Package org.hortonmachine.gears.utils.optimizers.particleswarm

Class Particle

java.lang.Object

org.hortonmachine.gears.utils.optimizers.particleswarm.Particle

public class Particle
extends Object

Class representing a particle in the swarm.

Author:

Andrea Antonello (www.hydrologis.com)

Constructor Summary

Constructors

Constructor	Description
Particle(double[][] ranges)	Create a new Particle with a given number of parameters dimension.

Method Summary

Modifier and Type	Method	Description
double[]	getInitialLocations()
double	getParticleBestFunction()	Calculated local best function value for the particle.
void	setParticleBestFunction(double particleBestFunction)	Setter for the local best function value of the particle.
void	setParticleLocalBeststoCurrent()	Setter to set the current positions to be the local best positions.
double[]	update(double w, double c1, double rand1, double c2, double rand2, double[] globalBest)	Particle swarming formula to update positions.

Methods inherited from class java.lang.Object

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

Constructor Detail

Particle

public Particle(double[][] ranges)

Create a new Particle with a given number of parameters dimension.

Parameters:

ranges - the parameters spaces ranges.

Method Detail

getInitialLocations

public double[] getInitialLocations()

Returns:

the initial swarm location.

update

public double[] update(double w,
                       double c1,
                       double rand1,
                       double c2,
                       double rand2,
                       double[] globalBest)

Particle swarming formula to update positions.

Parameters:

w - inertia weight (controls the impact of the past velocity of the particle over the current one).

c1 - constant weighting the influence of local best solutions.

rand1 - random factor introduced in search process.

c2 - constant weighting the influence of global best solutions.

rand2 - random factor introduced in search process.

globalBest - leader particle (global best) in all dimensions.

Returns:

the updated locations or null, if they are outside the ranges.

getParticleBestFunction

public double getParticleBestFunction()

Calculated local best function value for the particle.

Returns:

the local best function value for the particle.

setParticleBestFunction

public void setParticleBestFunction(double particleBestFunction)

Setter for the local best function value of the particle.

Parameters:

particleBestFunction - the new local best function value to set for the particle.

setParticleLocalBeststoCurrent

public void setParticleLocalBeststoCurrent()

Setter to set the current positions to be the local best positions.