org.jenetics

Class LineCrossover<G extends NumericGene<?,G>,C extends Comparable<? super C>>

java.lang.Object

org.jenetics.AbstractAlterer<G,C>

org.jenetics.Recombinator<G,C>

org.jenetics.Crossover<G,C>

org.jenetics.LineCrossover<G,C>

All Implemented Interfaces:

Alterer<G,C>

public class LineCrossover<G extends NumericGene<?,G>,C extends Comparable<? super C>>
extends Crossover<G,C>

This alterer takes two chromosome (treating it as vectors) and creates a linear combination of this vectors as result. The line-recombination depends on a variable p which determines how far out along the line (defined by the two multidimensional points/vectors) the children are allowed to be. If p = 0 then the children will be located along the line within the hypercube between the two points. If p > 0 then the children may be located anywhere on the line, even somewhat outside of the hypercube.

Points outside of the allowed numeric range are rejected and the original value are used instead. The strategy on how out-of-range points are handled, is the difference to the very similar IntermediateCrossover.

Since:

3.8

Version:

3.8

Author:

Franz Wilhelmstötter

See Also:

Essentials of Metaheuristic, page 42, IntermediateCrossover

Field Summary

Fields inherited from class org.jenetics.AbstractAlterer

_probability

Fields inherited from interface org.jenetics.Alterer

DEFAULT_ALTER_PROBABILITY

Constructor Summary

Constructors

Constructor and Description
LineCrossover() Creates a new linear-crossover with default recombination probability (Alterer.DEFAULT_ALTER_PROBABILITY) and a p value of zero, which restricts the recombined chromosomes within the hypercube of the selected chromosomes (vectors).
LineCrossover(double probability) Creates a new linear-crossover with the given recombination probability.
LineCrossover(double probability, double p) Creates a new linear-crossover with the given recombination probability and the line-scaling factor p.

Method Summary

Modifier and Type	Method and Description
protected int	crossover(MSeq<G> v, MSeq<G> w) Template method which performs the crossover.
boolean	equals(Object obj)
int	hashCode()
String	toString()

Methods inherited from class org.jenetics.Crossover

recombine

Methods inherited from class org.jenetics.Recombinator

alter, getOrder

Methods inherited from class org.jenetics.AbstractAlterer

getProbability

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface org.jenetics.Alterer

andThen, compose, of

Constructor Detail

LineCrossover

public LineCrossover(double probability,
                     double p)

Creates a new linear-crossover with the given recombination probability and the line-scaling factor p.

Parameters:

probability - the recombination probability.

p - defines the possible location of the recombined chromosomes. If p = 0 then the children will be located along the line within the hypercube between the two points. If p > 0 then the children may be located anywhere on the line, even somewhat outside of the hypercube.

Throws:

IllegalArgumentException - if the probability is not in the valid range of [0, 1] or if p is smaller then zero

LineCrossover

public LineCrossover(double probability)

Creates a new linear-crossover with the given recombination probability. The parameter p is set to zero, which restricts the recombined chromosomes within the hypercube of the selected chromosomes (vectors).

Parameters:

probability - the recombination probability.

Throws:

IllegalArgumentException - if the probability is not in the valid range of [0, 1]

LineCrossover

public LineCrossover()

Creates a new linear-crossover with default recombination probability (Alterer.DEFAULT_ALTER_PROBABILITY) and a p value of zero, which restricts the recombined chromosomes within the hypercube of the selected chromosomes (vectors).

Method Detail

crossover

protected int crossover(MSeq<G> v,
                        MSeq<G> w)

Description copied from class: Crossover

Template method which performs the crossover. The arguments given are mutable non null arrays of the same length.

Specified by:

crossover in class Crossover<G extends NumericGene<?,G>,C extends Comparable<? super C>>

Parameters:

v - the genes of the first chromosome

w - the genes of the other chromosome

Returns:

the number of altered genes

hashCode

public int hashCode()

Overrides:

hashCode in class AbstractAlterer<G extends NumericGene<?,G>,C extends Comparable<? super C>>

equals

public boolean equals(Object obj)

Overrides:

equals in class AbstractAlterer<G extends NumericGene<?,G>,C extends Comparable<? super C>>

toString

public String toString()

Overrides:

toString in class Object