Acceptance date

2022-01-19

Publication date

2022-08-01

Copyright date

2022

ISSN

1430-189X

eISSN

2567-3785

Language

  • en

Depositor

Dr Robert Mercas. Deposit date: 10 March 2022

Usage metrics

    Loughborough Publications

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC