Techniques for efficient lazy-grounding ASP solving

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Researchers

  • Lorenz Leutgeb
  • Antonius Weinzierl

Research units

  • Vienna University of Technology

Abstract

Answer-Set Programming (ASP) is a well-known and expressive logic programming paradigm based on efficient solvers. State-of-the-art ASP solvers require the ASP program to be variable-free, they thus ground the program upfront at the cost of a potential exponential explosion of the space required. Lazy-grounding, where solving and grounding are interleaved, circumvents this grounding bottleneck, but the resulting solvers lack many important search techniques and optimizations. The recently introduced ASP solver Alpha combines lazy-grounding with conflict-driven nogood learning (CDNL), a core technique of efficient ASP solving. This work presents how techniques for efficient propagation can be lifted to the lazy-grounding setting. The Alpha solver and its components are presented and detailed benchmarks comparing Alpha to other ASP solvers demonstrate the feasibility of this approach.

Details

Original languageEnglish
Title of host publicationDeclarative Programming and Knowledge Management - Conference on Declarative Programming, DECLARE 2017, Unifying INAP, WFLP, and WLP, 2017, Revised Selected Papers
Publication statusPublished - 1 Jan 2018
MoE publication typeA4 Article in a conference publication
EventInternational Conference on Declarative Programming - Wurzburg, Germany
Duration: 19 Sep 201722 Sep 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer
Volume10997 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

ConferenceInternational Conference on Declarative Programming
Abbreviated titleDECLARE
CountryGermany
CityWurzburg
Period19/09/201722/09/2017

ID: 28944955