/// <summary>
/// EQUATION 3
/// </summary>
private Vector <double> MotionDirection(Krill krill, int currentIteration)
{
Vector <double> alpha_local = AlphaLocal(krill); // i-th Krill
Vector <double> alpha_target = AlphaTarget(krill, currentIteration); // i-th Krill
return(alpha_local + alpha_target);
}
/// <summary>
/// EQUATION 15
/// </summary>
/// <returns></returns>
private Vector <double> Beta_i_best(Krill krill)
{
if (krill.BestFitness > krill.Fitness)
{
return(X_i_j(krill.Coordinates, krill.BestCoordinates).Multiply(K_i_j(krill.Fitness, krill.BestFitness)));
}
else
{
return(Vector <double> .Build.Dense(krill.Coordinates.Count));
}
}
/// <summary>
/// EQUATION 10
/// </summary>
public Vector <double> GetForagingMotion(Krill krill, int currentIteration, Vector <double> vf_position)
{
int lastIteration = currentIteration - 1;
Vector <double> F_i_old = ForagingSpeedHistory.ContainsKey(new Tuple <int, int>(lastIteration, krill.KrillNumber))
? ForagingSpeedHistory[new Tuple <int, int>(lastIteration, krill.KrillNumber)]
: Vector <double> .Build.Dense(krill.Coordinates.Count);
double Omega_f = (0.1 + (0.8 * (1 - currentIteration / MaxIteration)));
Vector <double> B_i = (Beta_i_best(krill) + Beta_i_food(krill, lastIteration, vf_position)); // EQUATION 11
Vector <double> F_i = B_i.Multiply(V_f) + F_i_old.Multiply(Omega_f); // EQUATION 10
ForagingSpeedHistory.Add(new Tuple <int, int>(currentIteration, krill.KrillNumber), F_i);
return(F_i);
}
/// <summary>
/// EQUATION 7
/// </summary>
private double SensingDistance(Krill krill)
{
double N = krillPopulation.Population.Count;
double firstPart = (1 / (5 * N));
double secondPart = 0;
foreach (var otherKrill in krillPopulation.Population)
{
secondPart += Distance.Euclidean(krill.Coordinates, otherKrill.Coordinates);
}
double result = firstPart * secondPart;
return(Math.Round(result, 2));
}
private void PopulationInitialization()
{
// Population initialization
List <Krill> Krills = new List <Krill>();
for (int i = 1; i <= Parameters.Population; i++)
{
Krill krill = new Krill(i)
{
Coordinates = Vector <double> .Build.Random(Parameters.Dimensions, new ContinuousUniform(Parameters.LB, Parameters.UB))
};
Krills.Add(krill);
}
Population = new KrillPopulation(Krills);
}
/// <summary>
/// EQUATION 8
/// </summary>
private Vector <double> AlphaTarget(Krill krill, int currentIteration)
{
Krill bestKrill = KrillPopulation.GetBestKrill();
if (bestKrill.Fitness > krill.Fitness)
{
double K_i_best = K_i_j(krill.Fitness, bestKrill.Fitness);
Vector <double> X_i_best = X_i_j(krill.Coordinates, bestKrill.Coordinates);
double C_best = EffectiveCoefficient(currentIteration);
return(X_i_best.Multiply(C_best * K_i_best));
}
else
{
return(Vector <double> .Build.Dense(krill.Coordinates.Count));
}
}
/// <summary>
/// EQUATION 2
/// </summary>
public Vector <double> GetInducedMotion(Krill krill, int currentIteration)
{
int lastIteration = currentIteration - 1;
Vector <double> Alpha_i = MotionDirection(krill, currentIteration);
Vector <double> N_old = InducedSpeedHistory.ContainsKey(new Tuple <int, int>(lastIteration, krill.KrillNumber))
? InducedSpeedHistory[new Tuple <int, int>(lastIteration, krill.KrillNumber)]
: Vector <double> .Build.Dense(krill.Coordinates.Count);
double Omega_n = (0.1 + (0.8 * (1 - currentIteration / MaxIteration)));
Vector <double> N_new = N_max * Alpha_i + Omega_n * N_old;
// We add a krill to the history to know what the value of N_old is in later iterations
InducedSpeedHistory.Add(new Tuple <int, int>(currentIteration, krill.KrillNumber), N_new);
return(N_new);
}
/// <summary>
/// EQUATION 13
/// </summary>
private Vector <double> Beta_i_food(Krill krill, int currentIteration, Vector <double> vf_position)
{
Vector <double> result = Vector <double> .Build.Dense(krill.Coordinates.Count);
var K_food = FitnessFunction.Invoke(vf_position.ToArray());
if (K_food > krill.Fitness)
{
double C_food = EffectiveFoodCoefficient(currentIteration);
double K_i_food = K_i_j(krill.Fitness, K_food);
Vector <double> X_i_food = X_i_j(krill.Coordinates, vf_position);
result = X_i_food.Multiply(K_i_food * C_food);
}
return(result);
}
/// <summary>
/// EQUATION 4
/// </summary>
private Vector <double> AlphaLocal(Krill krill)
{
Vector <double> alphaLocal = Vector <double> .Build.Dense(krill.Coordinates.Count);
var neighbourhood = Neighbourhood.GetNeighbourhood(krill);
if (neighbourhood.Count == 0)
{
return(alphaLocal);
}
foreach (var neighbour in neighbourhood) // NN
{
var K_i_j_value = K_i_j(krill.Fitness, neighbour.Fitness);
alphaLocal = alphaLocal.Add(X_i_j(krill.Coordinates, neighbour.Coordinates).Multiply(K_i_j_value)); // i-th Krill, j-th Krill Neighbour
}
return(alphaLocal);
}
/// <summary>
/// Creates a neighborhood for a given krill
/// </summary>
public List <Krill> GetNeighbourhood(Krill krill)
{
List <Krill> neighbourhood = new List <Krill>();
double sensingDistance = SensingDistance(krill);
foreach (var neighbour in krillPopulation.Population)
{
if (neighbour.KrillNumber != krill.KrillNumber)
{
// calculate the distance between one individual and the other
double distance = Distance.Euclidean(krill.Coordinates, neighbour.Coordinates);
// If the distance between two krills is less than "sensingDistance", then we add this krill to the neighborhood
if (distance < sensingDistance)
{
neighbourhood.Add(neighbour);
}
}
}
return(neighbourhood);
}
