Package org.moeaframework.algorithm.pso

Class OMOPSO

java.lang.Object

org.moeaframework.algorithm.AbstractAlgorithm

org.moeaframework.algorithm.pso.AbstractPSOAlgorithm

org.moeaframework.algorithm.pso.OMOPSO

All Implemented Interfaces:

Algorithm, Configurable, Stateful

public class OMOPSO
extends AbstractPSOAlgorithm

Implementation of OMOPSO, a multi-objective particle swarm optimizer (MOPSO). According to [2], OMOPSO is one of the top-performing PSO algorithms.

References:

1. Sierra, M. R. and C. A. Coello Coello (2005). Improving PSO-based Multi-Objective Optimization using Crowding, Mutation and ε-Dominance. Evolutionary Multi-Criterion Optimization, pp. 505-519.
2. Durillo, J. J., J. Garc�a-Nieto, A. J. Nebro, C. A. Coello Coello, F. Luna, and E. Alba (2009). Multi-Objective Particle Swarm Optimizers: An Experimental Comparison. Evolutionary Multi-Criterion Optimization, pp. 495-509.

Field Summary

Fields inherited from class org.moeaframework.algorithm.pso.AbstractPSOAlgorithm

archive, dominanceComparator, leaderComparator, leaders, localBestParticles, mutation, particles, velocities

Fields inherited from class org.moeaframework.algorithm.AbstractAlgorithm

initialized, numberOfEvaluations, problem, terminated

Constructor Summary

Constructors

Constructor	Description
OMOPSO(Problem problem, int maxIterations)	Constructs a new OMOPSO instance with default settings.
OMOPSO(Problem problem, int swarmSize, int leaderSize, Epsilons epsilons, double mutationProbability, double mutationPerturbation, int maxIterations)	Constructs a new OMOPSO instance.

Method Summary

Modifier and Type	Method	Description
void	applyConfiguration(TypedProperties properties)	Applies the properties to this instance.
protected EpsilonBoxDominanceArchive	getArchive()	Returns the archive of non-dominated solutions; or null of no external archive is used.
TypedProperties	getConfiguration()	Gets the current configuration of this instance.
double	getMutationProbability()	Returns the mutation probability used by the uniform and non-uniform mutation operators.
double	getPerturbationIndex()	Returns the perturbation index used by uniform and non-uniform mutation.
protected void	mutate(int i)	Applies the mutation operator to an individual particle.
protected void	setArchive(EpsilonBoxDominanceArchive archive)
protected void	setMaxIterations(int maxIterations)	Sets the maximum number of iterations for scaling the non-uniform mutation.
void	setMutationProbability(double mutationProbability)	Sets the mutation probability used by the uniform and non-uniform mutation operators.
void	setPerturbationIndex(double perturbationIndex)	Sets the perturbation index used by uniform and non-uniform mutation.

Methods inherited from class org.moeaframework.algorithm.pso.AbstractPSOAlgorithm

copyToList, getLeaders, getLeaderSize, getLocalBestParticles, getMutation, getParticles, getResult, getSwarmSize, initialize, iterate, loadState, mutate, saveState, selectLeader, setArchive, setLeaderSize, setSwarmSize, updateLocalBest, updatePosition, updatePositions, updateVelocities, updateVelocity

Methods inherited from class org.moeaframework.algorithm.AbstractAlgorithm

assertNotInitialized, evaluate, evaluateAll, evaluateAll, getNumberOfEvaluations, getProblem, isInitialized, isTerminated, step, terminate

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.moeaframework.core.Algorithm

run, run

Constructor Details

OMOPSO

public OMOPSO(Problem problem,
 int maxIterations)

Constructs a new OMOPSO instance with default settings.

Parameters:

problem - the problem

maxIterations - the maximum number of iterations for scaling non-uniform mutation; typically this should be maxEvaluations / swarmSize

OMOPSO

public OMOPSO(Problem problem,
 int swarmSize,
 int leaderSize,
 Epsilons epsilons,
 double mutationProbability,
 double mutationPerturbation,
 int maxIterations)

Constructs a new OMOPSO instance.

Parameters:

problem - the problem

swarmSize - the number of particles

leaderSize - the number of leaders

epsilons - the ε-values used in the external archive

mutationProbability - the mutation probability for uniform and non-uniform mutation

mutationPerturbation - the perturbation index for uniform and non-uniform mutation

maxIterations - the maximum iterations for scaling the non-uniform mutation

Method Details

getMutationProbability

public double getMutationProbability()

Returns the mutation probability used by the uniform and non-uniform mutation operators.

Returns:

the mutation probability

setMutationProbability

public void setMutationProbability(double mutationProbability)

Sets the mutation probability used by the uniform and non-uniform mutation operators. The default value is 1 / N, where N is the number of decision variables.

Parameters:

mutationProbability - the mutation probability

getPerturbationIndex

public double getPerturbationIndex()

Returns the perturbation index used by uniform and non-uniform mutation.

Returns:

the perturbation index

setPerturbationIndex

public void setPerturbationIndex(double perturbationIndex)

Sets the perturbation index used by uniform and non-uniform mutation. The default value is 0.5.

Parameters:

perturbationIndex - the perturbation index

getArchive

protected EpsilonBoxDominanceArchive getArchive()

Description copied from class: AbstractPSOAlgorithm

Returns the archive of non-dominated solutions; or null of no external archive is used.

Overrides:

getArchive in class AbstractPSOAlgorithm

Returns:

the archive or null

setArchive

protected void setArchive(EpsilonBoxDominanceArchive archive)

setMaxIterations

protected void setMaxIterations(int maxIterations)

Sets the maximum number of iterations for scaling the non-uniform mutation. Typically this should be set to maxEvaluations / swarmSize.

Parameters:

maxIterations - the maximum number of iterations

mutate

protected void mutate(int i)

Description copied from class: AbstractPSOAlgorithm

Applies the mutation operator to an individual particle.

Overrides:

mutate in class AbstractPSOAlgorithm

Parameters:

i - the index of the particle

applyConfiguration

public void applyConfiguration(TypedProperties properties)

Description copied from interface: Configurable

Applies the properties to this instance. It is strongly recommended to apply a configuration immediately after creating the instance, as some properties can not be changed after the class is used. Exceptions may be thrown if attempting to set such properties. After calling this method, we encourage users to call TypedProperties.warnIfUnaccessedProperties() to verify all properties were processed. This can identify simple mistakes like typos. If overriding this method, properties should only be updated if a new value is provided. Additionally, if updating any Configurable objects inside this object, they should be updated before calling super.applyConfiguration(properties).

Parameters:

properties - the user-defined properties

getConfiguration

public TypedProperties getConfiguration()

Description copied from interface: Configurable

Gets the current configuration of this instance. In theory, these properties should be able to create a duplicate instance. Note however, they are unlikely to behave identically due to random numbers and other transient fields.

Returns:

the properties defining this instance