Package org.apache.wayang.profiler.log

Class GeneticOptimizer

java.lang.Object

org.apache.wayang.profiler.log.GeneticOptimizer

public class GeneticOptimizer
extends Object

Implementation of the genetic optimization technique for finding good LoadProfileEstimators.

Constructor Summary

Constructors

Constructor	Description
GeneticOptimizer(OptimizationSpace optimizationSpace, Collection<PartialExecution> observations, Map<String,DynamicLoadProfileEstimator> estimators, Map<Platform,Variable> platformOverheads, Configuration configuration)	Creates a new instance.

Method Summary

Modifier and Type	Method	Description
double	calculateObservationBasedWeight(PartialExecution observation)
double	calculateRuntimeBasedWeight(PartialExecution observation)
List<Individual>	createInitialPopulation()	Creates a population of random Individuals.
List<Individual>	evolve(List<Individual> population)
Bitmask	getActivatedGenes()
Configuration	getConfiguration()
Collection<PartialExecution>	getData()
HashMap<Object,Integer>	getNumObservations()
Collection<PartialExecution>	getObservations()
OptimizationSpace	getOptimizationSpace()
Map<Platform,Variable>	getPlatformOverheads()
static double	getSelectionProbability(double score1, double score2, double minScore)
void	updateFitness(List<Individual> individuals)	Update the fitness of the Individuals w.r.t. to this instance and sort them according to their new fitness.

Methods inherited from class java.lang.Object

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

Constructor Details

GeneticOptimizer

public GeneticOptimizer(OptimizationSpace optimizationSpace,
 Collection<PartialExecution> observations,
 Map<String,DynamicLoadProfileEstimator> estimators,
 Map<Platform,Variable> platformOverheads,
 Configuration configuration)

Creates a new instance.

Method Details

createInitialPopulation

public List<Individual> createInitialPopulation()

Creates a population of random Individuals.

Returns:

the Individuals ordered by their fitness

updateFitness

public void updateFitness(List<Individual> individuals)

Update the fitness of the Individuals w.r.t. to this instance and sort them according to their new fitness.

Parameters:

individuals - the Individuals

evolve

public List<Individual> evolve(List<Individual> population)

getSelectionProbability

public static double getSelectionProbability(double score1,
 double score2,
 double minScore)

getActivatedGenes

public Bitmask getActivatedGenes()

getData

public Collection<PartialExecution> getData()

getConfiguration

public Configuration getConfiguration()

getOptimizationSpace

public OptimizationSpace getOptimizationSpace()

getObservations

public Collection<PartialExecution> getObservations()

getNumObservations

public HashMap<Object,Integer> getNumObservations()

getPlatformOverheads

public Map<Platform,Variable> getPlatformOverheads()

calculateObservationBasedWeight

public double calculateObservationBasedWeight(PartialExecution observation)

calculateRuntimeBasedWeight

public double calculateRuntimeBasedWeight(PartialExecution observation)