opt4j

Type Members

trait CoolingSchedule extends AnyRef
case class DifferentialEvolution(generations: Int = 1000, alpha: Int = 100, scalingFactor: Double = 0.5) extends Opt4JAlgorithm with _EvolutionaryAlgorithm with Product with Serializable

Represents algorithm for differential evolution.
Represents algorithm for differential evolution. Caution: seems to only support real-valued parameters so far.
generations
the number of generations
alpha
number of individuals per generation
scalingFactor
the scaling factor

See also
DifferentialEvolutionModule
class DoubleGenotypeCreator extends Creator[DoubleGenotype]
class DoubleGenotypeNeighbor extends Neighbor[Genotype]
class DoubleGenotypeToSimpleParamsDecoder extends Decoder[DoubleGenotype, SimpleParameters]
case class EvolutionaryAlgorithm(generations: Int = 1000, alpha: Int = 100, mu: Int = 25, lambda: Int = 25, rate: Double = 0.95) extends Opt4JAlgorithm with _EvolutionaryAlgorithm with Product with Serializable

Represents an evolutionary algorithm.
Represents an evolutionary algorithm.
generations
number of generations
alpha
number of individuals per generation
mu
number of parents per generation
lambda
number of offspring per generation
rate
rate of the crossover operation

See also
EvolutionaryAlgorithmModule
case class ExponentialCoolingSchedule(initialTemperature: Double = 10, finalTemperature: Double = 0, alpha: Double = 0.995) extends CoolingSchedule with Product with Serializable
case class HyperbolicCoolingSchedule(initialTemperature: Double = 10, finalTemperature: Double = 0) extends CoolingSchedule with Product with Serializable
class IterationListener extends OptimizerIterationListener

Listens to the end of an iteration within the optimizer and calls the corresponding event handling methods from Opt4JSetup.
Listens to the end of an iteration within the optimizer and calls the corresponding event handling methods from Opt4JSetup.

See also
Opt4JSetup
case class LinearCoolingSchedule(initialTemperature: Double = 10, finalTemperature: Double = 0) extends CoolingSchedule with Product with Serializable
trait Opt4JAlgorithm extends AnyRef

Common interface of all optimization algorithms in Opt4J.
Common interface of all optimization algorithms in Opt4J. Their parameters and default values as taken from the original source code.
class Opt4JSetup extends AbstractOptimizerSetup with Logging

Support for Opt4J.
case class ParticleSwarmOptimization(iterations: Int = 100, archiveSize: Int = 10, particles: Int = 10, perturbation: Double = 0.5) extends Opt4JAlgorithm with Product with Serializable
case class RandomSearch(iterations: Int = 1000, batchsize: Int = 25) extends Opt4JAlgorithm with _RandomSearch with Product with Serializable

Represents an optimizer using random search.
Represents an optimizer using random search.
iterations
the number of iterations
batchsize
the batch size (number of replications for the randomly sampled individual of the given iteration)

See also
RandomSearchModule
case class SearchSpace(searchSpace: List[BoundedSearchSpaceDimension[Double]]) extends Product with Serializable
class SearchSpaceModule extends AbstractModule
class SimpleParameterCreator extends Creator[CompositeGenotype[String, Genotype]]

Represents genotype by simply creating random parameter combinations.
Represents genotype by simply creating random parameter combinations. It is important that Opt4J's built-in genotypes are used, since this allows to re-use the built-in operators for mutation and recombination.
See http://opt4j.sourceforge.net/documentation/3.0/tutorial.xhtml for details.

See also
org.opt4j.core.problem.Genotype
class SimpleParameterDecoder extends Decoder[CompositeGenotype[String, Genotype], SimpleParameters]

Decodes genotype into phenotype.
Decodes genotype into phenotype.
Here, this is simply a matter of reading out the parameter values and putting them into a map representing a parameter assignment.

See also
org.opt4j.core.problem.Decoder
class SimpleParameterEvaluator extends Evaluator[SimpleParameters] with Logging

Evaluates phenotypes.
Evaluates phenotypes.

See also
org.opt4j.core.problem.Evaluator
case class SimulatedAnnealing(iterations: Int = 100000, schedule: CoolingSchedule = LinearCoolingSchedule()) extends Opt4JAlgorithm with _SimulatedAnnealing with Product with Serializable

Represents an optimizer based on simulated annealing.
Represents an optimizer based on simulated annealing.
iterations
the number of iterations
schedule
The type of the Cooling Schedule (Linear, Hyperbolic, Exponential)

See also
SimulatedAnnealingModule

Value Members

object DoubleNeighborModule extends AbstractModule
object Opt4JSetup

Companion object for the setup.
Companion object for the setup. Since Opt4J uses the GUICE framework for dependency injection, which does not seen to support the binding of anonymous inner classes, this object contains all problem-specific information that is required by the creator/decoder/evaluator classes.
To avoid bugs (race conditions etc.), the setup has a dedicated owner, i.e. in the current implementation Opt4J cannot be used to optimize multiple problems in parallel, only sequentially (this should not be too big a deal, as each objective function evaluation may be parallelized, and also the evaluation of different individuals).

See also
AbstractOptimizerSetup
SimpleParameterEvaluator
SimpleParameterDecoder
SimpleParameterCreator

package opt4j

Type Members

trait CoolingSchedule extends AnyRef

case class DifferentialEvolution(generations: Int = 1000, alpha: Int = 100, scalingFactor: Double = 0.5) extends Opt4JAlgorithm with _EvolutionaryAlgorithm with Product with Serializable

class DoubleGenotypeCreator extends Creator[DoubleGenotype]

class DoubleGenotypeNeighbor extends Neighbor[Genotype]

class DoubleGenotypeToSimpleParamsDecoder extends Decoder[DoubleGenotype, SimpleParameters]

case class EvolutionaryAlgorithm(generations: Int = 1000, alpha: Int = 100, mu: Int = 25, lambda: Int = 25, rate: Double = 0.95) extends Opt4JAlgorithm with _EvolutionaryAlgorithm with Product with Serializable

case class ExponentialCoolingSchedule(initialTemperature: Double = 10, finalTemperature: Double = 0, alpha: Double = 0.995) extends CoolingSchedule with Product with Serializable

case class HyperbolicCoolingSchedule(initialTemperature: Double = 10, finalTemperature: Double = 0) extends CoolingSchedule with Product with Serializable

class IterationListener extends OptimizerIterationListener

case class LinearCoolingSchedule(initialTemperature: Double = 10, finalTemperature: Double = 0) extends CoolingSchedule with Product with Serializable

trait Opt4JAlgorithm extends AnyRef

class Opt4JSetup extends AbstractOptimizerSetup with Logging

case class ParticleSwarmOptimization(iterations: Int = 100, archiveSize: Int = 10, particles: Int = 10, perturbation: Double = 0.5) extends Opt4JAlgorithm with Product with Serializable

case class RandomSearch(iterations: Int = 1000, batchsize: Int = 25) extends Opt4JAlgorithm with _RandomSearch with Product with Serializable

case class SearchSpace(searchSpace: List[BoundedSearchSpaceDimension[Double]]) extends Product with Serializable

class SearchSpaceModule extends AbstractModule

class SimpleParameterCreator extends Creator[CompositeGenotype[String, Genotype]]

class SimpleParameterDecoder extends Decoder[CompositeGenotype[String, Genotype], SimpleParameters]

class SimpleParameterEvaluator extends Evaluator[SimpleParameters] with Logging

case class SimulatedAnnealing(iterations: Int = 100000, schedule: CoolingSchedule = LinearCoolingSchedule()) extends Opt4JAlgorithm with _SimulatedAnnealing with Product with Serializable

Value Members

object DoubleNeighborModule extends AbstractModule

object Opt4JSetup

Ungrouped