Module de.learnlib.common.util

Package de.learnlib.util

Class Experiment<A extends FiniteRepresentation,I,D>

java.lang.Object

de.learnlib.util.Experiment<A,I,D>

Type Parameters:

A - the automaton type

I - input symbol type

D - output domain type

Direct Known Subclasses:

DFAExperiment, MealyExperiment, MooreExperiment

public class Experiment<A extends FiniteRepresentation,I,D>
extends Object

Runs a learning experiment.

Field Summary

Fields

Modifier and Type	Field	Description
protected final EquivalenceOracle<? super A,I,D>	equivalenceAlgorithm
protected final Alphabet<I>	inputs
static final StatisticsKey	KEY_DUR_CEX	The StatisticsKey this class uses for clocking the duration of the counterexample search of the equivalence oracle.
static final StatisticsKey	KEY_DUR_LEARN	The StatisticsKey this class uses for clocking the duration of the exploration phase of the learning algorithm.
static final StatisticsKey	KEY_FINAL_SIZE	The StatisticsKey this class uses for counting the size of the final hypothesis.
static final StatisticsKey	KEY_ROUNDS	The StatisticsKey this class uses for counting the number of learning rounds of this experiment.
protected final LearningAlgorithm<? extends A,I,D>	learningAlgorithm
protected static final org.slf4j.Logger	LOGGER
protected final @Nullable InputModelSerializer<I,? super A>	serializer
protected final StatisticsService	statistics

Constructor Summary

Constructors

Constructor	Description
Experiment(LearningAlgorithm<? extends A,I,D> learningAlgorithm, EquivalenceOracle<? super A,I,D> equivalenceAlgorithm, Alphabet<I> inputs)	Constructor.
Experiment(LearningAlgorithm<? extends A,I,D> learningAlgorithm, EquivalenceOracle<? super A,I,D> equivalenceAlgorithm, Alphabet<I> inputs, @Nullable InputModelSerializer<I,? super A> serializer)	Constructor.

Method Summary

Modifier and Type	Method	Description
final A	getFinalHypothesis()	Returns the final hypothesis model.
protected final int	getRound()	Utility method to access to current learning round.
protected void	initializeLearning()	Initializes the learning process.
protected void	postRefinementHook()	Called upon calling LearningAlgorithm.refineHypothesis(DefaultQuery).
final A	run()	Run the experiment, once.

Methods inherited from class java.lang.Object

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

Field Details

KEY_DUR_LEARN

public static final StatisticsKey KEY_DUR_LEARN

The StatisticsKey this class uses for clocking the duration of the exploration phase of the learning algorithm.

KEY_DUR_CEX

public static final StatisticsKey KEY_DUR_CEX

The StatisticsKey this class uses for clocking the duration of the counterexample search of the equivalence oracle.

KEY_ROUNDS

public static final StatisticsKey KEY_ROUNDS

The StatisticsKey this class uses for counting the number of learning rounds of this experiment.

KEY_FINAL_SIZE

public static final StatisticsKey KEY_FINAL_SIZE

The StatisticsKey this class uses for counting the size of the final hypothesis.

LOGGER

protected static final org.slf4j.Logger LOGGER

learningAlgorithm

protected final LearningAlgorithm<? extends A extends FiniteRepresentation,I,D> learningAlgorithm

equivalenceAlgorithm

protected final EquivalenceOracle<? super A extends FiniteRepresentation,I,D> equivalenceAlgorithm

inputs

protected final Alphabet<I> inputs

serializer

protected final @Nullable InputModelSerializer<I,? super A extends FiniteRepresentation> serializer

statistics

protected final StatisticsService statistics

Constructor Details

Experiment

public Experiment(LearningAlgorithm<? extends A,I,D> learningAlgorithm,
 EquivalenceOracle<? super A,I,D> equivalenceAlgorithm,
 Alphabet<I> inputs)

Constructor. Delegates to Experiment(LearningAlgorithm, EquivalenceOracle, Alphabet, InputModelSerializer) using null for serializer.

Parameters:

learningAlgorithm - the learning algorithm to use in this experiment

equivalenceAlgorithm - the strategy for finding counterexamples

inputs - the inputs to consider for exploration

See Also:

• Experiment(LearningAlgorithm, EquivalenceOracle, Alphabet, InputModelSerializer)

Experiment

public Experiment(LearningAlgorithm<? extends A,I,D> learningAlgorithm,
 EquivalenceOracle<? super A,I,D> equivalenceAlgorithm,
 Alphabet<I> inputs,
 @Nullable InputModelSerializer<I,? super A> serializer)

Constructor. Creates a new experiment to run.

Parameters:

learningAlgorithm - the learning algorithm to use in this experiment

equivalenceAlgorithm - the strategy for finding counterexamples

inputs - the inputs to consider for exploration

serializer - the serializer for logging intermediate hypotheses (may be null in case no such logging is wanted)

Method Details

getFinalHypothesis

public final A getFinalHypothesis()

Returns the final hypothesis model.

Returns:

the final hypothesis model

Throws:

IllegalStateException - if the experiment has not been run yet

run

public final A run()

Run the experiment, once.

Returns:

the final hypothesis

Throws:

IllegalStateException - if invoked more than once

getRound

protected final int getRound()

Utility method to access to current learning round.

Returns:

the current learning round

initializeLearning

protected void initializeLearning()

Initializes the learning process. By default, this method calls LearningAlgorithm.startLearning().

postRefinementHook

protected void postRefinementHook()

Called upon calling LearningAlgorithm.refineHypothesis(DefaultQuery).