//3.1 DE/selbest/1 scheme
public void d_evolve_selbest_1(DecisionVector v, DecisionVector sbest, DecisionVector v1, DecisionVector v2, double F, double X)
{
for (int i = 0; i < v.Dimension; i++)
{
v.TrialVector[i] = sbest.CurrentVector[i] + F * (v1.CurrentVector[i] - v2.CurrentVector[i]);
}
}
//4.DE/best/2 scheme
public void d_evolve_best_2(DecisionVector v, DecisionVector vbest, DecisionVector v1, DecisionVector v2, DecisionVector v3, DecisionVector v4, double F)
{
for (int i = 0; i < v.Dimension; i++)
{
v.TrialVector[i] = vbest.CurrentVector[i] + F * (v1.CurrentVector[i] + v2.CurrentVector[i] - v3.CurrentVector[i] - v4.CurrentVector[i]);
}
}
//1. DE/rand/1 scheme
public void d_evole_r_1(DecisionVector v, DecisionVector v1, DecisionVector v2, DecisionVector v3, double F)
{
for (int i = 0; i < v.Dimension; i++)
{
v.TrialVector[i] = v1.CurrentVector[i] + F * (v2.CurrentVector[i] - v3.CurrentVector[i]);
}
}
//2.1 DE/current to selbest/1 scheme
public void d_evolve_rand_selbest_1(DecisionVector v, DecisionVector sbest, DecisionVector v2, DecisionVector v3, double F, double X)
{
for (int i = 0; i < v.Dimension; i++)
{
v.TrialVector[i] = v.CurrentVector[i] + X * (sbest.CurrentVector[i] - v.CurrentVector[i]) + F * (v2.CurrentVector[i] - v3.CurrentVector[i]);
}
}
//11. combine linear weight between randtolocalbest to rand/1
public void d_evolve_randtolocalbest_to_rand_1(DecisionVector v, DecisionVector lbest, DecisionVector v1, DecisionVector v2, DecisionVector v3, DecisionVector v4, DecisionVector v5, double F, double X, double W)
{
for (int i = 0; i < v.Dimension; i++)
{
v.TrialVector[i] = (1 - W) * (lbest.CurrentVector[i] + X * (lbest.CurrentVector[i] - v.CurrentVector[i]) + F * (v1.CurrentVector[i] - v2.CurrentVector[i])) + W * (v3.CurrentVector[i] + F * (v4.CurrentVector[i] - v5.CurrentVector[i]));
}
}
//12. combine linear weight between localbest to rand/1
public void d_evolve_rand1_to_localbest(DecisionVector v, DecisionVector lbest, DecisionVector v1, DecisionVector v2, DecisionVector v3, DecisionVector v4, DecisionVector v5, double F, double X, double W)
{
for (int i = 0; i < v.Dimension; i++)
{
v.TrialVector[i] = W * lbest.CurrentVector[i] + (1 - W) * v1.CurrentVector[i] + F * (v1.CurrentVector[i] - v2.CurrentVector[i]);
}
}
//9. combine linear weight between randtolocalBest and randtoBest
public void d_evolve_rand_localBest_TO_rand_to_Best(DecisionVector v, DecisionVector lbest, DecisionVector vbest, DecisionVector v1, DecisionVector v2, DecisionVector v3, DecisionVector v4, double X, double F, double W)
{
for (int i = 0; i < v.Dimension; i++)
{
v.TrialVector[i] = (1 - W) * (v.CurrentVector[i] + X * (lbest.CurrentVector[i] - v.CurrentVector[i]) + F * (v2.CurrentVector[i] - v3.CurrentVector[i])) + W * (v.CurrentVector[i] + X * (vbest.CurrentVector[i] - v.CurrentVector[i]) + F * (v2.CurrentVector[i] - v3.CurrentVector[i]));
}
}
//5.DE/rand/2 scheme
public void d_evolve_rand_2(DecisionVector v, DecisionVector v1, DecisionVector v2, DecisionVector v3, DecisionVector v4, DecisionVector v5, double F)
{
for (int i = 0; i < v.Dimension; i++)
{
v.TrialVector[i] = v1.CurrentVector[i] + F * (v2.CurrentVector[i] - v3.CurrentVector[i]) + F * (v4.CurrentVector[i] - v5.CurrentVector[i]);
}
}
public void randomtrialvector(DecisionVector v)
// diff way to generate new vector
{
Random rnd = new Random();
for (int i = 0; i < v.Dimension; i++)
{
v.TrialVector[i] = rnd.NextDouble();
}
}
public void Crossover_1point(DecisionVector v, double coRate, Random rnd)
{
for (int i = 0; i < v.Dimension; i++) // consider each dimension เลยนะ
{
if (rnd.NextDouble() >= coRate)
{
v.TrialVector[i] = v.CurrentVector[i];
}
}
}
public Population(int nVec, int nDim)
{ // construct a population with nVec vector,
// each vector with nDim dimension
Member = nVec;
Vector = new DecisionVector[Member]; // declare size of Vector (array)
lBestVectorIndex = new int[Member];
sBestVectorIndex = new int[Member];
for (int i = 0; i < Member; i++)
{ //Initialize vectors and each vector has nDim dimension
Vector[i] = new DecisionVector(nDim);
}
}
public void Crossover_2point(DecisionVector v, double coRate, Random rnd)
{
int start_p = rnd.Next(v.Dimension);
int L = 0;
do
{
L++;
}while ((rnd.NextDouble() <= coRate) && (start_p + L < v.Dimension));
for (int i = start_p; i < start_p + L; i++)
{
v.TrialVector[i] = v.CurrentVector[i];
}
}
public override double Objective(DecisionVector P, int trial)
{
double obj = 0;
//evaluate the objective value corresponding to the position of the Vector
if (trial == 0)
{
for (int i = 0; i < P.Dimension; i++)
{
obj += 0.001 * Math.Pow(P.CurrentVector[i], 2) + 2 * Math.Sin(P.CurrentVector[i]);
}
}
if (trial == 1)
{
for (int i = 0; i < P.Dimension; i++)
{
obj += 0.001 * Math.Pow(P.TrialVector[i], 2) + 2 * Math.Sin(P.TrialVector[i]);
}
}
return(obj);
}
public virtual void LocalSearchVector(DecisionVector p, ref Random rnd)
{
}
public virtual double Objective(DecisionVector p, int trial)
{ //empty function to be override in the problem specific definition
//to calculate objective function of a particle
return(0);
}
