public double[][] getReferenceSet() {
openga.applications.data.parallelMachine data1 = new openga.applications.data.parallelMachine();
if (numberOfJob == 35) {
return data1.getParetoJob35M10();
} else if (numberOfJob == 50) {
return data1.getParetoJob50M15();
} else { // for job = 65
return data1.getParetoJob65M18();
}
}
public void initiateVars() {
// initiate scheduling data. We use the benchmark problem.
openga.applications.data.parallelMachine data1 = new openga.applications.data.parallelMachine();
dueDay = data1.getTestData3_DueDay(numberOfJob);
processingTime = data1.getTestData3_processingTime(numberOfJob);
weights = new double[numberOfSubPopulations][numberOfObjs];
objectiveMinimization = new boolean[numberOfObjs];
objectiveMinimization[0] = true;
objectiveMinimization[1] = true;
encodeType = true;
// GA objects
GaMain = new MainWeightScalarizationI[numberOfSubPopulations];
Population = new populationI[numberOfSubPopulations];
Selection = new SelectI[numberOfSubPopulations];
Crossover = new CrossoverI[numberOfSubPopulations];
// Crossover2 = new CrossoverI[numberOfSubPopulations];
Mutation = new MutationI[numberOfSubPopulations];
// Mutation2 = new MutationI[numberOfSubPopulations];
// clone1 = new cloneI[numberOfSubPopulations];
ObjectiveFunction = new ObjectiveFunctionScheduleI[numberOfSubPopulations][numberOfObjs];
Fitness = new FitnessI[numberOfSubPopulations];
for (int i = 0; i < numberOfSubPopulations; i++) {
GaMain[i] =
new SPGAwithSharedParetoSet(); // singleThreadGA SPGAwithSharedParetoSet adaptiveGA
Population[i] = new population();
if (tournamentSize
== 2) { // binaryTournament similaritySelection2 tenaryTournament varySizeTournament
Selection[i] = new binaryTournament();
} else if (tournamentSize == 5) {
Selection[i] = new quintetTournament();
} else if (tournamentSize == 10) {
Selection[i] = new tenaryTournament();
} else {
Selection[i] = new twentyTournament();
}
// random generate solutionVectorClone solutionVectorCloneWithMutation
// solutionVectorCloneAdjacentSwapWithMutation solutionVectorCloneWithMiningGene
clone1 = new solutionVectorCloneWithMutation();
Crossover[i] =
new twoPointCrossover2(); // oneByOneChromosomeCrossover twoPointCrossover2
// twoPointCrossover2withAdpative
Crossover2 = new PMX();
Mutation[i] = new swapMutation(); // swapMutation shiftMutation swapMutationWithAdaptive
Mutation2 = new shiftMutation(); // shiftMutation
ObjectiveFunction[i] = new ObjectiveFunctionScheduleI[numberOfObjs];
ObjectiveFunction[i][0] = new ObjectiveMakeSpan(); // the first objective, makespan
ObjectiveFunction[i][1] = new ObjectiveTardiness(); // the second one.
Fitness[i] = new FitnessByScalarizedM_objectives();
// set schedule data to the objectives
ObjectiveFunction[i][0].setScheduleData(processingTime, numberOfMachines);
ObjectiveFunction[i][1].setScheduleData(dueDay, processingTime, numberOfMachines);
// set the data to the GA main program.
GaMain[i].setData(
Population[i],
Selection[i],
Crossover[i],
Mutation[i],
ObjectiveFunction[i],
Fitness[i],
firstIteration,
DEFAULT_PopSize * 2,
DEFAULT_PopSize,
numberOfJob,
DEFAULT_crossoverRate,
DEFAULT_mutationRate,
objectiveMinimization,
numberOfObjs,
encodeType,
elitism);
// set weight data
weights[i] = calcWeightsForEachSubPop(i);
GaMain[i].setWeight(weights[i]);
// set secondary crossover and mutation operator.
GaMain[i].setSecondaryCrossoverOperator(Crossover2, applySecCRX);
GaMain[i].setSecondaryMutationOperator(Mutation2, applySecMutation);
// GaMain[i].setCloneOperatpr(clone1, applyClone);
}
}