com.numericalmethod.suanshu.optimization.geneticalgorithm.minimizer.simplegrid

Class SimpleGridMinimizer.Solution

java.lang.Object

com.numericalmethod.suanshu.optimization.geneticalgorithm.GeneticAlgorithm

com.numericalmethod.suanshu.optimization.geneticalgorithm.minimizer.simplegrid.SimpleGridMinimizer.Solution

All Implemented Interfaces:

IterativeMethod<Vector>, IterativeMinimizer<Vector>, MinimizationSolution<Vector>, java.io.Serializable

Direct Known Subclasses:

DEOptim.Solution

Enclosing class:

SimpleGridMinimizer

protected class SimpleGridMinimizer.Solution
extends GeneticAlgorithm
implements IterativeMinimizer<Vector>

This is the solution to a minimization problem using SimpleGridMinimizer.

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected RealScalarFunction	f
protected SimpleCellFactory	factory
protected double	fmin
protected double	fminLast
protected Vector[]	initials
protected int	iteration
protected int	nNoChanges
protected Vector	xmin

Fields inherited from class com.numericalmethod.suanshu.optimization.geneticalgorithm.GeneticAlgorithm

parallel, population, uniform

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	Solution(RealScalarFunction f)

Method Summary

Modifier and Type	Method and Description
protected java.util.ArrayList<Chromosome>	initialization() The initial population is generated by putting a uniform mesh/grid/net over the entire region.
protected boolean	isConverged() This genetic algorithm terminates if the minimum does not improve for a fixed number of iterations, or the maximum number of iterations is exceeded.
ImmutableVector	minimizer() Get the minimizer (solution) to the minimization problem.
double	minimum() Get the (approximate) minimum found.
Vector	search(Vector... initials) Search for a solution that optimizes the objective function from the given starting points.
void	setInitials(Vector... initials) Supply the starting points for the search.
java.lang.Object	step() Run a step in genetic algorithm: produce the next generation of chromosome pool.

Methods inherited from class com.numericalmethod.suanshu.optimization.geneticalgorithm.GeneticAlgorithm

getBest, getChild, getNewPool, getNextGeneration, getOne, nChildren, nPopulation, run

Methods inherited from class java.lang.Object

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

Field Detail

initials

protected Vector[] initials

iteration

protected int iteration

nNoChanges

protected int nNoChanges

fminLast

protected double fminLast

fmin

protected double fmin

xmin

protected Vector xmin

f

protected final RealScalarFunction f

factory

protected final SimpleCellFactory factory

Constructor Detail

Solution

protected Solution(RealScalarFunction f)

Method Detail

initialization

protected java.util.ArrayList<Chromosome> initialization()

The initial population is generated by putting a uniform mesh/grid/net over the entire region. The grid points are the chromosomes in the first population. The population size is proportional to the number of available cores. The region bounds are determined from the initial guesses.

Specified by:

initialization in class GeneticAlgorithm

Returns:

the first population

isConverged

protected boolean isConverged()

This genetic algorithm terminates if

• the minimum does not improve for a fixed number of iterations, or
• the maximum number of iterations is exceeded.

Specified by:

isConverged in class GeneticAlgorithm

Returns:

true if a convergence is found

setInitials

public void setInitials(Vector... initials)

Description copied from interface: IterativeMethod

Supply the starting points for the search. This can also initialize the state of the algorithm for a new search.

Specified by:

setInitials in interface IterativeMethod<Vector>

Parameters:

initials - the initial guesses

step

public java.lang.Object step()

Description copied from class: GeneticAlgorithm

Run a step in genetic algorithm: produce the next generation of chromosome pool.

Specified by:

step in interface IterativeMethod<Vector>

Overrides:

step in class GeneticAlgorithm

Returns:

true

search

public Vector search(Vector... initials)

Description copied from interface: IterativeMethod

Search for a solution that optimizes the objective function from the given starting points. This method typically calls first #setInitials(S[]) and then iteratively IterativeMethod.step(). It implements a default convergence criterion.

Specified by:

search in interface IterativeMethod<Vector>

Parameters:

initials - the initial guesses

Returns:

an (approximate) optimizer

minimum

public double minimum()

Description copied from interface: MinimizationSolution

Get the (approximate) minimum found.

Specified by:

minimum in interface MinimizationSolution<Vector>

Returns:

the (approximate) minimum found

minimizer

public ImmutableVector minimizer()

Description copied from interface: MinimizationSolution

Get the minimizer (solution) to the minimization problem.

Specified by:

minimizer in interface MinimizationSolution<Vector>

Returns:

the minimizer