void Start()
{
// create our neural networks for the genomes
species = new GeneticController(numGenomes, mutationRate);
cars = new GameObject[numSimulate];
carController = new AICarController[numSimulate];
bestCar = carController[0];
CarCheckPoint checkpoint = carFab.GetComponent <CarCheckPoint>();
checkpoint.track = track;
// Instantiate the number of cars we are going to simulate
// Add assign them an AI Controller
for (int i = 0; i < numSimulate; i++)
{
cars[i] = Instantiate(carFab, startingPos, carFab.transform.rotation);
carController[i] = cars[i].GetComponent <AICarController>();
carController[i].network = species.population[i];
}
currentGenome = numSimulate;
batchSimulate = numSimulate;
Network_GUI = Instantiate(Network_GUI);
UI_Genetics genetics = Network_GUI.GetComponentInChildren <UI_Genetics>();
genetics.academy = this;
networkUI = Network_GUI.GetComponentInChildren <UI_Network>();
networkUI.Display(carController[0].network);
}
void Start()
{
gController = new GeneticController(genomes, 60, brainStructure, outputs);
agents = new GameObject[genomes];
fitness = new double[genomes];
gController.createGeneration();
setEnvironment(49, 49);
Debug.Log("Generation: " + gController.getGeneration());
}
void Awake()
{
if (instance == null)
{
instance = this;
}
else if (instance != this)
{
Destroy(gameObject);
}
}
void Awake()
{
geneticController = new GeneticController <ChromosomeSalesman> ();
// Setting basic info to Chromosome
/*ChromosomeSalesman.cities = new List<SalesmanCity>();
* ChromosomeSalesman.cities.Add (new SalesmanCity (0));
* ChromosomeSalesman.cities.Add (new SalesmanCity (1));
* ChromosomeSalesman.cities.Add (new SalesmanCity (2));
*
* ChromosomeSalesman solution = geneticController.Calculate ();
* Debug.Log ( solution.toString() );*/
}
public GeneticBrain(int inputSize, int hiddenLayerSize, int outputSize, int minRandom, int maxRandom, GeneticController control)
{
this.inputSize = inputSize;
this.hiddenLayerSize = hiddenLayerSize;
this.outputSize = outputSize;
ihWeights = new Matrix(hiddenLayerSize, inputSize);
hoWeights = new Matrix(outputSize, hiddenLayerSize);
ihWeights.Randomize(minRandom, maxRandom);
hoWeights.Randomize(minRandom, maxRandom);
hBias = new Matrix(hiddenLayerSize, 1);
oBias = new Matrix(outputSize, 1);
hBias.Randomize(minRandom, maxRandom);
oBias.Randomize(minRandom, maxRandom);
this.fitness = 0;
this.control = control;
}
static void Main(string[] args)
{
Console.Title = "Genetic Algorithm + Neural Network ~ Testing";
// Testing Genetic Algorithm combined with Neural Network of 200 neurons
GeneticController GController = new GeneticController(12, 3, new int[] { 80, 120 }, 1);
// Learning process for 100 generations
for (int i = 0; i < 100; i++)
{
GController.createGeneration();
printResultsAndSetFitness(GController);
}
Console.ForegroundColor = ConsoleColor.White;
Console.Write("\nPress a key to exit...");
Console.ReadLine();
}
private void Form1_Load(object sender, EventArgs e)
{
ColumnHeader logHeader = new ColumnHeader();
geneticController = new GeneticController(this);
bestInds = new List <BitIndividual>();
averageInds = new List <BitIndividual>();
worstInds = new List <BitIndividual>();
bestIndsBackup = new List <BitIndividual>();
averageIndsBackup = new List <BitIndividual>();
worstIndsBackup = new List <BitIndividual>();
LogListView.Columns.Add(logHeader);
LogListView.Scrollable = true;
LogListView.HeaderStyle = ColumnHeaderStyle.None;
LogListView.View = View.Details;
LogListView.AutoResizeColumns(ColumnHeaderAutoResizeStyle.HeaderSize);
geneticChart.ChartAreas[0].AxisX.ScaleView.Zoom(0, 50);
geneticChart.ChartAreas[0].CursorX.IsUserEnabled = true;
geneticChart.ChartAreas[0].CursorX.IsUserSelectionEnabled = true;
geneticChart.ChartAreas[0].AxisX.ScaleView.Zoomable = true;
geneticChart.ChartAreas[0].AxisX.ScrollBar.IsPositionedInside = true;
}
static void printResultsAndSetFitness(GeneticController g_ctrl)
{
Console.Clear();
double[][] result = g_ctrl.executeGeneration(new double[] { 0, 1, -.5 });
double[] fitness = new double[result.Length];
int index = 0;
foreach (double[] r in result)
{
if (r[0] <= .4)
{
Console.ForegroundColor = ConsoleColor.Red;
}
else if (r[0] >= .6)
{
Console.ForegroundColor = ConsoleColor.Green;
}
else
{
Console.ForegroundColor = ConsoleColor.White;
}
Console.WriteLine(r[0]);
if (r[0] > .4 && r[0] < .6)
{
fitness[index] = 1 - (10 * Math.Abs(.5 - r[0])); // Funzione di fitness
}
else
{
fitness[index] = 0;
}
index++;
}
g_ctrl.setFitness(fitness);
Console.WriteLine();
}
private void MakeGenetic()
{
geneticController = Instantiate(geneticControllerPrefab);
Attach(geneticController.gameObject);
}
