Package ai.timefold.solver.core.impl.exhaustivesearch.decider

Class ListVariableExhaustiveSearchDecider<Solution_,Score_ extends Score<Score_>>

java.lang.Object

ai.timefold.solver.core.impl.exhaustivesearch.decider.AbstractExhaustiveSearchDecider<Solution_,Score_>

ai.timefold.solver.core.impl.exhaustivesearch.decider.ListVariableExhaustiveSearchDecider<Solution_,Score_>

All Implemented Interfaces:

ExhaustiveSearchPhaseLifecycleListener<Solution_>, SolverLifecycleListener<Solution_>, EventListener

public final class ListVariableExhaustiveSearchDecider<Solution_,Score_ extends Score<Score_>>
extends AbstractExhaustiveSearchDecider<Solution_,Score_>

Field Summary

Fields inherited from class ai.timefold.solver.core.impl.exhaustivesearch.decider.AbstractExhaustiveSearchDecider

acceptUninitializedSolutions, bestSolutionRecaller, manualEntityMimicRecorder, scoreBounderEnabled, sourceEntitySelector

Constructor Summary

Constructors

Constructor	Description
ListVariableExhaustiveSearchDecider(String logIndentation, BestSolutionRecaller<Solution_> bestSolutionRecaller, PhaseTermination<Solution_> termination, EntitySelector<Solution_> sourceEntitySelector, ManualEntityMimicRecorder<Solution_> manualEntityMimicRecorder, MoveRepository<Solution_> moveRepository, boolean scoreBounderEnabled, ScoreBounder<?> scoreBounder)

Method Summary

Modifier and Type	Method	Description
void	expandNode(ExhaustiveSearchStepScope<Solution_> stepScope)	The method updates the exploration of the solution's space when using a list variable.
boolean	isEntityReinitializable(Object entity)
boolean	isSolutionComplete(ExhaustiveSearchNode<Solution_> expandingNode)
void	phaseEnded(ExhaustiveSearchPhaseScope<Solution_> phaseScope)
void	phaseStarted(ExhaustiveSearchPhaseScope<Solution_> phaseScope)
void	restoreWorkingSolution(ExhaustiveSearchStepScope<Solution_> stepScope, boolean assertWorkingSolutionScoreFromScratch, boolean assertExpectedWorkingSolutionScore)

Methods inherited from class ai.timefold.solver.core.impl.exhaustivesearch.decider.AbstractExhaustiveSearchDecider

doMove, enableAcceptUninitializedSolutions, expandNode, fillLayerList, getScoreBounder, initStartNode, setAssertExpectedUndoMoveScore, setAssertMoveScoreFromScratch, solvingEnded, solvingStarted, stepEnded, stepStarted

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface ai.timefold.solver.core.impl.solver.event.SolverLifecycleListener

solvingError

Constructor Details

ListVariableExhaustiveSearchDecider

public ListVariableExhaustiveSearchDecider(String logIndentation,
 BestSolutionRecaller<Solution_> bestSolutionRecaller,
 PhaseTermination<Solution_> termination,
 EntitySelector<Solution_> sourceEntitySelector,
 ManualEntityMimicRecorder<Solution_> manualEntityMimicRecorder,
 MoveRepository<Solution_> moveRepository,
 boolean scoreBounderEnabled,
 ScoreBounder<?> scoreBounder)

Method Details

expandNode

public void expandNode(ExhaustiveSearchStepScope<Solution_> stepScope)

The method updates the exploration of the solution's space when using a list variable. The logic will start adding all possible search nodes that belong to all available layers. For example, considering two entities [e0, e1] and the values [v0, v1], the result is: (layer 0, e0, [v0]), (layer 0, e0, [v1]), (layer 1, e1, [v0]), (layer 1, e1, [v1]). From these initial nodes, the method will be able to explore all possible solutions.

When a solver adds another value to a list, the layer remains unchanged, ensuring that the depth remains equal. This allows the solver to correctly sort the nodes using either breadth-first or depth-first approaches. For example, the following search node has three values in the list e0[v0, v1, v2], and its node tree is given as follows:

  (layer 0, e0, [v0])
              |
              |
      (layer 0, e0, [v1])
                  |
                  |
          (layer 0, e0, [v2])
  

After selecting a search node, all potential moves for that node's layer and the subsequent layers are generated. This step is essential to prevent the reevaluation of already visited solutions. It's important to note that the previous layers have already evaluated all possible permutations.

Specified by:

expandNode in class AbstractExhaustiveSearchDecider<Solution_,Score_ extends Score<Score_>>

isSolutionComplete

public boolean isSolutionComplete(ExhaustiveSearchNode<Solution_> expandingNode)

Specified by:

isSolutionComplete in class AbstractExhaustiveSearchDecider<Solution_,Score_ extends Score<Score_>>

isEntityReinitializable

public boolean isEntityReinitializable(Object entity)

Specified by:

isEntityReinitializable in class AbstractExhaustiveSearchDecider<Solution_,Score_ extends Score<Score_>>

restoreWorkingSolution

public void restoreWorkingSolution(ExhaustiveSearchStepScope<Solution_> stepScope,
 boolean assertWorkingSolutionScoreFromScratch,
 boolean assertExpectedWorkingSolutionScore)

Specified by:

restoreWorkingSolution in class AbstractExhaustiveSearchDecider<Solution_,Score_ extends Score<Score_>>

phaseStarted

public void phaseStarted(ExhaustiveSearchPhaseScope<Solution_> phaseScope)

Specified by:

phaseStarted in interface ExhaustiveSearchPhaseLifecycleListener<Solution_>

Overrides:

phaseStarted in class AbstractExhaustiveSearchDecider<Solution_,Score_ extends Score<Score_>>

phaseEnded

public void phaseEnded(ExhaustiveSearchPhaseScope<Solution_> phaseScope)

Specified by:

phaseEnded in interface ExhaustiveSearchPhaseLifecycleListener<Solution_>

Overrides:

phaseEnded in class AbstractExhaustiveSearchDecider<Solution_,Score_ extends Score<Score_>>