private void doMutation (Solution x)
{
double rnd, delta1, delta2, mut_pow, deltaq;
double y, yl, yu, val, xy;
double eta_m = distributionIndex_;
for (int var = 0; var < x.numberOfVariables_; var++) {
if (PseudoRandom.Instance ().NextDouble () <= mutationProbability_) {
y = x.variable_[var].value_;
yl = x.variable_[var].lowerBound_;
yu = x.variable_[var].upperBound_;
delta1 = (y - yl) / (yu - yl);
delta2 = (yu - y) / (yu - yl);
rnd = PseudoRandom.Instance ().NextDouble ();
mut_pow = 1.0 / (eta_m + 1.0);
if (rnd <= 0.5) {
xy = 1.0 - delta1;
val = 2.0 * rnd + (1.0 - 2.0 * rnd) * (Math.Pow (xy, (eta_m + 1.0)));
deltaq = Math.Pow (val, mut_pow) - 1.0;
} else {
xy = 1.0 - delta2;
val = 2.0 * (1.0 - rnd) + 2.0 * (rnd - 0.5) * (Math.Pow (xy, (eta_m + 1.0)));
deltaq = 1.0 - (Math.Pow (val, mut_pow));
}
y = y + deltaq * (yu - yl);
if (y < yl)
y = yl;
if (y > yu)
y = yu;
x.variable_[var].value_ = y;
}
}
}
public override void evaluate(Solution solution)
{
// <pex>
if (solution == (Solution)null)
throw new ArgumentNullException("solution");
if (solution.variable_ == (Variable[])null)
throw new ArgumentNullException("solution");
// </pex>
Variable[] x = solution.variable_;
double[] fx = new double[numberOfObjectives_];
double sum1 = 0.0;
sum1 = Math.Sqrt(Math.Pow((x[0].value_ -2),2) + Math.Pow((x[1].value_ -3) ,2)) ;
fx[0] = sum1;
//for (int var = 0; var < numberOfVariables_; var++)
//{
// sum1 += Math.Pow(x[var].value_
// - (1.0 / Math.Sqrt((double)numberOfVariables_)), 2.0);
//}
//double exp1 = Math.Exp((-1.0) * sum1);
//fx[0] = 1 - exp1;
//double sum2 = 0.0;
//sum2 = (x[0].value_ + x[1].value_ + x[2].value_)*(-2);
//fx[1] = sum2;
//double sum2 = 0.0;
//for (int var = 0; var < numberOfVariables_; var++)
//{
// sum2 += Math.Pow(x[var].value_
// + (1.0 / Math.Sqrt((double)numberOfVariables_)), 2.0);
//}
//double exp2 = Math.Exp((-1.0) * sum2);
//fx[1] = 1 - exp2;
solution.objective_[0] = fx[0];
//solution.objective_[1] = fx[1];
}
/// <summary>
/// Generated new the children population list based on the parents population list.
/// This node needs to be used in a loop along with Assign Fitness Function node and
/// Sorting node.
/// </summary>
/// <param name="populationList">Parents population list.</param>
/// <param name="lowerLimits">List of lower limits for new generation.</param>
/// <param name="upperLimits">List of upper limits for new generation.</param>
/// <returns>Generated population list.</returns>
public static List<List<double>> GenerationAlgorithm(List<List<double>> populationList, List<double> lowerLimits, List<double> upperLimits)
{
int numVar = lowerLimits.Count;
int numObj = populationList.Count - numVar;
//create the NSGA-II algorithm
Algorithm algorithm = CreateAlgorithm(numObj, lowerLimits, upperLimits);
//change solutions list to solutionSet
SolutionSet population = SolutionListToPop(populationList, algorithm, numVar);
Operator mutationOperator = algorithm.operators["mutation"];
Operator crossoverOperator = algorithm.operators["crossover"];
Operator selectionOperator = algorithm.operators["selection"];
int populationSize = population.size();
SolutionSet offspringPopulation = new SolutionSet(populationSize);
Solution[] parents = new Solution[2];
for (int i = 0; i < (populationSize / 2); i++)
{
// selection
parents[0] = (Solution)selectionOperator.execute(population);
parents[1] = (Solution)selectionOperator.execute(population);
// crossover
Solution[] offSpring = (Solution[])crossoverOperator.execute(parents);
// mutation
mutationOperator.execute(offSpring[0]);
mutationOperator.execute(offSpring[1]);
offspringPopulation.@add(offSpring[0]);
offspringPopulation.add(offSpring[1]);
}
return PopToSolutionList(offspringPopulation);
}
/// <summary>
/// Creates the initial solution lists of variables and objectives.
/// The objectives will be assigned after evaluation using
/// NSGA_II Assigne Fitness Function Results node.
/// </summary>
/// <param name="populationSize">Size that will be used to generate population.</param>
/// <param name="numObjectives">Number of objectives that you are trying to optimize. If your problem is single objective optimization, set this number to one.</param>
/// <param name="lowerLimits">List of lower limits for all variable. The number of lower limits should be the same as the number of variables.</param>
/// <param name="upperLimits">List of upper limits for all variable. The number of upper limits should be the same as the number of variables.</param>
/// <returns>An initial solution list.</returns>
public static List<List<double>> InitialSolutionList(double populationSize, double numObjectives, List<double> lowerLimits, List<double> upperLimits)
{
//number of varables plus number of objectives
int numObj = Convert.ToInt32(numObjectives);
int numVar = lowerLimits.Count;
Algorithm algorithm = CreateAlgorithm(numObjectives, lowerLimits, upperLimits);
//Initializing variables
SolutionSet population = new SolutionSet(Convert.ToInt32(populationSize));
// Creating the initial solutionSet
Solution newSolution;
for (int i = 0; i < populationSize; i++)
{
newSolution = new Solution(algorithm.problem_);
population.add(newSolution);
}
List<List<double>> test = PopToSolutionList(population);
return PopToSolutionList(population);
}
//changes solution list to solution set
private static SolutionSet SolutionListToPop(List<List<double>> populationList, Algorithm algorithm, int numVar)
{
int popSize = populationList[0].Count;
int numObj = populationList.Count - numVar;
SolutionSet population = new SolutionSet(popSize);
Solution newSolution;
for (int j = 0; j < popSize; j++)
{
newSolution = new Solution(algorithm.problem_);
for (int i = 0; i < numVar; i++)
{
newSolution.variable_[i].value_ = populationList[i][j];
}
population.add(newSolution);
}
for (int j = 0; j < population.size(); j++)
{
for (int i = 0; i < numObj; i++)
{
population[j].objective_[i] = populationList[i + numVar][j];
}
}
return population;
}
public override object execute (object obj)
{
Object[] parameters = (Object[])obj;
Solution current = (Solution)parameters[0];
Solution[] parent = (Solution[])parameters[1];
//if ((Array.Find (validTypes, n => n == parent[0].type_.GetType ()) == null) ||
// (Array.Find (validTypes, n => n == parent[1].type_.GetType ()) == null) ||
// (Array.Find (validTypes, n => n == parent[2].type_.GetType ()) == null))
// //Console.WriteLine ("DiffentialEvolution. Types are incompatible");
//else
// ;
//Console.WriteLine ("Differential evolution. Types are Compatible");
int numberOfVariables = current.numberOfVariables_;
double jrand = PseudoRandom.Instance ().Next (numberOfVariables);
Solution child = new Solution (current);
if (DE_Variant_.CompareTo ("rand/1/bin") == 0) {
for (int j = 0; j < numberOfVariables; j++) {
if (PseudoRandom.Instance ().NextDouble () < CR_ || j == jrand) {
double val;
val = parent[2].variable_[j].value_ + F_ * (parent[0].variable_[j].value_ - parent[1].variable_[j].value_);
if (val < child.variable_[j].lowerBound_)
val = child.variable_[j].value_ = child.variable_[j].lowerBound_;
if (val > child.variable_[j].upperBound_)
val = child.variable_[j].value_ = child.variable_[j].upperBound_;
child.variable_[j].value_ = val;
} else {
child.variable_[j].value_ = current.variable_[j].value_ ;
} // else
} // for
} // if
return child;
}
// union
public void replace(int position, Solution solution) {
if (position > _solutionsList.Count) {
_solutionsList.Add(solution);
} // if
_solutionsList.RemoveAt(position);
_solutionsList.Insert(position,solution);
} // replace
// SolutionSet
/// <summary>
/// Adds a new solution into the solution set
/// </summary>
/// <param name="solution">
/// A <see cref="Solution"/>
/// </param>
/// <returns>
/// A <see cref="System.Boolean"/>
/// </returns>
public bool add (Solution solution)
{
if (_solutionsList.Count == capacity_) {
return false;
// if
} else {
_solutionsList.Add (solution);
return true;
}
// else
}
// NSGAII
public override SolutionSet execute()
{
int populationSize;
int maxEvaluations;
int evaluations;
SolutionSet population;
SolutionSet offspringPopulation;
SolutionSet union;
SolutionSet allTest = new SolutionSet();
Operator mutationOperator;
Operator crossoverOperator;
Operator selectionOperator;
populationSize = (int)inputParameters_["populationSize"];
maxEvaluations = (int)inputParameters_["maxEvaluations"];
// Initializing variables
population = new SolutionSet(populationSize);
evaluations = 0;
//System.Console.WriteLine("Solves Name:" + problem_.problemName_);
//System.Console.WriteLine("Pop size : " + populationSize);
//System.Console.WriteLine("Max Evals: " + maxEvaluations);
// Reading operators
mutationOperator = operators["mutation"];
crossoverOperator = operators["crossover"];
selectionOperator = operators["selection"];
//System.Console.WriteLine("Crossover parameters: " + crossoverOperator);
//System.Console.WriteLine("Mutation parameters: " + mutationOperator);
// Creating the initial solutionSet
Solution newSolution;
for (int i = 0; i < populationSize; i++)
{
newSolution = new Solution(problem_);
//newSolution.variable_[0] = 1;
problem_.evaluate(newSolution);
//newSolution.objective_[0] = 10;
//Mohammad
//slnLst.Add(newSolution);
//System.Console.WriteLine ("" + i + ": " + newSolution);
//problem_.evaluateConstraints(newSolution);
evaluations++;
population.add(newSolution);
}
//for
// Main loop
while (evaluations < maxEvaluations)
{
// Creating the offSpring solutionSet
offspringPopulation = new SolutionSet(populationSize);
Solution[] parents = new Solution[2];
for (int i = 0; i < (populationSize / 2); i++)
{
if (evaluations < maxEvaluations)
{
//if ((evaluations % 1000) == 0)
// Console.WriteLine("Evals: " + evaluations);
// selection
parents[0] = (Solution)selectionOperator.execute(population);
parents[1] = (Solution)selectionOperator.execute(population);
// crossover
Solution[] offSpring = (Solution[])crossoverOperator.execute(parents);
// mutation
mutationOperator.execute(offSpring[0]);
mutationOperator.execute(offSpring[1]);
//Environment.Exit(0);
// evaluation
problem_.evaluate(offSpring[0]);
problem_.evaluate(offSpring[1]);
offspringPopulation.@add(offSpring[0]);
offspringPopulation.add(offSpring[1]);
evaluations += 2;
}
// if
}
// for
// Creating the solutionSet union of solutionSet and offSpring
union = ((SolutionSet)population).union(offspringPopulation);
//Mohammad
allTest = ((SolutionSet)allTest).union(union);
//System.Console.WriteLine ("Union size:" + union.size ());
// Ranking the union
Ranking ranking = new Ranking(union);
int remain = populationSize;
int index = 0;
SolutionSet front = null;
//Distance distance = new Distance ();
population.clear();
// Obtain the next front
front = ranking.getSubfront(index);
//*
while ((remain > 0) && (remain >= front.size()))
{
//Assign crowding distance to individuals
Distance.crowdingDistanceAssignment(front, problem_.numberOfObjectives_);
//Add the individuals of this front
for (int k = 0; k < front.size(); k++)
{
population.@add(front[k]);
}
// for
//Decrement remain
remain = remain - front.size();
//Obtain the next front
index++;
if (remain > 0)
{
front = ranking.getSubfront(index);
}
// if
}
// while
// Remain is less than front(index).size, insert only the best one
if (remain > 0)
{
// front contains individuals to insert
Distance.crowdingDistanceAssignment(front, problem_.numberOfObjectives_);
IComparer comp = new CrowdingDistanceComparator();
front.solutionList_.Sort(comp.Compare);
for (int k = 0; k < remain; k++)
{
population.@add(front[k]);
}
// for
remain = 0;
}
// if
}
// while
// Return as output parameter the required evaluations
outputParameters_["evaluations"] = evaluations;
// Return the first non-dominated front
Ranking ranking2 = new Ranking(population);
//Console.WriteLine(slnLst[0].objective_.GetValue(0));
//File.WriteAllLines("C:/text.txt",slnLst.ConvertAll(Convert.ToString));
return ranking2.getSubfront(0);
}
Solution[] doCrossover (Solution parent1, Solution parent2)
{
Solution[] offSpring = new Solution[2];
offSpring[0] = new Solution (parent1);
offSpring[1] = new Solution (parent2);
int i;
double rand;
double y1, y2, yL, yu;
double c1, c2;
double alpha, beta, betaq;
double valueX1, valueX2;
Solution x1 = parent1;
Solution x2 = parent2;
Solution offs1 = offSpring[0];
Solution offs2 = offSpring[1];
int numberOfVariables = x1.numberOfVariables_;
if (PseudoRandom.Instance ().NextDouble () <= crossoverProbability_) {
for (i = 0; i < numberOfVariables; i++) {
valueX1 = x1.variable_[i].value_;
valueX2 = x2.variable_[i].value_;
if (PseudoRandom.Instance ().NextDouble () <= 0.5) {
if (Math.Abs (valueX1 - valueX2) > EPS) {
if (valueX1 < valueX2) {
y1 = valueX1;
y2 = valueX2;
} else {
y1 = valueX2;
y2 = valueX1;
}
// if
yL = x1.variable_[i].lowerBound_;
yu = x1.variable_[i].upperBound_;
rand = PseudoRandom.Instance ().NextDouble ();
beta = 1.0 + (2.0 * (y1 - yL) / (y2 - y1));
alpha = 2.0 - Math.Pow (beta, -(distributionIndex_ + 1.0));
if (rand <= (1.0 / alpha)) {
betaq = Math.Pow ((rand * alpha), (1.0 / (distributionIndex_ + 1.0)));
} else {
betaq = Math.Pow ((1.0 / (2.0 - rand * alpha)), (1.0 / (distributionIndex_ + 1.0)));
}
// if
c1 = 0.5 * ((y1 + y2) - betaq * (y2 - y1));
beta = 1.0 + (2.0 * (yu - y2) / (y2 - y1));
alpha = 2.0 - Math.Pow (beta, -(distributionIndex_ + 1.0));
if (rand <= (1.0 / alpha)) {
betaq = Math.Pow ((rand * alpha), (1.0 / (distributionIndex_ + 1.0)));
} else {
betaq = Math.Pow ((1.0 / (2.0 - rand * alpha)), (1.0 / (distributionIndex_ + 1.0)));
}
// if
c2 = 0.5 * ((y1 + y2) + betaq * (y2 - y1));
if (c1 < yL)
c1 = yL;
if (c2 < yL)
c2 = yL;
if (c1 > yu)
c1 = yu;
if (c2 > yu)
c2 = yu;
if (PseudoRandom.Instance ().NextDouble () <= 0.5) {
offs1.variable_[i].value_ = c2;
offs2.variable_[i].value_ = c1;
} else {
offs1.variable_[i].value_ = c1;
offs2.variable_[i].value_ = c2;
}
// if
} else {
offs1.variable_[i].value_ = valueX1;
offs2.variable_[i].value_ = valueX2;
}
// if
} else {
offs1.variable_[i].value_ = valueX2;
offs2.variable_[i].value_ = valueX1;
}
// if
}
// if
}
// if
return offSpring;
}
// Solution
/// <summary>
/// Copy constructor
/// </summary>
/// <param name="solution">
/// A <see cref="Solution"/>
/// </param>
public Solution (Solution solution)
{
// <pex>
if (solution == (Solution)null)
throw new ArgumentNullException("solution");
if (solution.type_ == (SolutionType)null)
throw new ArgumentNullException("solution");
// </pex>
problem_ = solution.problem_;
type_ = solution.type_;
numberOfObjectives_ = solution.numberOfObjectives_;
objective_ = new double[numberOfObjectives_];
for (int i = 0; i < numberOfObjectives_; i++) {
objective_[i] = solution.objective_[i];
}
// for
//<-
variable_ = type_.copyVariables (solution.variable_);
overallConstraintViolation_ = solution.overallConstraintViolation_;
numberOfViolatedConstraints_ = solution.numberOfViolatedConstraints_;
//distanceToSolutionSet_ = solution.getDistanceToSolutionSet();
crowdingDistance_ = solution.crowdingDistance_;
//kDistance_ = solution.getKDistance();
//fitness_ = solution.getFitness();
//marked_ = solution.isMarked ();
rank_ = solution.rank_;
//location_ = solution.getLocation();
}
public virtual void evaluateConstraints (Solution solution)
{
}
// Solves public abstract void evaluate (Solution solution);
