Packagecom.suckatmath.machinelearning.genetic.core
Classpublic class GeneticEngine
InheritanceGeneticEngine Inheritance Object
Implements EvolvableComparator

performs actual genetic algorithms


Public Properties
 PropertyDefined By
  numNewPerGeneration : int = 1
Number of new random genomes per generation.
GeneticEngine
Public Methods
 MethodDefined By
  
GeneticEngine(f:EvolvableFactory, mutate:Boolean = true, crossover:Boolean = true, npg:int = 1, mprob:Number = 0.5, cparents:int = 2, ccross:int = 1)
new engine constructor
GeneticEngine
  
compareFitness(a:Evolvable, b:Evolvable):int
Used both in ROULETTE and as default EvolvableComparator function for TOURNAMENT selection
GeneticEngine
  
getAvgPopFitness():Number
return the average fitness of the population.
GeneticEngine
  
getPopulation():Array
gets population.
GeneticEngine
  
nextGeneration():void
main work method.
GeneticEngine
  
randomPopulation(psize:int):void
generate a new population of psize.
GeneticEngine
  
setCarryOverNum(n:int):void
how many of the fittest of previous generation to put in new pop.
GeneticEngine
  
setCrossoverNumParents(n:int):void
set how many parents to use for crossover - should be <= number of crossover points +1
GeneticEngine
  
setCrossoverNumPoints(n:int):void
set number of crossover points
GeneticEngine
  
setFactory(fac:EvolvableFactory):void
set EvolvableFactory to use to create new Evolvables
GeneticEngine
  
setMutateProbability(n:Number):void
set probability of mutation in generating new individuals
GeneticEngine
  
setNumNewPerGeneration(n:int):void
set number of new random individuals per generation.
GeneticEngine
  
setPopulation(pop:Array):void
set population to array of Evolvables
GeneticEngine
  
setSelectionMode(s:int):void
set selectionMode to ROULETTE or TOURNAMENT
GeneticEngine
  
setTournamentComparator(e:EvolvableComparator):void
set EvolvableCOmparator to use during TOURNAMENT selection
GeneticEngine
  
useCrossover(c:Boolean):void
set whether to use crossover in generating new individuals
GeneticEngine
  
useMutate(m:Boolean):void
set whether to use mutation in generating new individuals
GeneticEngine
Public Constants
 ConstantDefined By
  ROULETTE : int = 1
[static] constant for ROULETTE selection : proportional selection.
GeneticEngine
  TOURNAMENT : int = 2
[static] constant for TOURNAMENT selection
GeneticEngine
Property Detail
numNewPerGenerationproperty
public var numNewPerGeneration:int = 1

Number of new random genomes per generation.

Constructor Detail
GeneticEngine()Constructor
public function GeneticEngine(f:EvolvableFactory, mutate:Boolean = true, crossover:Boolean = true, npg:int = 1, mprob:Number = 0.5, cparents:int = 2, ccross:int = 1)

new engine constructor

Parameters
f:EvolvableFactory — EvolvableFactory to generate new Evolvables from Genomes
 
mutate:Boolean (default = true) — Boolean whether to use mutation
 
crossover:Boolean (default = true) — Boolean whether to use crossover
 
npg:int (default = 1) — int number of new random individuals per generation
 
mprob:Number (default = 0.5) — Number probablity of mutation
 
cparents:int (default = 2) — int Number of parents to use for crossover
 
ccross:int (default = 1) — int Number of points to cross over in crossover operation
Method Detail
compareFitness()method
public function compareFitness(a:Evolvable, b:Evolvable):int

Used both in ROULETTE and as default EvolvableComparator function for TOURNAMENT selection

Parameters

a:Evolvable
 
b:Evolvable

Returns
int —
getAvgPopFitness()method 
public function getAvgPopFitness():Number

return the average fitness of the population. This may be an EXPENSIVE operation if getFitness is expensive

Returns
Number —
getPopulation()method 
public function getPopulation():Array

gets population.

Returns
Array —
nextGeneration()method 
public function nextGeneration():void

main work method. Calls to this will change the population array.

randomPopulation()method 
public function randomPopulation(psize:int):void

generate a new population of psize. OR if population is already an existing array, replace it with an array of the same size which is full of random individuals

Parameters

psize:int — int

setCarryOverNum()method 
public function setCarryOverNum(n:int):void

how many of the fittest of previous generation to put in new pop. Not used in TOURNAMENT selection

Parameters

n:int

setCrossoverNumParents()method 
public function setCrossoverNumParents(n:int):void

Parameters

n:int

setCrossoverNumPoints()method 
public function setCrossoverNumPoints(n:int):void

set number of crossover points

Parameters

n:int

setFactory()method 
public function setFactory(fac:EvolvableFactory):void

set EvolvableFactory to use to create new Evolvables

Parameters

fac:EvolvableFactory

setMutateProbability()method 
public function setMutateProbability(n:Number):void

set probability of mutation in generating new individuals

Parameters

n:Number

setNumNewPerGeneration()method 
public function setNumNewPerGeneration(n:int):void

set number of new random individuals per generation. 1 or 0 usually.

Parameters

n:int

setPopulation()method 
public function setPopulation(pop:Array):void

set population to array of Evolvables

Parameters

pop:Array

setSelectionMode()method 
public function setSelectionMode(s:int):void

set selectionMode to ROULETTE or TOURNAMENT

Parameters

s:int

setTournamentComparator()method 
public function setTournamentComparator(e:EvolvableComparator):void

set EvolvableCOmparator to use during TOURNAMENT selection

Parameters

e:EvolvableComparator

useCrossover()method 
public function useCrossover(c:Boolean):void

set whether to use crossover in generating new individuals

Parameters

c:Boolean

useMutate()method 
public function useMutate(m:Boolean):void

set whether to use mutation in generating new individuals

Parameters

m:Boolean

Constant Detail
ROULETTEConstant
public static const ROULETTE:int = 1

constant for ROULETTE selection : proportional selection. This is the default selection mode

TOURNAMENTConstant 
public static const TOURNAMENT:int = 2

constant for TOURNAMENT selection




Information on this site is by default visible to everyone within GODPAPER. Please ensure you apply appropriate access controls when posting GODPAPER Confidential information.Copyright © by GODPAPER. All material on this collaboration platform is the property of GODPAPER.

Thu Jan 16 2014, 01:11 PM +08:00