DifferentialRecombination (Morphokinetics)

java.lang.Object
- geneticAlgorithm.recombination.DifferentialRecombination

All Implemented Interfaces:: IRecombination

public class DifferentialRecombination
extends java.lang.Object
implements IRecombination

Field Summary

Fields
Modifier and Type	Field and Description
`private RichMatrix`	`B` B defines the coordinate system.
`private RichMatrix`	`C` Covariance matrix C.
`private double`	`c1`
`private double`	`cc`
`private double`	`chiN` Expectation of \|\|N(0,I)\|\| == norm(randn(N,1)).
`private double`	`cmu`
`private int`	`counteval` Counter for the number of evaluations.
`private double`	`cSigma` C_sigma.
`private RichArray`	`D` D contains the standard deviations.
`private double`	`dampSigma` Damping for sigma.
`private int`	`dimensions` Number of objective variables/problem dimension.
`private double`	`eigeneval` Track update of B and D.
`private int`	`errorsNumber`
`private RichMatrix`	`invsqrtC` C^-1/2
`private int`	`mu` Number of parents for recombination in CMA-ES.
`private double`	`muEffective` Variance effective selection mass.
`private RichArray`	`pc` P_c of sigma.
`private RichArray`	`pSigma` P of sigma.
`private double`	`sigma` Step size in CMA-ES.
`private RichArray`	`weights` Array for weighted recombination in CMA-ES.
`private RichArray`	`xMean` Mean of objective variables.

Constructor Summary

Constructors
Constructor and Description

DifferentialRecombination(int populationSize, int dimensions)

Constructors
Constructor and Description
`DifferentialRecombination(int populationSize, int dimensions)`

Method Summary

All Methods Instance Methods Concrete Methods
Modifier and Type	Method and Description
`private void`	`createWeightsArray()`
`void`	`initialise(Population population)` Initialises the first population and assigns values to sigma and xMean.
`boolean`	`isDtooLarge()` Condition exceeds 1e14
`Population`	`recombinate(Population population, IndividualGroup[] groups)` Creates a new offspring population.

Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

- Field Detail
  - cc
```
private final double cc
```
  - cSigma
```
private final double cSigma
```
    C_sigma.
  - c1
```
private final double c1
```
  - cmu
```
private final double cmu
```
  - dampSigma
```
private final double dampSigma
```
    Damping for sigma.
  - pc
```
private RichArray pc
```
    P_c of sigma. Evolution paths for C and sigma
  - pSigma
```
private RichArray pSigma
```
    P of sigma. Used for step size control.
  - sigma
```
private double sigma
```
    Step size in CMA-ES. It stores the standard deviation of all objective variables.
  - B
```
private RichMatrix B
```
    B defines the coordinate system. B is an orthogonal matrix. Columns of B are eigenvectors of C with unit length and correspond to the diagonal elements of D.
  - C
```
private RichMatrix C
```
    Covariance matrix C. Covariance matrix at generation g.
  - D
```
private RichArray D
```
    D contains the standard deviations. Diagonal D defines the scaling. A diagonal matrix (thus, represented with an array). The diagonal elements of D are square roots of eigenvalues of C and correspond to the respective columns of B.
  - eigeneval
```
private double eigeneval
```
    Track update of B and D.
  - invsqrtC
```
private RichMatrix invsqrtC
```
    C^-1/2
  - chiN
```
private final double chiN
```
    Expectation of ||N(0,I)|| == norm(randn(N,1)).
  - dimensions
```
private final int dimensions
```
    Number of objective variables/problem dimension.
  - xMean
```
private RichArray xMean
```
    Mean of objective variables. Recombination, new mean value in CMA-ES.
  - mu
```
private final int mu
```
    Number of parents for recombination in CMA-ES.
  - muEffective
```
private final double muEffective
```
    Variance effective selection mass. It holds 1 <= muEffective <= mu
  - weights
```
private RichArray weights
```
    Array for weighted recombination in CMA-ES. Positive weight coefficients for recombination.
  - counteval
```
private int counteval
```
    Counter for the number of evaluations. A good optimisation method will minimise this number.
  - errorsNumber
```
private final int errorsNumber
```
- Constructor Detail
  - DifferentialRecombination
```
public DifferentialRecombination(int populationSize,
                                 int dimensions)
```
    Parameters:
    
    populationSize - Number of individuals of the population
    
    dimensions - Number of different genes
- Method Detail
  - createWeightsArray
```
private void createWeightsArray()
```
  - initialise
```
public void initialise(Population population)
```
    Initialises the first population and assigns values to sigma and xMean.
    
    Specified by:
    
    initialise in interface IRecombination
    
    Parameters:
    
    population -
  - recombinate
```
public Population recombinate(Population population,
                              IndividualGroup[] groups)
```
    Creates a new offspring population. The new population size has to be equal to the original population size.
    
    Specified by:
    
    recombinate in interface IRecombination
    
    Parameters:
    
    population - original population. For the moment only used to get the fitness errors
    
    groups - new individuals
    
    Returns:
    
    new Population
  - isDtooLarge
```
public boolean isDtooLarge()
```
    Condition exceeds 1e14
    
    Specified by:
    
    isDtooLarge in interface IRecombination
    
    Returns:
    
    true if D > 1e14

Class DifferentialRecombination

Field Summary

Constructor Summary

Method Summary

Methods inherited from class java.lang.Object

Field Detail

cc

cSigma

c1

cmu

dampSigma

pc

pSigma

sigma

B

C

D

eigeneval

invsqrtC

chiN

dimensions

xMean

mu

muEffective

weights

counteval

errorsNumber

Constructor Detail

DifferentialRecombination

Method Detail

createWeightsArray

initialise

recombinate

isDtooLarge