gpsys.symreg
Class SymRegFitness

java.lang.Object
  |
  +--gpsys.Fitness
        |
        +--gpsys.symreg.SymRegFitness
public class SymRegFitness
extends Fitness

The Fitness class for the symbolic regression problem. The fitness is measured using two values. The first is the error between the real function and the function generated by the GP system. The second value is the complexity of the GP program. The fitness of one GP is is deemed better than another if either the error is smaller or if the error is equal, but the complexity is smaller. The termination criteria is met when the error is zero. This problem involves finding the function :- 

			x^3 + 2x^2 +3x + 5

Version:
$Revision: 1.1 $, $Date: 2000/07/05 16:25:51 $
Author:
Adil Qureshi
Department of Computer Science,
University College London,
Gower St,
London WC1E 6BT,
UK.
See Also:
Serialized Form

Field Summary
(package private)  int complexity
          The complexity of the GP i.e.
(package private)  double fitness
          The error between the ideal function and the GP.
static float from
          The minimum value of x to be used for testing the evolved function.
static int samples
          The number of samples used test the evolved function.
static float to
          The maximum value of x to be used for testing the evolved function.
 
Constructor Summary
SymRegFitness()
          Construct a Fitness object with default Fitness.
SymRegFitness(GPParameters gpParameters, Individual i)
          Constructs a Fitness object by evaluating an Individual.
 
Method Summary
 void add(Fitness f)
          Adds a fitness value to this fitness value.
 void divide(int divisor)
          Divide the fitness by the specified integer.
 boolean equals(Fitness f)
          Tests if this fitness is equal to another fitness.
 boolean greaterThan(Fitness f)
          Tests if this fitness value is greater than another fitness value.
 Fitness instance()
          Creates a new instance of the Fitness object with a default fitness.
 Fitness instance(GPParameters gpParameters, Individual i)
          Creates a new instance of the Fitness object which represents the fitness of the specified individual.
 boolean lessThan(Fitness f)
          Tests if this fitness is less than another fitness.
 boolean terminationCondition()
          Tests whether this fitness meets the termination criteria.
 java.lang.String toString()
          Converts the fitness into a String suitable for printing.
 
Methods inherited from class java.lang.Object
, clone, equals, finalize, getClass, hashCode, notify, notifyAll, registerNatives, wait, wait, wait
 

Field Detail

fitness

double fitness
The error between the ideal function and the GP.

complexity

int complexity
The complexity of the GP i.e. the number of Genes it contains.

from

public static float from
The minimum value of x to be used for testing the evolved function.

to

public static float to
The maximum value of x to be used for testing the evolved function.

samples

public static int samples
The number of samples used test the evolved function.
Constructor Detail

SymRegFitness

public SymRegFitness()
Construct a Fitness object with default Fitness.

SymRegFitness

public SymRegFitness(GPParameters gpParameters,
                     Individual i)
Constructs a Fitness object by evaluating an Individual.
Parameters:
gpParameters - The GP parameters for this run.
i - The individual to be evaluated.
Method Detail

add

public void add(Fitness f)
Adds a fitness value to this fitness value.
Parameters:
f - the fitness value to be added.
Overrides:
add in class Fitness

divide

public void divide(int divisor)
Divide the fitness by the specified integer. This is used by the GP system to calculate the average fitness of the population.
Parameters:
divisor - the integer to divide the fitness by.
Overrides:
divide in class Fitness

greaterThan

public boolean greaterThan(Fitness f)
Tests if this fitness value is greater than another fitness value.
Parameters:
f - the fitness with which to compare.
Returns:
true if f has higher fitness, false otherwise.
Overrides:
greaterThan in class Fitness

lessThan

public boolean lessThan(Fitness f)
Tests if this fitness is less than another fitness.
Parameters:
f - the fitness with which to compare.
Returns:
true if f has less fitness, false otherwise.
Overrides:
lessThan in class Fitness

equals

public boolean equals(Fitness f)
Tests if this fitness is equal to another fitness.
Parameters:
f - the fitness with which to compare.
Returns:
true if f has the same fitness, false otherwise.
Overrides:
equals in class Fitness

instance

public Fitness instance()
Creates a new instance of the Fitness object with a default fitness.
Returns:
an instance of a Fitness object with default fitness.
Overrides:
instance in class Fitness

instance

public Fitness instance(GPParameters gpParameters,
                        Individual i)
Creates a new instance of the Fitness object which represents the fitness of the specified individual.
Parameters:
gpParameters - the parameters for this GP run.
i - the individual to be evaluated.
Returns:
The Fitness of the specified individual.
Overrides:
instance in class Fitness

terminationCondition

public boolean terminationCondition()
Tests whether this fitness meets the termination criteria.
Returns:
true if the termination criteria has been met, false otherwise.
Overrides:
terminationCondition in class Fitness

toString

public java.lang.String toString()
Converts the fitness into a String suitable for printing.
Returns:
A String representing the fitness.
Overrides:
toString in class Fitness