public void Initialize(ProblemDefinition problem, Randomization randomObj)
{
for (int i = 0; i < IndList.Count; i++)
{
IndList[i].Initialize(problem, randomObj);
}
}
public void Evaluate(ProblemDefinition problemObj)
{
for (int i = 0; i < this.IndList.Count; i++)
{
IndList[i].Evaluate(problemObj);
}
}
public void Initialize(ProblemDefinition problem, Randomization randomObj)
{
for (int i = 0; i < IndList.Count; i++)
{
IndList[i].Initialize(problem, randomObj);
}
}
public void Evaluate(ProblemDefinition problemObj)
{
for (int i = 0; i < this.IndList.Count; i++)
{
IndList[i].Evaluate(problemObj);
}
}
/* Function to print the information of a population in a file */
public void ReportPopulation(ProblemDefinition problemObj, string name, string info, int generation = 0)
{
Directory.CreateDirectory("report");
var file = File.OpenWrite($"report\\{problemObj.Title}_{name}_pop.out");
StreamWriter writer = new StreamWriter(file);
writer.WriteLine($"# {info}");
writer.WriteLine($"# of objectives = {problemObj.ObjectiveCount}, # of constraints = {problemObj.ConstraintCount}, # of real_var = {problemObj.RealVariableCount}, # of bits of bin_var = {problemObj.TotalBinaryBitLength}, constr_violation, rank, crowding_distance");
if (generation > 0)
{
writer.WriteLine($"# Generation no: {generation}/{problemObj.MaxGeneration}");
}
for (int i = 0; i < problemObj.PopulationSize; i++)
{
for (int j = 0; j < problemObj.ObjectiveCount; j++)
{
writer.Write($"{IndList[i].Obj[j].ToString(CultureInfo.InvariantCulture)}\t");
}
if (problemObj.ConstraintCount != 0)
{
for (int j = 0; j < problemObj.ConstraintCount; j++)
{
writer.Write($"{IndList[i].Constr[j].ToString("E")}\t");
}
}
if (problemObj.RealVariableCount != 0)
{
for (int j = 0; j < problemObj.RealVariableCount; j++)
{
writer.Write($"{IndList[i].Xreal[j].ToString("E")}\t");
}
}
if (problemObj.BinaryVariableCount != 0)
{
for (int j = 0; j < problemObj.BinaryVariableCount; j++)
{
for (int k = 0; k < problemObj.Nbits[j]; k++)
{
writer.Write($"{IndList[i].Gene[j][ k]}\t");
}
}
}
writer.Write($"{IndList[i].ConstrViolation.ToString("E")}\t");
writer.Write($"{IndList[i].Rank}\t");
writer.Write($"{IndList[i].CrowdDist.ToString("E")}\n");
}
writer.Flush();
writer.Close();
}
public double HillClimb(ProblemDefinition problemObj)
{
double retVal = 0;
for (int i = 0; i < IndList.Count; i++)
{
retVal += IndList[i].HillClimb(problemObj);
}
return retVal;
}
public void Decode(ProblemDefinition problem)
{
if (problem.BinaryVariableCount == 0)
return;
for (int i = 0; i < IndList.Count; i++)
{
IndList[i].Decode(problem);
}
}
public double HillClimb(ProblemDefinition problemObj)
{
double retVal = 0;
for (int i = 0; i < IndList.Count; i++)
{
retVal += IndList[i].HillClimb(problemObj);
}
return(retVal);
}
public void Decode(ProblemDefinition problem)
{
if (problem.BinaryVariableCount == 0)
{
return;
}
for (int i = 0; i < IndList.Count; i++)
{
IndList[i].Decode(problem);
}
}
public void Merge( Population pop1, Population pop2, ProblemDefinition problem)
{
IndList.Clear();
for (int i = 0; i < pop1.IndList.Count; i++)
{
IndList.Add( new Individual(pop1.IndList[i], problem));
}
for (int i = 0; i < pop2.IndList.Count; i++)
{
IndList.Add(new Individual(pop2.IndList[i], problem));
}
}
public void Merge(Population pop1, Population pop2, ProblemDefinition problem)
{
IndList.Clear();
for (int i = 0; i < pop1.IndList.Count; i++)
{
IndList.Add(new Individual(pop1.IndList[i], problem));
}
for (int i = 0; i < pop2.IndList.Count; i++)
{
IndList.Add(new Individual(pop2.IndList[i], problem));
}
}
public double HillClimb(ProblemDefinition problemObj)
{
double tResultImprovement = 0;
if (CollisionList.Count > 0)
{
bool continueClimb;
Individual original = new Individual(this, problemObj);
original.Decode(problemObj);
original.Evaluate(problemObj);
int seeder = 7;
Random rnd = new Random(seeder);
int climbRetryCount = 0;
do
{
climbRetryCount++;
bool climb = HillClimber(problemObj, rnd);
if (climb)
return 0;
//HillClimberF22(problemObj, rnd,22);
Decode(problemObj);
Evaluate(problemObj);
var obj0Imp = original.Obj[0] - Obj[0];
var obj1Imp = original.Obj[1] - Obj[1];
var obj2Imp = original.Obj[2] - Obj[2];
if (obj0Imp + obj1Imp + obj2Imp == 0)
{
if (obj0Imp > 0)
{
original = new Individual(this, problemObj);
original.Decode(problemObj);
original.Evaluate(problemObj);
}
continueClimb = false;
}
else if (obj0Imp + obj1Imp + obj2Imp > 0)
{
continueClimb = false;
original = new Individual(this, problemObj);
original.Decode(problemObj);
original.Evaluate(problemObj);
}
else
{
continueClimb = false;
}
tResultImprovement += obj0Imp + obj1Imp + obj2Imp;
} while (continueClimb && climbRetryCount < 10);
if (tResultImprovement < 0) //todo check sanki geri almıyor.
{
Copy(original, problemObj);
Decode(problemObj);
Evaluate(problemObj);
tResultImprovement = 0;
}
}
return tResultImprovement;
}
public void Evaluate(ProblemDefinition problemObj)
{
EvaluateProblem(problemObj);
#region constraint kısmı yok
if (problemObj.ConstraintCount == 0)
{
ConstrViolation = 0.0;
}
else
{
ConstrViolation = 0.0;
for (int j = 0; j < problemObj.ConstraintCount; j++)
{
if (Constr[j] < 0.0)
{
ConstrViolation += Constr[j];
}
}
}
#endregion
}
/* Function to perform mutation of an individual */
static void MutateIndividual(Individual ind, ProblemDefinition problemObj, Randomization randomizationObj)
{
if (problemObj.RealVariableCount != 0)
{
RealMutate(ind, problemObj, randomizationObj);
}
if (problemObj.BinaryVariableCount != 0)
{
BinaryMutate(ind, problemObj, randomizationObj);
}
}
/* Routine to compute crowding distance based on objective function values when the population in in the form of an array */
static void assign_crowding_distance_indices(Population pop, int c1, int c2, ProblemDefinition problemObj, Randomization randomizationObj)
{
int[][] objArray;
int[] dist;
int i, j;
int frontSize;
frontSize = c2 - c1 + 1;
if (frontSize == 1)
{
pop.IndList[c1].CrowdDist = problemObj.Inf;
return;
}
if (frontSize == 2)
{
pop.IndList[c1].CrowdDist = problemObj.Inf;
pop.IndList[c2].CrowdDist = problemObj.Inf;
return;
}
dist = new int[frontSize];
objArray = new int[problemObj.ObjectiveCount][];
//obj_array = (int**)malloc(ProblemObj.ObjectiveCount * sizeof(int*));
for (i = 0; i < problemObj.ObjectiveCount; i++)
{
objArray[i] = new int[frontSize];
}
for (j = 0; j < frontSize; j++)
{
dist[j] = c1++;
}
assign_crowding_distance(pop, dist, objArray, frontSize, problemObj, randomizationObj);
}
/* Routine to compute crowding distance based on ojbective function values when the population in in the form of a list */
static void assign_crowding_distance_list(Population pop, Lists lst, int frontSize, ProblemDefinition problemObj, Randomization randomizationObj)
{
int[][] objArray;
int[] dist;
int i, j;
Lists temp;
temp = lst;
if (frontSize == 1)
{
pop.IndList[lst.index].CrowdDist = problemObj.Inf;
return;
}
if (frontSize == 2)
{
pop.IndList[lst.index].CrowdDist = problemObj.Inf;
pop.IndList[lst.child.index].CrowdDist = problemObj.Inf;
return;
}
dist = new int[frontSize];
objArray = new int[problemObj.ObjectiveCount][];
//obj_array = (int**)malloc(ProblemObj.ObjectiveCount * sizeof(int*));
for (i = 0; i < problemObj.ObjectiveCount; i++)
{
objArray[i] = new int[frontSize];
}
for (j = 0; j < frontSize; j++)
{
dist[j] = temp.index;
temp = temp.child;
}
assign_crowding_distance(pop, dist, objArray, frontSize, problemObj, randomizationObj);
}
///* Function to display the current population for the subsequent generation */
public void PlotPopulation(Population pop, ProblemDefinition problem, int genNo = 0, List <Individual> best = null)
{
if (!Use3D) //2D
{
var xr = new double[problem.PopulationSize];
var yr = new double[problem.PopulationSize];
for (int x = 0; x < problem.PopulationSize; x++)
{
xr[x] = pop.IndList[x].Obj[GnuplotObjective1];
yr[x] = pop.IndList[x].Obj[GnuplotObjective2];
}
if (best != null)
{
GnuPlot.HoldOn();
}
GnuPlot.Plot(xr, yr, $"title 'Generation #{genNo} of {problem.MaxGeneration}' with points pointtype 8 lc rgb 'blue'");
if (best != null)
{
var x = new double[best.Count];
var y = new double[best.Count];
for (int i = 0; i < best.Count; i++)
{
x[i] = best[i].Obj[GnuplotObjective1];
y[i] = best[i].Obj[GnuplotObjective2];
}
GnuPlot.Plot(x, y, $"title 'best ({best.First().TotalResult})' with points pointtype 6 lc rgb 'red'");
}
GnuPlot.Set($"xlabel \"obj[{GnuplotObjective1 }]\"");
GnuPlot.Set($"ylabel \"obj[{GnuplotObjective2 }]\"");
}
else //3D
{
var rank1Count = pop.IndList.Count(x => x.Rank == 1);
var rankOtherCount = pop.IndList.Count(x => x.Rank != 1);
if (rank1Count > 0)
{
var xr = new double[rank1Count];
var yr = new double[rank1Count];
var zr = new double[rank1Count];
int index = 0;
foreach (var ind in pop.IndList.Where(x => x.Rank == 1))
{
xr[index] = ind.Obj[GnuplotObjective1];
yr[index] = ind.Obj[GnuplotObjective2];
zr[index] = ind.Obj[GnuplotObjective3];
index++;
}
if (best != null || rankOtherCount > 0)
{
GnuPlot.HoldOn();
}
GnuPlot.SPlot(xr, yr, zr, $"title 'Rank 1 individuals' with points pointtype 8 lc rgb 'red'");
}
if (rankOtherCount > 0)
{
var xr = new double[rankOtherCount];
var yr = new double[rankOtherCount];
var zr = new double[rankOtherCount];
int index = 0;
foreach (var ind in pop.IndList.Where(x => x.Rank != 1))
{
xr[index] = ind.Obj[GnuplotObjective1];
yr[index] = ind.Obj[GnuplotObjective2];
zr[index] = ind.Obj[GnuplotObjective3];
index++;
}
GnuPlot.SPlot(xr, yr, zr, $"title 'Rank 2-{pop.IndList.Max(x=>x.Rank)} individuals' with points pointtype 8 lc rgb 'blue'");
}
if (best != null)
{
var x = new double[best.Count];
var y = new double[best.Count];
var z = new double[best.Count];
for (int i = 0; i < best.Count; i++)
{
x[i] = best[i].Obj[GnuplotObjective1];
y[i] = best[i].Obj[GnuplotObjective2];
z[i] = best[i].Obj[GnuplotObjective3];
}
GnuPlot.SPlot(x, y, z, $"title 'best ({best.First().TotalResult})' with points pointtype 6 lc rgb 'green'");
}
GnuPlot.Set($"title 'Generation #{genNo} of {problem.MaxGeneration}'");
GnuPlot.Set($"xlabel \"obj[{GnuplotObjective1}]\"");
GnuPlot.Set($"ylabel \"obj[{GnuplotObjective2}]\"");
GnuPlot.Set($"zlabel \"obj[{GnuplotObjective3}]\"");
}
}
private void ChangeGene(int geneId, int value, ProblemDefinition problem)
{
double minbin = problem.MinBinvar[geneId];
double maxbin = problem.MaxBinvar[geneId];
double nbit = problem.Nbits[geneId];
//int valueToAdd = (int)((value - minbin) * ((Math.Pow(2, nbit) - 1) / (maxbin - minbin)));
double sum = (double)((value - minbin) / ((maxbin - minbin) / (Math.Pow(2, nbit) - 1)));
int valueToAdd = (int)sum;
if (sum > (int)sum)
valueToAdd++;
int bits = problem.Nbits[geneId];
for (int k = bits - 1; k >= 0; k--)
{
int divident = (int)Math.Pow(2, bits - 1 - k);
int modulus = (int)Math.Pow(2, bits - k);
if (valueToAdd % modulus == divident)
{
valueToAdd = valueToAdd - divident;
Gene[geneId][k] = 1;
}
else
{
Gene[geneId][k] = 0;
}
}
}
/* Routine for two point binary Crossover */
static void BinaryCrossover(Individual parent1, Individual parent2, Individual child1, Individual child2, ProblemDefinition problemObj, Randomization randomizationObj)
{
int i, j;
double rand;
int temp, site1, site2;
for (i = 0; i < problemObj.BinaryVariableCount; i++)
{
rand = randomizationObj.RandomPercent();
if (rand <= problemObj.BinaryCrossoverProbability)
{
problemObj.BinaryCrossoverCount++;
site1 = randomizationObj.RandomInteger(0, problemObj.Nbits[i] - 1);
site2 = randomizationObj.RandomInteger(0, problemObj.Nbits[i] - 1);
if (site1 > site2)
{
temp = site1;
site1 = site2;
site2 = temp;
}
for (j = 0; j < site1; j++)
{
child1.Gene[i][j] = parent1.Gene[i][j];
child2.Gene[i][j] = parent2.Gene[i][j];
}
for (j = site1; j < site2; j++)
{
child1.Gene[i][j] = parent2.Gene[i][j];
child2.Gene[i][j] = parent1.Gene[i][j];
}
for (j = site2; j < problemObj.Nbits[i]; j++)
{
child1.Gene[i][j] = parent1.Gene[i][j];
child2.Gene[i][j] = parent2.Gene[i][j];
}
}
else
{
for (j = 0; j < problemObj.Nbits[i]; j++)
{
child1.Gene[i][j] = parent1.Gene[i][j];
child2.Gene[i][j] = parent2.Gene[i][j];
}
}
}
}
/* Routine for binary mutation of an individual */
static void BinaryMutate(Individual ind, ProblemDefinition problemObj, Randomization randomizationObj)
{
int j, k;
double prob;
for (j = 0; j < problemObj.BinaryVariableCount; j++)
{
for (k = 0; k < problemObj.Nbits[j]; k++)
{
prob = randomizationObj.RandomPercent();
if (prob <= problemObj.BinaryMutationProbability)
{
if (ind.Gene[j][k] == 0)
{
ind.Gene[j][k] = 1;
}
else
{
ind.Gene[j][k] = 0;
}
problemObj.BinaryMutationCount += 1;
}
}
}
}
/* Routine for usual non-domination checking
It will return the following values
1 if a dominates b
-1 if b dominates a
0 if both a and b are non-dominated */
static int CheckDominance(Individual a, Individual b, ProblemDefinition problemObj)
{
int i;
int flag1;
int flag2;
flag1 = 0;
flag2 = 0;
if (a.ConstrViolation < 0 && b.ConstrViolation < 0)
{
if (a.ConstrViolation > b.ConstrViolation)
{
return 1;
}
if (a.ConstrViolation < b.ConstrViolation)
{
return -1;
}
return 0;
}
if (a.ConstrViolation < 0 && b.ConstrViolation == 0)
{
return -1;
}
if (a.ConstrViolation == 0 && b.ConstrViolation < 0)
{
return 1;
}
for (i = 0; i < problemObj.ObjectiveCount; i++)
{
if (a.Obj[i] < b.Obj[i])
{
flag1 = 1;
}
else
{
if (a.Obj[i] > b.Obj[i])
{
flag2 = 1;
}
}
}
if (flag1 == 1 && flag2 == 0)
{
return 1;
}
if (flag1 == 0 && flag2 == 1)
{
return -1;
}
return 0;
}
/* Function to cross two individuals */
static void Crossover(Individual parent1, Individual parent2, Individual child1, Individual child2, ProblemDefinition problemObj, Randomization randomizationObj)
{
if (problemObj.RealVariableCount != 0)
{
RealCrossover(parent1, parent2, child1, child2, problemObj, randomizationObj);
}
if (problemObj.BinaryVariableCount != 0)
{
BinaryCrossover(parent1, parent2, child1, child2, problemObj, randomizationObj);
}
}
/* Routine to fill a population with individuals in the decreasing order of crowding distance */
static void crowding_fill(Population mixedPop, Population newPop, int count, int frontSize, Lists elite, ProblemDefinition ProblemObj, Randomization RandomizationObj)
{
int[] dist;
Lists temp;
int i, j;
assign_crowding_distance_list(mixedPop, elite.child, frontSize, ProblemObj, RandomizationObj);
dist = new int[frontSize];
temp = elite.child;
for (j = 0; j < frontSize; j++)
{
dist[j] = temp.index;
temp = temp.child;
}
quicksort_dist(mixedPop, dist, frontSize, RandomizationObj);
for (i = count, j = frontSize - 1; i < ProblemObj.PopulationSize; i++, j--)
{
newPop.IndList[i].Copy(mixedPop.IndList[dist[j]], ProblemObj);
//CopyIndividual(mixedPop.IndList[dist[j]], newPop.IndList[i]);
//newPop.IndList[i] = new Individual(mixedPop.IndList[dist[j]], ProblemObj);
}
}
/* Function to perform mutation in a population */
static void MutatePopulation(Population pop, ProblemDefinition problemObj, Randomization randomizationObj)
{
int i;
for (i = 0; i < problemObj.PopulationSize; i++)
{
MutateIndividual(pop.IndList[i], problemObj, randomizationObj);
}
}
/* Function to initialize an individual randomly */
public void Initialize(ProblemDefinition problem, Randomization randomObj)
{
int j;
if (problem.RealVariableCount != 0)
{
for (j = 0; j < problem.RealVariableCount; j++)
{
Xreal[j] = randomObj.RandomDouble(problem.MinRealvar[j], problem.MaxRealvar[j]);
}
}
if (problem.BinaryVariableCount != 0)
{
for (j = 0; j < problem.BinaryVariableCount; j++)
{
for (int k = 0; k < problem.Nbits[j]; k++)
{
if (randomObj.RandomPercent() <= 0.5)
{
Gene[j][k] = 0;
}
else
{
Gene[j][k] = 1;
}
}
}
}
}
public Individual(Individual ind, ProblemDefinition problem)
{
CollisionList = new List<Collision>();
FacultySections = new List<FacultySection>();
DiffSemesterFacultySections = new List<FacultySection>();
ElectiveFacultySections = new List<FacultySection>();
TotalResult = 0;
NRealVar = ind.NRealVar;
NBinVar = ind.NBinVar;
NMaxBit = ind.NMaxBit;
NObj = ind.NObj;
NCons = ind.NCons;
if (ind.NRealVar != 0)
Xreal = new double[NRealVar];
if (ind.NBinVar != 0)
{
SlotId = new int[ind.NBinVar];
Gene = new List<List<int>>();
for (int i = 0; i < ind.NBinVar; i++)
{
Gene.Add(new List<int>(ind.NBinVar));
for (int j = 0; j < NMaxBit; j++)
{
Gene[i].Add(0);
}
}
//Gene = new int[ind._nBinVar, ind._nMaxBit];
}
Obj = new double[ind.NObj];
if (ind.NCons != 0)
Constr = new double[ind.NCons];
Rank = ind.Rank;
ConstrViolation = ind.ConstrViolation;
CrowdDist = ind.CrowdDist;
if (ind.NRealVar > 0)
{
for (int i = 0; i < ind.NRealVar; i++)
{
Xreal[i] = ind.Xreal[i];
}
}
if (ind.NBinVar > 0)
{
for (int i = 0; i < ind.NBinVar; i++)
{
SlotId[i] = ind.SlotId[i];
for (int j = 0; j < problem.Nbits[i]; j++)
{
Gene[i][j] = ind.Gene[i][j];
}
}
}
for (int i = 0; i < ind.NObj; i++)
{
Obj[i] = ind.Obj[i];
}
if (ind.NCons > 0)
{
for (int i = 0; i < ind.NCons; i++)
{
Constr[i] = ind.Constr[i];
}
}
Timetable = new List<List<Slot>>();
}
/* Routine to compute crowding distances */
static void assign_crowding_distance(Population pop, int[] dist, int[][] objArray, int frontSize, ProblemDefinition problemObj, Randomization randomizationObj)
{
int i, j;
for (i = 0; i < problemObj.ObjectiveCount; i++)
{
for (j = 0; j < frontSize; j++)
{
objArray[i][j] = dist[j];
}
quicksort_front_obj(pop, i, objArray[i], frontSize, randomizationObj);
}
for (j = 0; j < frontSize; j++)
{
pop.IndList[dist[j]].CrowdDist = 0.0;
}
for (i = 0; i < problemObj.ObjectiveCount; i++)
{
pop.IndList[objArray[i][0]].CrowdDist = problemObj.Inf;
}
for (i = 0; i < problemObj.ObjectiveCount; i++)
{
for (j = 1; j < frontSize - 1; j++)
{
if (pop.IndList[objArray[i][j]].CrowdDist != problemObj.Inf)
{
if (pop.IndList[objArray[i][frontSize - 1]].Obj[i] == pop.IndList[objArray[i][0]].Obj[i])
{
pop.IndList[objArray[i][j]].CrowdDist += 0.0;
}
else
{
pop.IndList[objArray[i][j]].CrowdDist += (pop.IndList[objArray[i][j + 1]].Obj[i] - pop.IndList[objArray[i][j - 1]].Obj[i]) / (pop.IndList[objArray[i][frontSize - 1]].Obj[i] - pop.IndList[objArray[i][0]].Obj[i]);
}
}
}
}
for (j = 0; j < frontSize; j++)
{
if (pop.IndList[dist[j]].CrowdDist != problemObj.Inf)
{
pop.IndList[dist[j]].CrowdDist = pop.IndList[dist[j]].CrowdDist / problemObj.ObjectiveCount;
}
}
}
private int DecodeGene(ProblemDefinition problem, int j)
{
var sum = 0;
for (int k = 0; k < problem.Nbits[j]; k++)
{
if (Gene[j][k] == 1)
{
sum += (int)Math.Pow(2, problem.Nbits[j] - 1 - k);
}
}
double minbin = problem.MinBinvar[j];
double maxbin = problem.MaxBinvar[j];
double nbit = problem.Nbits[j];
return (int)(minbin + sum * (maxbin - minbin) / (Math.Pow(2, nbit) - 1));
}
private void EvaluateProblem(ProblemDefinition problemObj)
{
TotalResult = 0;
CollisionList.Clear();
#region faculty section lists
FacultySections.Clear();
foreach (var course in problemObj.FacultyCourseList.OrderBy(x => x.Code))
{
if (!FacultySections.Any(x => x.Code == course.Code && x.Section == course.Section))
{
var temp = new FacultySection
{
Code = course.Code,
Section = course.Section
};
FacultySections.Add(temp);
}
}
DiffSemesterFacultySections.Clear();
foreach (var course in problemObj.FacultyCourseList.OrderBy(x => x.Code))
{
if (!DiffSemesterFacultySections.Any(x => x.Code == course.Code && x.Section == course.Section))
{
var temp = new FacultySection
{
Code = course.Code,
Section = course.Section
};
DiffSemesterFacultySections.Add(temp);
}
}
ElectiveFacultySections.Clear();
foreach (var course in problemObj.FacultyCourseList.OrderBy(x => x.Code))
{
if (!ElectiveFacultySections.Any(x => x.Code == course.Code && x.Section == course.Section))
{
var temp = new FacultySection
{
Code = course.Code,
Section = course.Section
};
ElectiveFacultySections.Add(temp);
}
}
#endregion
#region fill variables
Slot[] easy = new Slot[50];
Timetable.Clear();
int idcounter = 1;
for (int x = 0; x < 5; x++)
{
Timetable.Add(new List<Slot>());
for (int y = 0; y < 10; y++)
{
Timetable[x].Add(new Slot(problemObj.TeacherList.Count, idcounter));
easy[idcounter - 1] = Timetable[x][y];
idcounter++;
}
}
Obj[0] = 0;
Obj[1] = 0;
Obj[2] = 0;
for (int j = 0; j < problemObj.CourseList.Count; j++) //ders sayisi kadar,
{
int slotId = SlotId[j];
AddToTimetable(slotId, problemObj.CourseList[j]); //dersleri timetable'a yerleştirdim.
}
foreach (Course facultyCourse in problemObj.FacultyCourseList)
{
for (int i = 0; i < facultyCourse.Duration; i++)
{
easy[facultyCourse.SlotId + i - 1].facultyCourses.Add(facultyCourse);
}
}
for (int x = 0; x < 5; x++)
{
for (int y = 0; y < 9; y++)
{
if (problemObj.Meeting[x][y] > 0)
Timetable[x][y].meetingHour = true; // meeting hours
if (problemObj.LabScheduling[x][y] > 0)
Timetable[x][y].facultyLab = problemObj.LabScheduling[x][y];
}
}
#endregion
#region calc. collisions
#region Base vs Faculty same semester
//dönem ici dekanlik/bolum dersi cakismasi todo: scheduling'in normal slot halinde gelmesi lazım
{
List<Collision> col = CollisionBaseVsFaculty(0);
foreach (var collision in col)
{
double changeResult = 0;
var crashingfacultycourses = collision.CrashingCourses.Where(x => x.FacultyCourse);
foreach (var course in crashingfacultycourses)
{
if (FacultySections.Any(x => x.Code == course.Code && x.Crashing == false))
{
collision.Reason += $"; {course.Code} has noncrashing section";
changeResult += 0.01;
}
else
{
changeResult++;
}
}
collision.Result = changeResult;
}
var result = col.Sum(item => item.Result);
Obj[0] += result;
CollisionList.AddRange(col);
}
#endregion
#region Base vs Base same semester
//donem ici bolum dersi cakismasi
for (int j = 1; j < 9; j++)
{
List<Collision> col = CollisionBaseVsBase(Timetable, 1, j);
var result = col.Sum(item => item.Result);
Obj[0] += result;
CollisionList.AddRange(col);
}
#endregion
#region Base vs Faculty -1 +1 semester.
//dönemler arasi dekanlik/bolum dersi cakismasi todo: scheduling'in normal slot halinde gelmesi lazım
{
List<Collision> col = new List<Collision>();
for (int j = 1; j < 8; j++)
{
// 1-2 2-3 3-4 4-5 5-6 6-7 7-8
// 2-1 3-2 4-3 5-4 6-5 7-6 8-7 consecutive CSE&faculty courses
col.AddRange(CollisionBaseVsFacultyDiffSemester(0, j, j + 1, 1));
col.AddRange(CollisionBaseVsFacultyDiffSemester(0, j + 1, j, 1));
}
foreach (var collision in col)
{
double changeResult = 0;
var crashingfacultycourses = collision.CrashingCourses.Where(x => x.FacultyCourse);
foreach (var course in crashingfacultycourses)
{
if (DiffSemesterFacultySections.Any(x => x.Code == course.Code && x.Crashing == false))
{
collision.Reason += $"; {course.Code} has noncrashing section";
changeResult += 0.01;
}
else
{
changeResult++;
}
}
collision.Result = changeResult;
}
var result = col.Sum(item => item.Result);
Obj[1] += result;
CollisionList.AddRange(col);
}
#endregion
#region Base vs Base -1 +1 semester
//dönemler arası CSE çakışmaları
for (int j = 1; j < 8; j++)
{
List<Collision> col = CollisionBaseVsBaseDiffSemester(Timetable, 1, new List<int> { j, j + 1 }, 1);
var y = col.Sum(item => item.Result);
Obj[1] += y;
CollisionList.AddRange(col);
}
#endregion
#region Lab Count
//aynı saatte 4'ten fazla lab olmaması lazim
{
List<Collision> labcol = CollisionInLabs(Timetable, 4);
var y = labcol.Sum(item => item.Result);
Obj[0] += y;
CollisionList.AddRange(labcol);
}
//# of lab at most 4
#endregion
#region Teacher Collisions
for (int j = 0; j < problemObj.TeacherList.Count; j++) //Bütün hocalar için
{
if (problemObj.TeacherList[j].Equals("ASSISTANT"))
{
continue;
}
{
//og. gor. aynı saatte baska dersinin olmaması
List<Collision> col = CollisionTeacher(Timetable, j, 1);
var yy1 = col.Sum(item => item.Result);
Obj[0] += yy1;
CollisionList.AddRange(col);
}
{
//og. gor. gunluk 4 saatten fazla pespese dersinin olmamasi
List<Collision> col = CollisionTeacherConsicutive(Timetable, j, 4);
var y = col.Sum(item => item.Result);
Obj[2] += y;
CollisionList.AddRange(col);
//teacher have at most 4 consective lesson per day
}
{
//og. gor. boş gununun olması
List<Collision> col = CollisionTeacherFreeDay(Timetable, j);
var y = col.Sum(item => item.Result);
Obj[2] += y;
CollisionList.AddRange(col);
}
/* teacher have free day*/
}
#endregion
#region Lab vs Lecture different day collision
//lab ve lecture farklı günlerde olsun
{
List<Collision> col = CollisionLabLectureSameDay(Timetable);
var y = col.Sum(item => item.Result);
Obj[2] += y;
CollisionList.AddRange(col);
}
#endregion
#region Elecive vs Elective
//elective vs elective collision
{
List<Collision> col = CollisionElectiveVsElective(Timetable, 1);
var y = col.Sum(item => item.Result);
Obj[0] += y;
CollisionList.AddRange(col);
}
#endregion
#region Elective vs Faculty in semester 6 7 8
//elective vs faculty courses in semester
{
List<Collision> col = new List<Collision>();
col.AddRange(CollisionElectiveVsFacultyDiffSemester(6, 0));
col.AddRange(CollisionElectiveVsFacultyDiffSemester(7, 0));
col.AddRange(CollisionElectiveVsFacultyDiffSemester(8, 0));
foreach (var collision in col)
{
double changeResult = 0;
var crashingfacultycourses = collision.CrashingCourses.Where(x => x.FacultyCourse);
foreach (var course in crashingfacultycourses)
{
if (ElectiveFacultySections.Any(x => x.Code == course.Code && x.Crashing == false))
{
collision.Reason += $"; {course.Code} has noncrashing section";
changeResult += 0.01;
}
else
{
changeResult++;
}
}
collision.Result = changeResult;
}
var y = col.Sum(item => item.Result);
Obj[2] += y;
CollisionList.AddRange(col);
}
#endregion
#region Elective vs Base in semester 6 7 8
//elective vs base courses in semester
if (!problemObj.DisabledCollisions.HasFlag(UCTProblem.DisabledCollisions.ElectiveVsBase))
{
List<Collision> col = CollisionElectiveVsBaseDiffSemester(0, 6, 1);
var y = col.Sum(item => item.Result);
Obj[1] += y;
CollisionList.AddRange(col);
col.Clear();
col = CollisionElectiveVsBaseDiffSemester(0, 7);
y = col.Sum(item => item.Result);
Obj[0] += y;
CollisionList.AddRange(col);
col.Clear();
col = CollisionElectiveVsBaseDiffSemester(0, 8);
y = col.Sum(item => item.Result);
Obj[0] += y;
CollisionList.AddRange(col);
}
#endregion
#endregion
TotalResult = Obj[0] + Obj[1] + Obj[2];
}
/* Function to assign rank and crowding distance to a population of size pop_size*/
static void assign_rank_and_crowding_distance(Population newPopulation, ProblemDefinition problemObj, Randomization randomizationObj)
{
int flag;
int i;
int end;
int frontSize;
int rank = 1;
Lists orig;
Lists cur;
Lists temp1, temp2;
orig = new Lists();
cur = new Lists();
orig.index = -1;
orig.parent = null;
orig.child = null;
cur.index = -1;
cur.parent = null;
cur.child = null;
temp1 = orig;
for (i = 0; i < problemObj.PopulationSize; i++)
{
Insert(temp1, i);
temp1 = temp1.child;
}
do
{
if (orig.child != null && orig.child.child == null)
{
newPopulation.IndList[orig.child.index].Rank = rank;
newPopulation.IndList[orig.child.index].CrowdDist = problemObj.Inf;
break;
}
temp1 = orig.child;
Insert(cur, temp1.index);
frontSize = 1;
temp2 = cur.child;
temp1 = Delete(temp1);
temp1 = temp1.child;
do
{
temp2 = cur.child;
do
{
end = 0;
flag = CheckDominance(newPopulation.IndList[temp1.index], newPopulation.IndList[temp2.index], problemObj);
if (flag == 1)
{
Insert(orig, temp2.index);
temp2 = Delete(temp2);
frontSize--;
temp2 = temp2.child;
}
if (flag == 0)
{
temp2 = temp2.child;
}
if (flag == -1)
{
end = 1;
}
}
while (end != 1 && temp2 != null);
if (flag == 0 || flag == 1)
{
Insert(cur, temp1.index);
frontSize++;
temp1 = Delete(temp1);
}
temp1 = temp1.child;
}
while (temp1 != null);
temp2 = cur.child;
do
{
newPopulation.IndList[temp2.index].Rank = rank;
temp2 = temp2.child;
}
while (temp2 != null);
assign_crowding_distance_list(newPopulation, cur.child, frontSize, problemObj, randomizationObj);
temp2 = cur.child;
do
{
temp2 = Delete(temp2);
temp2 = temp2.child;
}
while (cur.child != null);
rank += 1;
}
while (orig.child != null);
}
private bool HillClimber(ProblemDefinition problemObj, Random rnd)
{
int maxClimb = 1;
int climbCount = 0;
var tempColl = CollisionList.ToList();
if (tempColl.Count > 0) //collision varsa
{
while (tempColl.Count > 0)
{
var randomColl = rnd.Next(tempColl.Count);
var collision = tempColl[randomColl];
tempColl.RemoveAt(randomColl);
//var collision = tempColl.First();
//tempColl.RemoveAt(0);
if (collision.CrashingCourses.Count == 0) //ögrentmen ise geçiver şimdilik.
continue;
List<int> fittingSlots = new List<int>();
List<int> semiFittingSlots = new List<int>();
#region Course type collisions
foreach (var courseToReposition in collision.CrashingCourses.Where(x => !x.FacultyCourse).OrderBy(x => x.Duration)) //önce küçügü koy bir yerlere
{
int maxSlot = courseToReposition.Duration == 3 ? 20 : 25;
List<int> testSlots = new List<int>(maxSlot);
for (int i = 0; i < maxSlot; i++)
{
if (i != SlotId[courseToReposition.Id]) //eski yerini deneneceklerden çıkaralım.
testSlots.Add(i);
}
bool fittingSlot = false;
int semester = courseToReposition.Semester;
int duration = courseToReposition.Duration;
while (testSlots.Count > 0)
{
var randomSlotToTest = rnd.Next(testSlots.Count);
int slotToTest = testSlots[randomSlotToTest];
testSlots.RemoveAt(randomSlotToTest);
var temp = GetSlot(slotToTest, duration);
int day = GetX(slotToTest, duration);
int hour = GetY(slotToTest, duration);
var daySlots = GetDay(day);
#region Check Fitting
for (int j = 0; j < duration; j++)
{
fittingSlot = true;
temp[j].Courses.RemoveAll(x => x.Id == courseToReposition.Id); //kendini çıkarıyorum ki kendinle çakışmasın.
#region base vs faculty same semester
if (!courseToReposition.Elective) //elektif degilse, o saatte fakülte dersi olmamalı.
{
if (temp[j].facultyCourses.Any(x => x.Semester == semester)) //base vs faculty collision, same semester.
{
fittingSlot = false;
break;
}
}
#endregion
#region base vs base same semester
if (!courseToReposition.Elective) //elektif degilse, o saatte başka bölüm dersi olmamalı
{
if (temp[j].Courses.Count(x => x.Semester == semester && !x.Elective) > 0) //base vs base collision, same semester.
{
fittingSlot = false;
break;
}
}
#endregion
#region base vs faculty +1 -1
if (!courseToReposition.Elective) //elektif degilse,
{
if (semester - 1 > 0) // o saate bir geri dönem fakülte dersi olmasın
{
if (temp[j].facultyCourses.Any(x => x.Semester == semester - 1)) //base vs faculty collision, -1 semester.
{
fittingSlot = false;
break;
}
}
if (semester <= 8) // o saate bir ileri dönem fakülte dersi olmasın
{
if (temp[j].facultyCourses.Any(x => x.Semester == semester + 1)) //base vs faculty collision, +1 semester.
{
fittingSlot = false;
break;
}
}
}
#endregion
#region base vs base +1 -1
if (!courseToReposition.Elective)
{
if (semester - 1 > 0)
{
if (temp[j].Courses.Count(x => x.Semester == semester - 1 && !x.Elective) > 0) //base vs base collision, -1 semester.
{
fittingSlot = false;
break;
}
}
if (semester + 1 < 9)
{
if (temp[j].Courses.Count(x => x.Semester == semester + 1 && !x.Elective) > 0) //base vs base faculty collision, +1 semester.
{
fittingSlot = false;
break;
}
}
}
#endregion
#region LAB base vs faculty
if (courseToReposition.Type == 1) //lab ise
{
if (temp[j].labCount + temp[j].facultyLab >= 4) //base vs faculty collision, same semester.
{
fittingSlot = false;
break;
}
}
#endregion
#region Teacher coll
if (courseToReposition.Teacher != "ASSISTANT")
{
if (temp[j].Courses.Count(x => x.TeacherId == courseToReposition.TeacherId || temp[j].meetingHour) > 0) //check teacher collision
{
fittingSlot = false;
break;
}
}
#endregion
#region Teacher consecutive coll
//todo: obj2 og. gor. gunluk 4 saatten fazla pespese dersinin olmamasi
if (courseToReposition.Teacher != "ASSISTANT")
{
int maxConsecutiveHour = 4;
int counter = 0;
for (int k = 0; k < 9; k++)
{
Slot tempSlot = daySlots[k];
if (k == hour)
counter++;
if (tempSlot.Courses.Any(x => x.TeacherId == courseToReposition.TeacherId) || tempSlot.meetingHour) //dersi veya meetingi varsa ++
{
counter++;
}
else
{
counter = 0; // ara varsa 0'la
}
if (counter > maxConsecutiveHour)
{
fittingSlot = false;
break;
}
}
}
#endregion
//todo: obj2 og. gor. boş gununun olması
//bunu duration loop dışına alabilirim.
#region lab lecture collision
if (courseToReposition.LabHour > 0) //lab'ı var. //todo: obj2 lab ve lecture farklı günlerde olsun
{
if (courseToReposition.Type == 1)//labı yerleştiriyoruz.
{
if (daySlots.Any(slot => slot.Courses.Count(x => x.Code == courseToReposition.Code && x.Type == 0) > 0))
{
fittingSlot = false;
break;
}
}
else //dersi yerleştiriyoruz
{
if (daySlots.Any(slot => slot.Courses.Count(x => x.Code == courseToReposition.Code && x.Type == 1) > 0))
{
fittingSlot = false;
break;
}
}
}
#endregion
#region Elective vs Elective
if (courseToReposition.Elective)
{
if (temp[j].Courses.Count(x => x.Elective) > 0) //elective vs elective collision, all semesters.
{
fittingSlot = false;
break;
}
}
#endregion
#region Elective vs Faculty in semester 6 7 8
//todo: obj2
if (courseToReposition.Elective)
{
if (temp[j].facultyCourses.Any(x => x.Semester == 6)) //base vs faculty collision, same semester.
{
fittingSlot = false;
break;
}
if (temp[j].facultyCourses.Any(x => x.Semester == 7)) //base vs faculty collision, same semester.
{
fittingSlot = false;
break;
}
if (temp[j].facultyCourses.Any(x => x.Semester == 8)) //base vs faculty collision, same semester.
{
fittingSlot = false;
break;
}
}
#endregion
#region Elective vs Base
//elective vs base courses in semester
if (courseToReposition.Elective)
{
if (temp[j].Courses.Count(x => x.Semester == 8 & !x.Elective) > 0) //elective vs base collision, #8 semester.
{
fittingSlot = false;
break;
}
if (temp[j].Courses.Count(x => x.Semester == 7 & !x.Elective) > 0) //elective vs base collision, #7 semester.
{
fittingSlot = false;
break;
}
if (temp[j].Courses.Count(x => x.Semester == 6 & !x.Elective) > 0) //todo: obj1 elective vs base collision, #6 semester.
{
fittingSlot = false;
break;
}
}
else if (courseToReposition.Semester == 8 || courseToReposition.Semester == 7 || courseToReposition.Semester == 6)
{
if (temp[j].Courses.Count(x => x.Elective) > 0) //elective vs base collision, #8 semester.
{
fittingSlot = false;
break;
}
if (temp[j].Courses.Count(x => x.Elective) > 0) //elective vs base collision, #7 semester.
{
fittingSlot = false;
break;
}
if (temp[j].Courses.Count(x => x.Elective) > 0) //todo: obj1 elective vs base collision, #6 semester.
{
fittingSlot = false;
break;
}
}
#endregion
} // DURATION LOOP, ders saati kadar ileriye bakıyor.
#endregion
if (fittingSlot)
{
fittingSlots.Add(slotToTest);
break;
}
}
if (fittingSlots.Count > 0)
{
int selectedSlot = fittingSlots[rnd.Next(fittingSlots.Count)];
int checkDay = GetX(selectedSlot, duration) + 1;
int checkHour = GetY(selectedSlot, duration) + 1;
int fixedObj = collision.Obj;
ChangeGene(courseToReposition.Id, selectedSlot, problemObj);
SlotId[courseToReposition.Id] = selectedSlot;
climbCount++;
break;
}
else
{
//no fit
//Console.WriteLine("0 fitting slot found :(");
}
} //CRASHING COURSE
#endregion
if (climbCount >= maxClimb)
{
return true;
}
}
if (climbCount > 0)
return true;
//hiçbirini çözemedi.
return false;
}
return false;
}
public void Copy(Individual ind, ProblemDefinition problem)
{
NRealVar = ind.NRealVar;
NBinVar = ind.NBinVar;
NMaxBit = ind.NMaxBit;
NObj = ind.NObj;
NCons = ind.NCons;
CollisionList.Clear();
FacultySections.Clear();
DiffSemesterFacultySections.Clear();
ElectiveFacultySections.Clear();
TotalResult = 0;
Rank = ind.Rank;
ConstrViolation = ind.ConstrViolation;
CrowdDist = ind.CrowdDist;
if (ind.NRealVar > 0)
{
for (int i = 0; i < ind.NRealVar; i++)
{
Xreal[i] = ind.Xreal[i];
}
}
if (ind.NBinVar > 0)
{
for (int i = 0; i < ind.NBinVar; i++)
{
SlotId[i] = ind.SlotId[i];
for (int j = 0; j < problem.Nbits[i]; j++)
{
Gene[i][j] = ind.Gene[i][j];
}
}
}
for (int i = 0; i < ind.NObj; i++)
{
Obj[i] = ind.Obj[i];
}
if (ind.NCons > 0)
{
for (int i = 0; i < ind.NCons; i++)
{
Constr[i] = ind.Constr[i];
}
}
}
public void Decode(ProblemDefinition problem)
{
TotalResult = 0;
CollisionList.Clear();
if(FacultySections == null)
FacultySections = new List<FacultySection>();
FacultySections.Clear();
if (DiffSemesterFacultySections == null)
DiffSemesterFacultySections = new List<FacultySection>();
DiffSemesterFacultySections.Clear();
if (ElectiveFacultySections == null)
ElectiveFacultySections = new List<FacultySection>();
ElectiveFacultySections.Clear();
if(Obj.Length == 0)
Obj = new double[problem.ObjectiveCount];
if (problem.BinaryVariableCount == 0)
return;
for (int j = 0; j < problem.BinaryVariableCount; j++)
{
SlotId[j] = DecodeGene(problem, j);
}
}
/* Function to print the information of feasible and non-dominated population in a file */
public void ReportFeasiblePopulation(ProblemDefinition problemObj, string name, string info)
{
Directory.CreateDirectory("report");
var file = File.OpenWrite($"report\\{problemObj.Title}_{name}_pop.out");
StreamWriter writer = new StreamWriter(file);
writer.WriteLine($"# {info}");
writer.WriteLine($"# of objectives = {problemObj.ObjectiveCount}, # of constraints = {problemObj.ConstraintCount}, # of real_var = {problemObj.RealVariableCount}, # of bits of bin_var = {problemObj.TotalBinaryBitLength}, constr_violation, rank, crowding_distance");
for (int i = 0; i < problemObj.PopulationSize; i++)
{
if (IndList[i].ConstrViolation == 0.0 && IndList[i].Rank == 1)
{
for (int j = 0; j < problemObj.ObjectiveCount; j++)
{
writer.Write($"{IndList[i].Obj[j].ToString(CultureInfo.InvariantCulture)}\t");
}
if (problemObj.ConstraintCount != 0)
{
for (int j = 0; j < problemObj.ConstraintCount; j++)
{
writer.Write($"{IndList[i].Constr[j].ToString("E")}\t");
}
}
if (problemObj.RealVariableCount != 0)
{
for (int j = 0; j < problemObj.RealVariableCount; j++)
{
writer.Write($"{IndList[i].Xreal[j].ToString("E")}\t");
}
}
if (problemObj.BinaryVariableCount != 0)
{
for (int j = 0; j < problemObj.BinaryVariableCount; j++)
{
for (int k = 0; k < problemObj.Nbits[j]; k++)
{
writer.Write($"{IndList[i].Gene[j][k]}\t");
}
}
}
writer.Write($"{IndList[i].ConstrViolation.ToString("E")}\t");
writer.Write($"{IndList[i].Rank}\t");
writer.Write($"{ IndList[i].CrowdDist.ToString("E")}\n");
}
}
writer.Flush();
writer.Close();
}
/* Routine to perform non-dominated sorting */
static void fill_nondominated_sort(Population mixedPop, Population newPop, ProblemDefinition problemObj, Randomization randomizationObj)
{
int flag;
int i, j;
int end;
int frontSize;
int archieveSize;
int rank = 1;
Lists pool;
Lists elite;
Lists temp1, temp2;
pool = new Lists();
elite = new Lists();
frontSize = 0;
archieveSize = 0;
pool.index = -1;
pool.parent = null;
pool.child = null;
elite.index = -1;
elite.parent = null;
elite.child = null;
temp1 = pool;
for (i = 0; i < 2 * problemObj.PopulationSize; i++)
{
Insert(temp1, i);
temp1 = temp1.child;
}
i = 0;
do
{
temp1 = pool.child;
Insert(elite, temp1.index);
frontSize = 1;
temp2 = elite.child;
temp1 = Delete(temp1);
temp1 = temp1.child;
do
{
temp2 = elite.child;
if (temp1 == null)
{
break;
}
do
{
end = 0;
flag = CheckDominance(mixedPop.IndList[temp1.index], mixedPop.IndList[temp2.index], problemObj);
if (flag == 1)
{
Insert(pool, temp2.index);
temp2 = Delete(temp2);
frontSize--;
temp2 = temp2.child;
}
if (flag == 0)
{
temp2 = temp2.child;
}
if (flag == -1)
{
end = 1;
}
}
while (end != 1 && temp2 != null);
if (flag == 0 || flag == 1)
{
Insert(elite, temp1.index);
frontSize++;
temp1 = Delete(temp1);
}
temp1 = temp1.child;
}
while (temp1 != null);
temp2 = elite.child;
j = i;
if (archieveSize + frontSize <= problemObj.PopulationSize)
{
do
{
newPop.IndList[i].Copy(mixedPop.IndList[temp2.index], problemObj);
//CopyIndividual(mixedPop.IndList[temp2.index], newPop.IndList[i]);
newPop.IndList[i].Rank = rank;
archieveSize += 1;
temp2 = temp2.child;
i += 1;
}
while (temp2 != null);
assign_crowding_distance_indices(newPop, j, i - 1, problemObj, randomizationObj);
rank += 1;
}
else
{
crowding_fill(mixedPop, newPop, i, frontSize, elite, problemObj, randomizationObj);
archieveSize = problemObj.PopulationSize;
for (j = i; j < problemObj.PopulationSize; j++)
{
newPop.IndList[j].Rank = rank;
}
}
temp2 = elite.child;
do
{
temp2 = Delete(temp2);
temp2 = temp2.child;
}
while (elite.child != null);
}
while (archieveSize < problemObj.PopulationSize);
}
///* Function to display the current population for the subsequent generation */
public void PlotPopulation(Population pop, ProblemDefinition problem, int genNo = 0, List<Individual> best = null)
{
if (!Use3D) //2D
{
var xr = new double[problem.PopulationSize];
var yr = new double[problem.PopulationSize];
for (int x = 0; x < problem.PopulationSize; x++)
{
xr[x] = pop.IndList[x].Obj[GnuplotObjective1];
yr[x] = pop.IndList[x].Obj[GnuplotObjective2];
}
if (best != null)
{
GnuPlot.HoldOn();
}
GnuPlot.Plot(xr, yr, $"title 'Generation #{genNo} of {problem.MaxGeneration}' with points pointtype 8 lc rgb 'blue'");
if (best != null)
{
var x = new double[best.Count];
var y = new double[best.Count];
for (int i = 0; i < best.Count; i++)
{
x[i] = best[i].Obj[GnuplotObjective1];
y[i] = best[i].Obj[GnuplotObjective2];
}
GnuPlot.Plot(x, y, $"title 'best ({best.First().TotalResult})' with points pointtype 6 lc rgb 'red'");
}
GnuPlot.Set($"xlabel \"obj[{GnuplotObjective1 }]\"");
GnuPlot.Set($"ylabel \"obj[{GnuplotObjective2 }]\"");
}
else //3D
{
var rank1Count = pop.IndList.Count(x => x.Rank == 1);
var rankOtherCount = pop.IndList.Count(x => x.Rank != 1);
if (rank1Count > 0)
{
var xr = new double[rank1Count];
var yr = new double[rank1Count];
var zr = new double[rank1Count];
int index = 0;
foreach (var ind in pop.IndList.Where(x => x.Rank == 1))
{
xr[index] = ind.Obj[GnuplotObjective1];
yr[index] = ind.Obj[GnuplotObjective2];
zr[index] = ind.Obj[GnuplotObjective3];
index++;
}
if (best != null || rankOtherCount > 0)
{
GnuPlot.HoldOn();
}
GnuPlot.SPlot(xr, yr, zr, $"title 'Rank 1 individuals' with points pointtype 8 lc rgb 'red'");
}
if (rankOtherCount > 0)
{
var xr = new double[rankOtherCount];
var yr = new double[rankOtherCount];
var zr = new double[rankOtherCount];
int index = 0;
foreach (var ind in pop.IndList.Where(x => x.Rank != 1))
{
xr[index] = ind.Obj[GnuplotObjective1];
yr[index] = ind.Obj[GnuplotObjective2];
zr[index] = ind.Obj[GnuplotObjective3];
index++;
}
GnuPlot.SPlot(xr, yr, zr, $"title 'Rank 2-{pop.IndList.Max(x=>x.Rank)} individuals' with points pointtype 8 lc rgb 'blue'");
}
if (best != null)
{
var x = new double[best.Count];
var y = new double[best.Count];
var z = new double[best.Count];
for (int i = 0; i < best.Count; i++)
{
x[i] = best[i].Obj[GnuplotObjective1];
y[i] = best[i].Obj[GnuplotObjective2];
z[i] = best[i].Obj[GnuplotObjective3];
}
GnuPlot.SPlot(x, y, z, $"title 'best ({best.First().TotalResult})' with points pointtype 6 lc rgb 'green'");
}
GnuPlot.Set($"title 'Generation #{genNo} of {problem.MaxGeneration}'");
GnuPlot.Set($"xlabel \"obj[{GnuplotObjective1}]\"");
GnuPlot.Set($"ylabel \"obj[{GnuplotObjective2}]\"");
GnuPlot.Set($"zlabel \"obj[{GnuplotObjective3}]\"");
}
}
public UCTProblem(double dSeed, int nPopulation, int nMaxGeneration, int nObjective, int nConstraint, bool useBinary, double crossProbability, double mutateProbability, bool usePlot = false, DisabledCollisions disableOptions = DisabledCollisions.None)
{
CurrentGeneration = 0;
Seed = dSeed;
if (Seed <= 0.0 || Seed >= 1.0)
{
Console.WriteLine("\n Entered seed value is wrong, seed value must be in (0,1) \n");
Seed = 0.75;
}
UsePlot = usePlot;
var title = RandomTitle.GetRandomTitle();
ProblemObj = new ProblemDefinition(title)
{
PopulationSize = nPopulation,
MaxGeneration = nMaxGeneration,
ObjectiveCount = nObjective,
ConstraintCount = nConstraint,
BinaryVariableCount = 0,
RealVariableCount = 0,
DisabledCollisions = disableOptions
};
ProblemObj.TeacherList.Add("ASSISTANT");
ReadCourseList();
if (useBinary)
{
ProblemObj.BinaryVariableCount = ProblemObj.CourseList.Count;
PrepareBinaryLimits();
ProblemObj.BinaryCrossoverProbability = crossProbability;
if (ProblemObj.BinaryCrossoverProbability < 0.0 || ProblemObj.BinaryCrossoverProbability > 1.0)
{
ProblemObj.BinaryCrossoverProbability = 0.75;
}
ProblemObj.BinaryMutationProbability = mutateProbability;
if (ProblemObj.BinaryMutationProbability < 0.0 || ProblemObj.BinaryMutationProbability > 1.0)
{
ProblemObj.BinaryMutationProbability = 0.0232558;
}
}
else
{
ProblemObj.BinaryVariableCount = ProblemObj.CourseList.Count;
ReadRealValues();
ProblemObj.BinaryCrossoverProbability = crossProbability;
if (ProblemObj.BinaryCrossoverProbability < 0.0 || ProblemObj.BinaryCrossoverProbability > 1.0)
{
ProblemObj.BinaryCrossoverProbability = 0.75;
}
ProblemObj.BinaryMutationProbability = mutateProbability;
if (ProblemObj.BinaryMutationProbability < 0.0 || ProblemObj.BinaryMutationProbability > 1.0)
{
ProblemObj.BinaryMutationProbability = 0.0232558;
}
}
WriteStartParams();
RandomizationObj = new Randomization(dSeed);
RandomizationObj.Randomize();
_displayObj = new Display();
InitDisplay(true, true, new[] { 0, 1, 2 });
ReadFacultyCourses();
ReadLab();
ReadMeeting();
ReadPreq();
CreatePopulationObject();
}