public void initiateVars() {
GaMain =
new singleThreadGAwithEDA(); // singleThreadGA singleThreadGAwithSecondFront
// singleThreadGAwithMultipleCrossover adaptiveGA
Population = new population();
Selection = new binaryTournament(); // binaryTournament
Crossover =
new twoPointCrossover2EDA(); // twoPointCrossover2 oneByOneChromosomeCrossover
// twoPointCrossover2withAdpative
// twoPointCrossover2withAdpativeThreshold
Mutation =
new swapMutationEDA(); // shiftMutation shiftMutationWithAdaptive
// shiftMutationWithAdaptiveThreshold
ObjectiveFunction = new ObjectiveFunctionTSPI[numberOfObjs];
ObjectiveFunction[0] = new ObjectiveFunctionMTSPHeu();
Fitness = new singleObjectiveFitness();
objectiveMinimization = new boolean[numberOfObjs];
objectiveMinimization[0] = true;
encodeType = true;
if (numberOfSalesmen >= length) {
System.out.println(
"The number of salesmen is "
+ numberOfSalesmen
+ ", which should be greater than the number of visiting locations.");
System.out.println("The program will exit.");
System.exit(0);
}
localSearch1 = new localSearchBy2OptForMTSP();
// localSearch1.setMTSPData(Population, distanceMatrix, numberOfSalesmen);
ObjectiveFunction[0].setTSPData(originalPoint, coordinates, numberOfSalesmen);
ObjectiveFunction[0].setObjectiveFunctionType("TotalDistance");
// set the data to the GA main program.
GaMain.setData(
Population,
Selection,
Crossover,
Mutation,
ObjectiveFunction,
Fitness,
DEFAULT_generations,
DEFAULT_PopSize,
DEFAULT_PopSize,
length,
DEFAULT_crossoverRate,
DEFAULT_mutationRate,
objectiveMinimization,
numberOfObjs,
encodeType,
elitism);
GaMain.setLocalSearchOperator(localSearch1, applyLocalSearch, 20);
GaMain.setSecondaryCrossoverOperator(Crossover2, false);
GaMain.setSecondaryMutationOperator(Mutation2, false);
GaMain.setEDAinfo(
lamda, numberOfCrossoverTournament, numberOfMutationTournament, startingGenDividen);
}
public static void main(String[] args) {
System.out.println("mTSPEDA_20150731DOE_totalDistance");
double crossoverRate[], mutationRate[];
crossoverRate = new double[] {1, 0.5}; // 1, 0.5
mutationRate = new double[] {0.1, 0.5}; // 1, 0.5
int counter = 0;
double elitism[] = new double[] {0.1};
int generations[] = new int[] {1000}; // 50000
int numInstances = 33; // 33
int numberOfSalesmen[] = new int[] {3, 5, 10, 20, 30}; // 2, 3, 5, 10, 20, 30
int repeat = 5;
// EDA parameters.
double[] lamda = new double[] {0.1}; // learning rate{0.1, 0.5, 0.9}
int numberOfCrossoverTournament[] = new int[] {2}; // {1, 2, 4}
int numberOfMutationTournament[] = new int[] {2}; // {1, 2, 4}
int startingGenDividen[] = new int[] {4}; // {2, 4}{4}
// to test different kinds of combinations.
for (int i = 0; i <= numInstances; i++) { // numInstances
// initiate scheduling data, we get the data from a program.
openga.applications.data.TSPInstances TSPInstances1 =
new openga.applications.data.TSPInstances();
String instanceName = TSPInstances1.getFileName(i); // getFileName getCaterFileName
TSPInstances1.setData(instanceName);
TSPInstances1.getDataFromFile();
TSPInstances1.calcEuclideanDistanceMatrix();
int length = TSPInstances1.getSize();
for (int j = 0; j < crossoverRate.length; j++) {
for (int k = 0; k < mutationRate.length; k++) {
for (int n = 0; n < elitism.length; n++) {
for (int p = 0; p < numberOfSalesmen.length; p++) {
for (int q = 0; q < lamda.length; q++) {
for (int m = 0; m < numberOfCrossoverTournament.length; m++) {
for (int r = 0; r < numberOfMutationTournament.length; r++) {
for (int s = 0; s < startingGenDividen.length; s++) {
for (int t = 0; t < repeat; t++) {
System.out.print("counter " + counter + "\t");
mTSPSGA_SGGA TSP1 = new mTSPSGA_SGGA();
TSP1.setParameter(
i,
crossoverRate[j],
mutationRate[k],
counter,
elitism[n],
generations[0],
numberOfSalesmen[p],
TSPInstances1.getOriginalPoint(),
TSPInstances1.getCoordinates(),
TSPInstances1.getDistanceMatrix(),
TSPInstances1.getSize(),
instanceName);
TSP1.setEDAinfo(
lamda[q],
numberOfCrossoverTournament[m],
numberOfMutationTournament[r],
startingGenDividen[s]);
TSP1.initiateVars();
TSP1.startMain();
counter++;
}
}
}
}
}
}
}
}
}
} // end for
System.exit(0);
}