public void SaveLearningProcess(string path)
{
if (learningProcess.Serialize(path))
{
Debug.Log("Zapisano w " + path + " , smochodzików " + learningProcess.PopulationCount);
}
else
{
Debug.Log("Błąd serializacji!");
}
}
public void Run()
{
Stopwatch t = new Stopwatch();
t.Start();
int counter = 0;
while (true)
{
double[] scores = new double[populationCount];
var testData = new[]
{
new[] { 0.2, -1.1, 0.9 },
new[] { 0.3, -2.1, -3 },
new[] { 1, -1, 1.0 },
new[] { 0, 0, 0.0 },
new[] { -0.2, 1.1, -0.9 },
};
double[,] scores2 = new double[populationCount, testData.Length];
for (int k = 0; k < testData.Length; k++)
{
for (int i = 0; i < populationCount; i++)
{
var result = process.Population[i].Calculate(testData[k]);
var a = FitnessFunction(testData[k], result);
scores[i] += a;
scores2[i, k] = a;
}
}
process.Learn(scores);
var data = process.HistoricalData.Last();
Console.WriteLine(
$"Gen: {data.GenerationIndex} Max: {data.BestScore} Min: {data.WorstScore} Avg: {data.AverageScore} Med: {data.MedianScore}");
//Thread.Sleep(1000);
if (counter == 50)
{
var test = new[]
{
rand.NextDouble() * 2 - 1,
rand.NextDouble() * 2 - 1,
rand.NextDouble() * 2 - 1
};
var spec = process.HistoricalData.Last().BestSpecimen;
for (int i = 0; i < testData.Length; i++)
{
Console.WriteLine("----------------------------");
var a = spec.Calculate(testData[i]);
Console.WriteLine($"network: {a[0]} {a[1]}");
var a2 = Problem(testData[i]);
Console.WriteLine($"exptected: {a2[0]} {a2[1]}");
Console.WriteLine("Fitness " + FitnessFunction(testData[i], a));
}
Console.WriteLine("----------------------------");
var a1 = spec.Calculate(test);
Console.WriteLine($"network: {a1[0]} {a1[1]}");
var fit = FitnessFunction(test, a1);
a1 = Problem(test);
Console.WriteLine($"exptected: {a1[0]} {a1[1]}");
var neuron0 = spec.Layers[0].Neurons[0].GetWeights();
var neuron1 = spec.Layers[0].Neurons[1].GetWeights();
Console.WriteLine("Fitness " + fit);
Console.WriteLine("----------------------------");
Console.WriteLine(
$"({Math.Round(neuron0[0], 2)}x1 {Math.Round(neuron0[1], 2)}x2 {Math.Round(neuron0[2], 2)}x3 {Math.Round(neuron0[3],3)})({Math.Round(neuron1[0], 2)}x1 {Math.Round(neuron1[1], 2)}x2 {Math.Round(neuron1[2], 2)}x3 {Math.Round(neuron1[3], 3)})");
Console.WriteLine("Expected: ");
Console.WriteLine($"(x1 + x2 - 2x3 + 0)(x1 - x2 + x3 + 1)");
t.Stop();
TimeSpan ts = t.Elapsed;
string elapsedTime = String.Format("{0:00}:{1:00}:{2:00}.{3:00}",
ts.Hours, ts.Minutes, ts.Seconds,
ts.Milliseconds / 10);
Console.WriteLine("RunTime " + elapsedTime);
var key = Console.ReadKey();
if (key.Key == ConsoleKey.S)
{
process.Serialize("leariningtest.xml");
}
counter = 0;
t.Reset();
t.Start();
}
counter++;
}
}
