org.encog.ml.anneal

Class SimulatedAnnealing<UNIT_TYPE>

java.lang.Object

org.encog.ml.anneal.SimulatedAnnealing<UNIT_TYPE>

Type Parameters:

UNIT_TYPE - What type of data makes up the solution.

Direct Known Subclasses:

NeuralSimulatedAnnealingHelper

public abstract class SimulatedAnnealing<UNIT_TYPE>
extends java.lang.Object

Simulated annealing is a common training method. This class implements a simulated annealing algorithm that can be used both for neural networks, as well as more general cases. This class is abstract, so a more specialized simulated annealing subclass will need to be created for each intended use. This book demonstrates how to use the simulated annealing algorithm to train feedforward neural networks, as well as find a solution to the traveling salesman problem. The name and inspiration come from annealing in metallurgy, a technique involving heating and controlled cooling of a material to increase the size of its crystals and reduce their defects. The heat causes the atoms to become unstuck from their initial positions (a local minimum of the internal energy) and wander randomly through states of higher energy; the slow cooling gives them more chances of finding configurations with lower internal energy than the initial one.

Constructor Summary

Constructors

Constructor and Description
SimulatedAnnealing()

Method Summary

Modifier and Type	Method and Description
abstract double	calculateScore() Subclasses should provide a method that evaluates the score for the current solution.
abstract UNIT_TYPE[]	getArray() Subclasses must provide access to an array that makes up the solution.
abstract UNIT_TYPE[]	getArrayCopy() Get a copy of the array.
int	getCycles()
double	getScore()
double	getStartTemperature()
double	getStopTemperature()
double	getTemperature()
boolean	isShouldMinimize()
void	iteration() Called to perform one cycle of the annealing process.
abstract void	putArray(UNIT_TYPE[] array) Store the array.
abstract void	randomize() Randomize the weight matrix.
void	setCycles(int theCycles)
void	setScore(double theScore) Set the score.
void	setShouldMinimize(boolean theShouldMinimize) Should the score be minimized.
void	setStartTemperature(double theStartTemperature)
void	setStopTemperature(double theStopTemperature)
void	setTemperature(double theTemperature)

Methods inherited from class java.lang.Object

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

Constructor Detail

SimulatedAnnealing

public SimulatedAnnealing()

Method Detail

calculateScore

public abstract double calculateScore()

Subclasses should provide a method that evaluates the score for the current solution. Those solutions with a lower score are better.

Returns:

Return the score.

getArray

public abstract UNIT_TYPE[] getArray()

Subclasses must provide access to an array that makes up the solution.

Returns:

An array that makes up the solution.

getArrayCopy

public abstract UNIT_TYPE[] getArrayCopy()

Get a copy of the array.

Returns:

A copy of the array.

getCycles

public int getCycles()

Returns:

the cycles

getScore

public double getScore()

Returns:

the globalScore

getStartTemperature

public double getStartTemperature()

Returns:

the startTemperature

getStopTemperature

public double getStopTemperature()

Returns:

the stopTemperature

getTemperature

public double getTemperature()

Returns:

the temperature

isShouldMinimize

public boolean isShouldMinimize()

Returns:

True if the score should be minimized.

iteration

public void iteration()

Called to perform one cycle of the annealing process.

putArray

public abstract void putArray(UNIT_TYPE[] array)

Store the array.

Parameters:

array - The array to be stored.

randomize

public abstract void randomize()

Randomize the weight matrix.

setCycles

public void setCycles(int theCycles)

Parameters:

theCycles - the cycles to set

setScore

public void setScore(double theScore)

Set the score.

Parameters:

theScore - The score to set.

setShouldMinimize

public void setShouldMinimize(boolean theShouldMinimize)

Should the score be minimized.

Parameters:

theShouldMinimize - True if the score should be minimized.

setStartTemperature

public void setStartTemperature(double theStartTemperature)

Parameters:

theStartTemperature - the startTemperature to set

setStopTemperature

public void setStopTemperature(double theStopTemperature)

Parameters:

theStopTemperature - the stopTemperature to set

setTemperature

public void setTemperature(double theTemperature)

Parameters:

theTemperature - the temperature to set