private void OutputAvg(GeneticTS geneticTS, List <List <double> > bestMaster, List <List <double> > avgMaster, string experimentName)
{
if (!File.Exists("output.csv"))
{
File.Create("output.csv").Close();
}
StreamWriter data = new StreamWriter("output.csv", true);
string[] expInfo = geneticTS.GetExperimentInfo();
foreach (var line in expInfo)
{
data.WriteLine(line);
}
data.WriteLine("");
data.WriteLine(experimentName);
data.WriteLine("Best Fitness " + "," + "Average Fitness");
//Depth
for (int i = 0; i < bestMaster.Max(x => x.Count) - 1; i++)
{
//Breadth
double avgBest = 0;
double avgAvg = 0;
for (int k = 0; k < bestMaster.Count; k++)
{
avgBest += (i >= bestMaster[k].Count) ? bestMaster[k].Last() : bestMaster[k][i];
avgAvg += (i >= avgMaster[k].Count) ? avgMaster[k].Last() : avgMaster[k][i];
}
data.WriteLine(avgBest / bestMaster.Count + "," + avgAvg / avgMaster.Count);
}
data.WriteLine("");
data.Close();
data.Dispose();
}
private void Output1(GeneticTS geneticTS, List <double> bestMaster, List <double> avgMaster, string experimentName)
{
if (!File.Exists("output.csv"))
{
File.Create("output.csv").Close();
}
StreamWriter data = new StreamWriter("output.csv", true);
string[] expInfo = geneticTS.GetExperimentInfo();
foreach (var line in expInfo)
{
data.WriteLine(line);
}
data.WriteLine("");
data.WriteLine(experimentName);
data.WriteLine("Best Fitness " + "," + "Average Fitness");
for (int i = 0; i < bestMaster.Count; i++)
{
data.WriteLine(bestMaster[i] + ", " + avgMaster[i]);
}
data.WriteLine("");
data.Close();
data.Dispose();
}
public Program()
{
//Setup
if (File.Exists("output.csv"))
{
File.Delete("output.csv");
}
List <double[, ]> datasets = new List <double[, ]> {
ulyssess22TSP, eil51TSP
};
string s = "";
do
{
Console.Clear();
Console.WriteLine("1. Output all permutations. 2. Output custom run");
s = Console.ReadLine();
}while (s != "1" && s != "2");
switch (s)
{
case "1":
double[,] set = eil51TSP;
int mPopSize = 1000;
int mGenerationSpan = 1000;
int tSize = 5;
int round = 5;
int startCity = 1;
bool useConvergence = true;
ElitismMode elitMode = ElitismMode.top10P;
CrossoverType crossoverType = CrossoverType.PMX;
double crossoverRate = 1.0;
double mutateRate = 0.1;
string seed = "SampleRandomSeed";
City[] cit = new City[set.GetLength(0)];
for (int i = 0; i < set.GetLength(0); i++)
{
cit[i] = new City(set[i, 1], set[i, 2]);
}
GeneticTS g1 = new GeneticTS(
elitMode,
crossoverType,
crossoverRate,
mutateRate,
mPopSize,
mGenerationSpan,
seed.GetHashCode(),
tSize,
startCity,
cit,
useConvergence,
round,
CalculateSafeZone(elitMode, mPopSize)
);
Output1(g1, g1.bestFitnesses, g1.avgFitnesses, "CustomRun");
break;
case "2":
List <double> crossovers = new List <double> {
1, .9
};
List <double> mutations = new List <double> {
0, .1, 1
};
int maxPopSize = 1000;
int maxGenerationSpan = 1000;
int tournamentSize = 5;
int roundDigits = 5;
int startCityIndex = 1;
bool allowConvergence = true;
foreach (double[,] dataset in datasets)
{
foreach (CrossoverType crossType in Enum.GetValues(typeof(CrossoverType)))
{
foreach (double crossRate in crossovers)
{
foreach (double mutRate in mutations)
{
foreach (ElitismMode eMode in Enum.GetValues(typeof(ElitismMode)))
{
GeneticTS geneticTS = null;
List <List <double> > setBest = new List <List <double> >();
List <List <double> > setAvg = new List <List <double> >();
string experimentName = string.Format("SET: {0} Elietism: {1} Crossover Type: {2} Crossover Rate: {3} Mutation Rate: {4}",
dataset.GetLength(0), eMode, crossType, crossRate, mutRate);
Console.WriteLine("Starting " + experimentName);
for (int b = 0; b < randomSeeds.Length; b++)
{
City[] cities = new City[dataset.GetLength(0)];
for (int i = 0; i < dataset.GetLength(0); i++)
{
cities[i] = new City(dataset[i, 1], dataset[i, 2]);
}
//Call GeneticTS
geneticTS = new GeneticTS(
eMode,
crossType,
crossRate,
mutRate,
maxPopSize,
maxGenerationSpan,
randomSeeds[b].GetHashCode(),
tournamentSize,
startCityIndex,
cities,
allowConvergence,
roundDigits,
CalculateSafeZone(eMode, maxPopSize)
);
setBest.Add(geneticTS.bestFitnesses);
setAvg.Add(geneticTS.avgFitnesses);
}
OutputAvg(geneticTS, setBest, setAvg, experimentName);
}
}
}
}
}
break;
}
}
public Crossover(GeneticTS geneticTS)
{
this.geneticTS = geneticTS;
}
