Class AbstractSatExtensionReasoner

All Implemented Interfaces:
ModelProvider<Argument,DungTheory,Extension<DungTheory>>, PostulateEvaluatable<Argument>, QualitativeReasoner<DungTheory,Argument>, Reasoner<Boolean,DungTheory,Argument>
Direct Known Subclasses:
SatCompleteReasoner, SatStableReasoner

public abstract class AbstractSatExtensionReasoner extends AbstractExtensionReasoner
Uses a provided SAT solver to solve reasoning problems in AAFs.
Author:
Matthias Thimm
  • Constructor Details

    • AbstractSatExtensionReasoner

      public AbstractSatExtensionReasoner(SatSolver solver)
      Instantiates a new reasoner that uses the given SAT solver
      Parameters:
      solver - some AT solver
  • Method Details

    • getModels

      public Collection<Extension<DungTheory>> getModels(DungTheory bbase)
      Description copied from interface: ModelProvider
      Returns a characterizing model of the given belief base
      Parameters:
      bbase - some belief base
      Returns:
      the (selected) models of the belief base
    • getPropositionalCharacterisation

      public PlBeliefSet getPropositionalCharacterisation(DungTheory aaf)
      Creates a propositional representation of the set of labelings of the given Dung theory that are consistent with the given semantics. This means that for every argument A in the theory three propositions are created: in_A, out_A, undec_A. For every attack A->B the formula "in_A => out_B" is added to the belief set. Depending on the actual semantics further propositional formulas are added. For example, for any admissable semantics and unattacked argument A, the constraint "\top=>in_A" is added; another constraint added for admissable semantics is, given any argument A and attackers B1...BN, add the constraint in_A => out_B1 ^ ... ^ out_BN.
      Parameters:
      aaf - a Dung Thery
      Returns:
      a propositional belief set.
    • getModel

      public Extension<DungTheory> getModel(DungTheory bbase)
      Description copied from interface: ModelProvider
      Returns a single (dedicated) model of the given belief base. If the implemented method allows for more than one dedicated model, the selection may be non-deterministic.
      Parameters:
      bbase - some belief base
      Returns:
      a selected model of the belief base.