private void Run()
{
GA ga = new GA();
double[][] input = new double[4][];
input[0] = new double[] { 1, 1 };
input[1] = new double[] { 0, 1 };
input[2] = new double[] { 1, 0 };
input[3] = new double[] { 0, 0 };
double[][] output = new double[4][];
output[0] = new double[] { 0 };
output[1] = new double[] { 1 };
output[2] = new double[] { 1 };
output[3] = new double[] { 0 };
List <NeuralNet> result = ga.DetermineFitness(ga.SpawnGeneration(1000000, 2, 0, 1), input, output);
result = ga.SelectParents(result);
for (int i = 0; i < 2; i++)
{
net = result[i];
PrintOut(0, 1);
PrintOut(1, 0);
PrintOut(0, 0);
PrintOut(1, 1);
}
/* Random random = new Random();
* int seed = 0;
*
* double[][] input = new double[4][];
* input[0] = new double[] { 1, 1 };
* input[1] = new double[] { 0, 1 };
* input[2] = new double[] { 1, 0 };
* input[3] = new double[] { 0, 0 };
*
* double[][] output = new double[4][];
* output[0] = new double[] { 0 };
* output[1] = new double[] { 1 };
* output[2] = new double[] { 1 };
* output[3] = new double[] { 0 };
*
* net.Init(3, 2, 4, 1);
* net.Train(input, output, 7, 8);
*
* PrintOut(0, 1);
* PrintOut(1, 0);
* PrintOut(0, 0);
* PrintOut(1, 1);*/
Console.Read();
}
public void CalculateErrors(NeuralNet neuralNet, double[] expectedOutput)
{
//Calculate outputlayer errors
for (int i = 0; i < neuralNet.outputLayer.neurons.Count; i++)
{
double temp = neuralNet.outputLayer.neurons[i].Output;
neuralNet.outputLayer.neurons[i].error = (expectedOutput[i] - temp) * temp * (1.0F - temp);
}
//Calculate hiddenlayer errors
for (int i = 0; i < neuralNet.hiddenLayer.neurons.Count; i++)
{
double error = 0;
Neuron temp = neuralNet.hiddenLayer.neurons[i];
for (int a = 0; a < neuralNet.outputLayer.neurons.Count; a++)
{
Neuron outputNode = neuralNet.outputLayer.neurons[a];
error += outputNode.error * outputNode.Input[i].Weight * temp.Output * (1.0 - temp.Output);
}
neuralNet.hiddenLayer.neurons[i].error = error;
}
}
private void AdjuestNet(NeuralNet neuralNet, double learningrate, double[] expected)
{
for (int i = 0; i < neuralNet.hiddenLayer.neurons.Count; i++)
{
Neuron node = neuralNet.hiddenLayer.neurons[i];
for (int i2 = 0; i2 < neuralNet.outputLayer.neurons.Count; i2++)
{
Neuron output = neuralNet.outputLayer.neurons[i2];
double momentum = 0;
for (int i3 = 0; i3 < expected.Count(); i3++)
{
momentum = Math.Abs(expected[i3] - output.Output);
}
output.Input[i].Weight += momentum * learningrate * output.error * node.Output;
output.delta += momentum * learningrate * output.error * output.weight;
}
}
for (int i = 0; i < neuralNet.inputLayer.neurons.Count; i++)
{
Neuron node = neuralNet.inputLayer.neurons[i];
for (int i2 = 0; i2 < neuralNet.hiddenLayer.neurons.Count; i2++)
{
Neuron output = neuralNet.hiddenLayer.neurons[i2];
double momentum = 0;
for (int i3 = 0; i3 < expected.Count(); i3++)
{
momentum = Math.Abs(expected[i3] - neuralNet.outputLayer.neurons[i3].Output);
}
output.Input[i].Weight += momentum * learningrate * output.error * node.Output;
output.delta += momentum * learningrate * output.error * output.weight;
}
}
}
