Cybernetics And Systems Analysis logo
Editorial Board Announcements Abstracts Authors Archive
Cybernetics And Systems Analysis
International Theoretical Science Journal
UDC 519.854

Mikhailyuk V.A.,1 Sergienko I.V.2
1V.M. Glushkov Institute of Cybernetics, National Academy of Sciences of Ukraine, Kyiv, Ukraine,
e-mail: mikhailyukvictor2@gmail.com
2V.M. Glushkov Institute of Cybernetics, National Academy of Sciences of Ukraine, Kyiv, Ukraine,
e-mail: aik@public.icyb.kiev.ua.

REOPTIMIZATION OF CONSTRAINT SATISFACTION PROBLEMS WITH APPROXIMATION RESISTANT PREDICATES

// Kibernetika i sistemnyj analiz. 2012. Vol. 48, N 1. P. 89–104.

Abstract. If k=O(log n) and a predicate Р is approximation resistant for the reoptimization of problem Max-EkCSP-P under insertion of a truth-value in the predicate and some constraint, then there exists a polynomial approximation algorithm with the ratio q(P)=1/(2-d(P)), where d(P)=2-k|P-1(1)| is a threshold «random» approximation ratio of Р. The approximation ratio q(P) is threshold. Refs: 23 titles.

Keywords: reoptimization, C-approximation algorithm, discrete Fourier analysis, PCP theorem, approximation resistant predicate.



FULL TEXT

© 2019 Kibernetika.org. All rights reserved.