public NeuralNetwork[] AddHidden(NeuralNetwork[] neuralNetworkArgs, int percentNegative, int percentHidden)
{
List<NeuralNetwork> neuralNetwork = new List<NeuralNetwork>();
NeuralNetwork neuralNetworkMethods = new NeuralNetwork(0, 0, 0);
int randomInt = random.Next(101);
for (int x = 0; x < neuralNetworkArgs.Length; x++)
{
if (randomInt < percentHidden)
{
List<Neuron> input = new List<Neuron>(), hidden = new List<Neuron>(), output = new List<Neuron>();
foreach (Neuron neuronArgs in neuralNetworkArgs[x].input)
input.Add(neuronArgs);
foreach (Neuron neuronArgs in neuralNetworkArgs[x].output)
output.Add(neuronArgs);
for (int y = 0; y < neuralNetworkArgs[x].hidden.Length; y++)
{
hidden.Add(neuralNetworkArgs[x].hidden[y]);
if (y == neuralNetworkArgs[x].hidden.Length - 1)
{
if (randomInt < percentNegative)
hidden.Add(neuralNetworkMethods.CreateHidden(neuralNetworkArgs[x].hidden.Length, true));
else
hidden.Add(neuralNetworkMethods.CreateHidden(neuralNetworkArgs[x].hidden.Length, false));
neuralNetworkMethods.MutateNeuron(hidden[y + 1], output.ToArray());
neuralNetworkMethods.ConnectInput(input.ToArray(), hidden[y + 1]);
}
}
neuralNetwork.Add(new NeuralNetwork(input.Count, hidden.Count, output.Count));
neuralNetwork[x].EstablishExistingNetwork(input.ToArray(), hidden.ToArray(), output.ToArray());
}
else
{
neuralNetwork.Add(neuralNetworkArgs[x]);
}
randomInt = random.Next(101);
}
return neuralNetwork.ToArray();
}
public void ResetFish()
{
brain = new NeuralNetwork(2, random.Next(1, 4), 3);
brain.EstablishRandomNetwork();
position = new Vector2(random.Next((Game1.screenWidth - Game1.theoreticalScreenWidth) / 2, ((Game1.screenWidth - Game1.theoreticalScreenWidth) / 2) + Game1.theoreticalScreenWidth - texture.Width), random.Next((Game1.screenHeight - Game1.theoreticalScreenHeight) / 2, ((Game1.screenHeight - Game1.theoreticalScreenHeight) / 2) + Game1.theoreticalScreenHeight - texture.Height));
rectangle = new Rectangle((int)(position.X), (int)(position.Y), texture.Width, texture.Height);
collisionCircle = new CollisionCircle((int)(position.X - origin.X), (int)(position.Y - origin.Y));
rotation = 90;
}
public Fish(NeuralNetwork newBrain, CollisionCircle newCollisionCircle)
{
texture = content.Load<Texture2D>("Sprites/fish.png");
brain = newBrain;
position = new Vector2(random.Next((Game1.screenWidth - Game1.theoreticalScreenWidth) / 2, ((Game1.screenWidth - Game1.theoreticalScreenWidth) / 2) + Game1.theoreticalScreenWidth - texture.Width), random.Next((Game1.screenHeight - Game1.theoreticalScreenHeight) / 2, ((Game1.screenHeight - Game1.theoreticalScreenHeight) / 2) + Game1.theoreticalScreenHeight - texture.Height));
rectangle = new Rectangle((int)(position.X), (int)(position.Y), texture.Width, texture.Height);
origin = new Vector2(texture.Width / 2, texture.Height / 2);
collisionCircle = newCollisionCircle;
rotation = 90;
proximitySensor[0] = new ProximitySensor((int)(position.X), (int)(position.Y));
proximitySensor[1] = new ProximitySensor((int)(position.X), (int)(position.Y));
proximitySensor[2] = new ProximitySensor((int)(position.X), (int)(position.Y));
}
public NeuralNetwork[] Mutate(NeuralNetwork[] neuralNetworkArgs, int percentAmountMutate, int percentMutate)
{
List<NeuralNetwork> neuralNetwork = new List<NeuralNetwork>();
NeuralNetwork neuralNetworkMethods = new NeuralNetwork(0, 0, 0);
int randomInt = random.Next(101);
for (int x = 0; x < neuralNetworkArgs.Length; x++)
{
if (randomInt < percentAmountMutate)
{
List<Neuron> input = new List<Neuron>(), hidden = new List<Neuron>(), output = new List<Neuron>();
randomInt = random.Next(101);
for (int y = 0; y < neuralNetworkArgs[x].input.Length; y++)
{
if (randomInt < percentMutate)
input.Add(neuralNetworkMethods.MutateNeuron(neuralNetworkArgs[x].input[y], neuralNetworkArgs[x].hidden));
else
input.Add(neuralNetworkArgs[x].input[y]);
randomInt = random.Next(101);
}
for (int y = 0; y < neuralNetworkArgs[x].hidden.Length; y++)
{
if (randomInt < percentMutate)
hidden.Add(neuralNetworkMethods.MutateNeuron(neuralNetworkArgs[x].hidden[y], neuralNetworkArgs[x].output));
else
hidden.Add(neuralNetworkArgs[x].hidden[y]);
randomInt = random.Next(101);
}
foreach (Neuron neuron in neuralNetworkArgs[x].output)
output.Add(neuron);
neuralNetwork.Add(new NeuralNetwork(input.Count, hidden.Count, output.Count));
neuralNetwork[x].EstablishExistingNetwork(input.ToArray(), hidden.ToArray(), output.ToArray());
}
else
{
neuralNetwork.Add(neuralNetworkArgs[x]);
}
randomInt = random.Next(101);
}
return neuralNetwork.ToArray();
}
public NeuralNetwork[] CrossMutate(NeuralNetwork[] neuralNetworkArgs, int total)
{
List<NeuralNetwork> neuralNetwork = new List<NeuralNetwork>();
NeuralNetwork neuralNetworkMethods = new NeuralNetwork(0, 0, 0);
int randomInt = random.Next(neuralNetworkArgs.Length);
for (int x = 0; x < total; x++)
{
List<Neuron> input = new List<Neuron>(), hidden = new List<Neuron>(), output = new List<Neuron>();
List<int> additionalConnections = new List<int>();
for (int y = 0; y < neuralNetworkArgs[0].output.Length; y++)
{
output.Add(neuralNetworkArgs[randomInt].output[y]);
randomInt = random.Next(neuralNetworkArgs.Length);
}
for (int y = 0; y < neuralNetworkArgs[0].hidden.Length; y++)
{
if (neuralNetworkArgs[0].hidden.Length > neuralNetworkArgs[randomInt].hidden.Length && y > neuralNetworkArgs[randomInt].hidden.Length - 1)
{
hidden.Add(neuralNetworkMethods.CloneConnections(neuralNetworkArgs[0].hidden[y], output.ToArray(), neuralNetworkArgs[0].hidden[y].negative));
additionalConnections.Add(randomInt);
}
else
hidden.Add(neuralNetworkMethods.CloneConnections(neuralNetworkArgs[randomInt].hidden[y], output.ToArray(), neuralNetworkArgs[randomInt].hidden[y].negative));
randomInt = random.Next(neuralNetworkArgs.Length);
}
for (int y = 0; y < neuralNetworkArgs[0].input.Length; y++)
{
bool passable = true;
if (additionalConnections.Count > 0)
{
while (passable == false)
{
passable = true;
foreach (int z in additionalConnections)
{
if (randomInt == z)
passable = false;
}
randomInt = random.Next(neuralNetworkArgs.Length);
}
input.Add(neuralNetworkMethods.CloneConnections(neuralNetworkArgs[randomInt].input[y], hidden.ToArray(), neuralNetworkArgs[randomInt].input[y].negative));
}
else
input.Add(neuralNetworkMethods.CloneConnections(neuralNetworkArgs[randomInt].input[y], hidden.ToArray(), neuralNetworkArgs[randomInt].input[y].negative));
randomInt = random.Next(neuralNetworkArgs.Length);
}
neuralNetwork.Add(new NeuralNetwork(input.Count, hidden.Count, output.Count));
neuralNetwork[x].EstablishExistingNetwork(input.ToArray(), hidden.ToArray(), output.ToArray());
}
return neuralNetwork.ToArray();
}
