Package de.uka.ipd.sdq.dsexplore.opt4j.optimizer.heuristic.startingPopulation.impl

Class StartingPopulationHeuristicImpl

java.lang.Object

de.uka.ipd.sdq.dsexplore.opt4j.optimizer.heuristic.startingPopulation.AbstractStartingPopulationHeuristic

de.uka.ipd.sdq.dsexplore.opt4j.optimizer.heuristic.startingPopulation.impl.StartingPopulationHeuristicImpl

All Implemented Interfaces:

IStartingPoulationHeuristic

public class StartingPopulationHeuristicImpl
extends AbstractStartingPopulationHeuristic

Class implements a basic strategy to implement a heuristic that generates a starting population. The user specifies minNumerOfResourceContainers and maxNumberOfResourceContainers. Then the algorithm will generate numberOfCandidatesPerAllocationLevel individuals for each resource container number ("allocation level") between minNumberOfResourceContainers and maxNumberOfResourceContainers using a random allocation of components (resp. allocation contexts) to resource containers and the default processing rate (as used by firstIndividual. For each allocation level all individuals are evaluated and the Pareto optimal candidates are put into the population. Then four new individuals are generated for each Pareto optimal individual in the population. These new individuals use four different processing rates: minimum processing rate, maximum processing rate, lower processing rate (mean between default and minimum) and higher processing rate (mean between default and maximum). All of them are put into the population as well.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	StartingPopulationHeuristicImpl.PROCESSING_RATE_LEVEL

Constructor Summary

Constructors

Constructor	Description
StartingPopulationHeuristicImpl(DSEWorkflowConfiguration configuration)

Method Summary

Modifier and Type	Method	Description
static ArrayList<Integer>	getNumberOfComponentsPerResourceContainer(int numberOfComponents, int numberOfResourceContainers)
Collection<DSEIndividual>	getStartingPopulation(org.opt4j.core.optimizer.IndividualCompleter completer, org.opt4j.core.IndividualFactory individualFactory, DSEIndividual baseIndividual)	Used to generate starting population as described in class documentation.

Methods inherited from class de.uka.ipd.sdq.dsexplore.opt4j.optimizer.heuristic.startingPopulation.AbstractStartingPopulationHeuristic

getAllocationContexts, getResourceContainers

Methods inherited from class java.lang.Object

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

Constructor Detail

StartingPopulationHeuristicImpl

public StartingPopulationHeuristicImpl(DSEWorkflowConfiguration configuration)

Parameters:

configuration - As defined by the user in the GUI.

Method Detail

getNumberOfComponentsPerResourceContainer

public static ArrayList<Integer> getNumberOfComponentsPerResourceContainer(int numberOfComponents,
                                                                           int numberOfResourceContainers)

Parameters:

numberOfComponents -

numberOfResourceContainers -

Returns:

Calculates the number of components on each resource container when using numberOfComponents allocation contexts and numberOfResourceContainers resource containers. Integer represents number of components/allocation contexts and each resource container. Size of array will be numberOfResourceContainers.

getStartingPopulation

public Collection<DSEIndividual> getStartingPopulation(org.opt4j.core.optimizer.IndividualCompleter completer,
                                                       org.opt4j.core.IndividualFactory individualFactory,
                                                       DSEIndividual baseIndividual)

Used to generate starting population as described in class documentation. FIXME: Must only use those servers that are allowed in designdecisions.

Parameters:

completer - Used to evaluate inidividuals

individualFactory - Used to copy individual

baseIndividual - This individual is used as a starting point for the creation of all other individuals

Returns: