protected virtual bool GetNextGen(Equation EvolvedEquation, Equation OldEquation, int toCalc)
{
bool BestCandEvolved = false;
Equation BestEvolvedEquation = BestCandidate.MakeClone(new Equation(EInfo, Randomizer));
Debug.WriteLine(BestCandidate.CreateFunction());
//int simplestCount = 0;
for (double i = 0; i < EInfo.CandidatesPerGen * EInfo.EvolvedCandidatesPerGen; i++)
{
BestCandidate.MakeClone(EvolvedEquation);
EvolveCand.EvolveCandidate(EInfo, EvolvedEquation);
EvolvedEquation.CalcPartialOffSet(toCalc);
//if (EvolvedEquation.CreateFunction() == "f(x) = x")
//{
// simplestCount++;
//}
//Debug.WriteLine(EvolvedEquation.CreateFunction());
//Thread.Sleep(500);
bool EvolvedToBetter = ChangeIfBetter(EvolvedEquation, OldEquation, BestEvolvedEquation);
BestCandEvolved = (EvolvedToBetter) ? true : BestCandEvolved;
ResetSingle(EvolvedEquation);
}
//Debug.WriteLine(simplestCount);
//Thread.Sleep(500);
int[] Indexes = SmartCand.CanSmartChangeNumbers(BestCandidate, OldEquation);
for (double i = 0; i < EInfo.CandidatesPerGen * EInfo.SmartCandidatesPerGen; i++)
{
BestCandidate.MakeClone(EvolvedEquation);
SmartCand.SmartifyCandidate(EInfo, EvolvedEquation, BestCandidate, OldEquation, Indexes);
EvolvedEquation.CalcPartialOffSet(toCalc);
bool EvolvedToBetter = ChangeIfBetter(EvolvedEquation, OldEquation, BestEvolvedEquation);
BestCandEvolved = (EvolvedToBetter) ? true : BestCandEvolved;
ResetSingle(EvolvedEquation);
}
for (double i = 0; i < EInfo.CandidatesPerGen * EInfo.RandomCandidatesPerGen; i++)
{
RandomCand.MakeRandomEquation(EvolvedEquation);
EvolvedEquation.CalcPartialOffSet(toCalc);
bool EvolvedToBetter = ChangeIfBetter(EvolvedEquation, OldEquation, BestEvolvedEquation);
BestCandEvolved = (EvolvedToBetter) ? true : BestCandEvolved;
ResetSingle(EvolvedEquation);
}
if (BestCandEvolved)
{
ResetSingle(OldEquation);
BestEvolvedEquation.MakeClone(OldEquation);
ResetSingle(BestCandidate);
BestEvolvedEquation.MakeClone(BestCandidate);
BestCandidate.CompressEquation();
}
return(BestCandEvolved);
}
public Family(EvolutionInfo eInfo, int parentCount)
{
parents = new Equation[parentCount];
children = new Equation[parentCount];
SynchronizedRandom random = new SynchronizedRandom();
for (int i = 0; i < parents.Length; i++)
{
parents[i] = new Equation(eInfo, random);
children[i] = new Equation(eInfo, random);
RandomCand.MakeValidRandomEquation(parents[i]);
parents[i].CalcTotalOffSet();
}
}
public override Genome EvolveSpecie()
{
StartFinding();
BestCandidate = new Equation(EInfo, Randomizer);
do
{
ResetSingle(BestCandidate);
RandomCand.MakeRandomEquation(BestCandidate);
BestCandidate.CalcTotalOffSet();
} while (!Tools.IsANumber(BestCandidate.OffSet));
Equation EvolvedEquation = new Equation(EInfo, Randomizer)
{
OffSet = float.NaN
};
Equation OldEquation = new Equation(EInfo, Randomizer)
{
OffSet = float.NaN
};
bool BestCandEvolved = false;
_toCalc = EInfo.coordInfo.expectedResults.Length;
while (_toCalc <= EInfo.coordInfo.expectedResults.Length)
{
int StuckCounter = 0;
do
{
BestCandEvolved = GetNextGen(EvolvedEquation, OldEquation, _toCalc);
StuckCounter = SetStuckCounter(StuckCounter, BestCandEvolved);
UpdateInfo();
SpecEnviroment.CheckBestCandidate(this.SpecInfo.GetCopy());
//} while (StuckCounter <= EInfo.MaxStuckGens && BestCandidate.OffSet != 0);
} while (StuckCounter <= EInfo.MaxStuckGens);
break;
//_toCalc++;
//BestCandidate.OffSet = double.MaxValue;
}
return(this);
}
