A portfolio approach to enumerating minimal correction subsets for satisfiability problems

Research output: Chapter in Book/Report/Conference proceedingsConference proceedingpeer-review

Abstract

Even when it has been shown that no solution exists for a particular constraint satisfaction problem, one may still aim to restore consistency by relaxing the minimal number of constraints. In the context of a Boolean formula like SAT, such a relaxation is referred to as a Minimal Correction Subset (MCS). In the context of SAT, identifying MCSs for an instance is relevant in a wide range of applications, including MaxSAT solution approximation and Minimal Unsatisfiable Subset (MUS) enumeration. However, while there are a number of existing approaches to this problem, in this paper we demonstrate how performance can be significantly improved by employing algorithm portfolios. Yet, instead of applying the standard approach of selecting a single solver for the instance at hand, we present a new technique that within a predetermined timeout switches between enumeration algorithms multiple times. Through experimental study, this new approach is shown to outperform any possible optimal portfolio that solely relies on solvers that run uninterrupted for the allotted time.

Original languageEnglish
Title of host publicationIntegration of AI and OR Techniques in Constraint Programming - 11th International Conference, CPAIOR 2014, Proceedings
PublisherSpringer Verlag
Pages368-376
Number of pages9
ISBN (Print)9783319070452
DOIs
Publication statusPublished - 2014
Event11th International Conference on the Integration of Artificial Intelligence (AI) and Operations Research (OR) Techniques in Constraint Programming, CPAIOR 2014 - Cork, Ireland
Duration: 19 May 201423 May 2014

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8451 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference11th International Conference on the Integration of Artificial Intelligence (AI) and Operations Research (OR) Techniques in Constraint Programming, CPAIOR 2014
Country/TerritoryIreland
CityCork
Period19/05/1423/05/14

Fingerprint

Dive into the research topics of 'A portfolio approach to enumerating minimal correction subsets for satisfiability problems'. Together they form a unique fingerprint.

Cite this