Package org.apache.commons.math4.legacy.optim.nonlinear.scalar

Class SimulatedAnnealing

java.lang.Object

org.apache.commons.math4.legacy.optim.nonlinear.scalar.SimulatedAnnealing

All Implemented Interfaces:

OptimizationData

public class SimulatedAnnealing
extends Object
implements OptimizationData

Simulated annealing setup.

Since:

4.0

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static interface	SimulatedAnnealing.CoolingSchedule	Specifies the cooling schedule.

Constructor Summary

Constructors

Constructor	Description
SimulatedAnnealing(int epoch, double startProb, double endProb, SimulatedAnnealing.CoolingSchedule cooling, org.apache.commons.rng.UniformRandomProvider random)

Method Summary

Modifier and Type	Method	Description
SimulatedAnnealing.CoolingSchedule	getCoolingSchedule()
double	getEndProbability()
int	getEpochDuration()
double	getStartProbability()
DoublePredicate	metropolis(double temperature)	Factory for the Metropolis check for accepting a worse state.

Methods inherited from class java.lang.Object

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

Constructor Detail

SimulatedAnnealing

public SimulatedAnnealing(int epoch,
                          double startProb,
                          double endProb,
                          SimulatedAnnealing.CoolingSchedule cooling,
                          org.apache.commons.rng.UniformRandomProvider random)

Parameters:

epoch - Number of iterations performed at fixed temperature.

startProb - Initial acceptance probablility.

endProb - Final acceptance probablility.

cooling - Computes the temperature as a function of the initial temperature and the epoch. It is called for computing a new temperature after each cycle of epoch iterations. Simulated annealing assumes that the function decreases monotically wrt the epoch (cf. provided implementation).

random - Random number generator.

Throws:

IllegalArgumentException - if epoch < 1 or startProb or endProb is outside the [0, 1] interval.

Method Detail

getEpochDuration

public int getEpochDuration()

Returns:

the epoch duration.

getStartProbability

public double getStartProbability()

Returns:

the acceptance probability at the beginning of the SA process.

getEndProbability

public double getEndProbability()

Returns:

the acceptance probability at the end of the SA process.

getCoolingSchedule

public SimulatedAnnealing.CoolingSchedule getCoolingSchedule()

Returns:

the cooling schedule.

metropolis

public DoublePredicate metropolis(double temperature)

Factory for the Metropolis check for accepting a worse state. \( e^{-|\Delta E| / T} \geq \mathrm{rand}(0, 1) \).

It is assumed that this check is performed after ensuring that the alternate state is worse than the current best state.

Parameters:

temperature - Current temperature.

Returns:

the acceptance test.