public Solution Solve(bool isContinueFromLastSolution = false, bool isOneGenerationOnly = false)
{
Solution solution;
CMAState cmaState;
if ((!(isContinueFromLastSolution)) || (_solution.Equals(default(Solution))))
{
solution = new Solution()
{
StopReason = StopReason.NotSet
};
cmaState = new CMAState(_initialCmaParams, new ColumnVector(this.InitialParameters), this.InitialVarianceEstimate);
}
else
{
solution = _solution;
solution.StopReason = StopReason.Restart;
cmaState = _solution.LatestCMAState;
}
int iteration = 0;
while ((solution.StopReason == StopReason.NotSet) || (solution.StopReason == StopReason.Restart))
{
Solve(iteration++, ref cmaState, ref solution, isOneGenerationOnly);
}
_solution = solution;
return(solution);
}
private bool IsRestart(ref Solution solution, ColumnVector fitnesses, double tolX, CMAState cmaState)
{
bool isAllNaN = fitnesses.All(x => double.IsNaN(x) || double.IsInfinity(x));
if (isAllNaN)
{
return(true);
}
// Stop if the condition number of the covariance matrix
//exceeds 1014 (conditioncov).
bool isCConditionNumberTooHigh = false;
if ((Utilities.Diag(cmaState.C, x => x).ToArray().Max() / Utilities.Diag(cmaState.C, x => x).ToArray().Min()) > 1e14)
{
isCConditionNumberTooHigh = true;
}
if (isCConditionNumberTooHigh)
{
return(true);
}
if (solution.History == null)
{
solution.History = new Queue <double>();
}
//Stop if the range of the best objective function values
//of the last 10 + [30n/lambda] generations is zero
//(equalfunvalhist), or the range of these function
//values and all function values of the recent generation
//is below Tolfun= 10^-12.
solution.History.Enqueue(fitnesses[0]);
if (solution.History.Count > (10.0 + ((30.0 * cmaState.p.N) / cmaState.p.Lambda)))
{
solution.History.Dequeue();
}
bool isObjectiveFunctionRangeTooLow = (solution.History.Count == (int)(10.0 + ((30.0 * cmaState.p.N) / cmaState.p.Lambda))) &&
((solution.History.Max() == solution.History.Min()) ||
(((solution.History.Max() - solution.History.Min()) < ObjectiveFunctionTolerance) && ((fitnesses.Max() - fitnesses.Min()) < ObjectiveFunctionTolerance)));
if (isObjectiveFunctionRangeTooLow)
{
return(true);
}
// Stop if the standard deviation of the normal distribution
//is smaller than TolX in all coordinates and [sigma]pc
//(the evolution path from Eq. 2 in [3]) is smaller than
//TolX in all components. We set TolX= 10^-12*[sigma]^(0).
bool isStandardDeviationTooSmall = (Utilities.IsTrueForAll(cmaState.C, x => Math.Abs(x) < tolX) && cmaState.ps.All(x => Math.Abs(x) < tolX));
if (isStandardDeviationTooSmall)
{
return(true);
}
// Stop if adding a 0.1-standard deviation vector in
//a principal axis direction of C^(g) does not change
//<x[vector]>w^(g) (noeffectaxis)3
int ith = (cmaState.gen % cmaState.p.N);
ColumnVector tmpXmean = (cmaState.mean + 0.1 * cmaState.sigma * cmaState.B * new ColumnVector(cmaState.d));
bool isNoEffectAxis = false;
if (ith < tmpXmean.Dimension)
{
isNoEffectAxis = (tmpXmean[ith] == cmaState.mean[ith]);
}
if (isNoEffectAxis)
{
return(true);
}
// Stop if adding 0.2-standard deviation in each coordinate
//does change <x[vector]>w^(g) (noeffectcoord).
SquareMatrix testC = 0.2 * cmaState.sigma * cmaState.C;
SquareMatrix vectors;
SquareMatrix values;
Utilities.EigAlgLib(testC, out vectors, out values, 1, isGetLower: false);
ColumnVector colValues = Utilities.Diag(values, x => Math.Sqrt(Math.Abs(x)));
tmpXmean = (cmaState.mean + cmaState.sigma * vectors * colValues);
bool isNoEffectCoord = true;
for (int i = 0; i < cmaState.mean.Dimension; i++)
{
if (cmaState.mean[i] != tmpXmean[i])
{
isNoEffectCoord = false;
break;
}
}
if (isNoEffectCoord)
{
return(true);
}
return(false);
}
private void Solve(int iteration, ref CMAState cmaState, ref Solution solution, bool isOnce = false)
{
//ColumnVector xmean = new ColumnVector(parameters);
double tolX = this.StandardDeviationToleranceMultiple * this.InitialVarianceEstimate;
int stopeval = this.StopEvals;
RectangularMatrix arx = new RectangularMatrix(cmaState.p.N, cmaState.p.Lambda);
ColumnVector arfitness = new ColumnVector(cmaState.p.Lambda);
bool isRestart = false;
bool isGoodEnough = false;
int counteval = 0;
while (counteval < stopeval)
{
// Generate and evaluate lambda offspring
for (int k = 0; k < cmaState.p.Lambda; k++)
{
double[] sample = cmaState.Sample(arx.Column(k));
for (int r = 0; r < cmaState.p.N; r++)
{
arx[r, k] = sample[r];
}
arfitness[k] = this.ObjectiveFunction(arx.Column(k).ToArray(), this.UserData);
counteval++;
}
// Sort by fitness and compute weighted mean into xmean
// minimization
int[] arindex;
arfitness = Utilities.Sort(arfitness, out arindex); // one based index
int mu = cmaState.p.Mu;
cmaState.ReEstimate(
(RectangularMatrix)Utilities.GetColumns(arx, arindex.Take(cmaState.p.Mu)),
arfitness.First(),
arfitness.SkipWhile((f, i) => i < (mu - 1)).First());
cmaState.UpdateEigenSystem(1, 0);
if ((solution.StopReason == StopReason.NotSet) || (arfitness[0] < solution.MinimizedFitness))
{
solution.StopReason = StopReason.Incomplete;
solution.MinimizedFitness = arfitness[0];
solution.BestCMAState = cmaState;
solution.BestNumEvals = counteval;
solution.BestNumRestarts = iteration;
solution.BestParameters = cmaState.mean.ToArray();
}
if (arfitness[0] <= this.StopFitnessMinimization)
{
isGoodEnough = true;
break;
}
if (IsRestart(ref solution, arfitness, tolX, cmaState))
{
isRestart = true;
break;
}
if (isOnce)
{
break;
}
}
if ((isRestart) && (iteration > this.MaxRestarts))
{
solution.StopReason = StopReason.MaxRestarts;
}
else if ((isRestart) && ((cmaState.p.Lambda * this.RestartPopulationMultiple) > MAX_POPULATIONSIZE))
{
solution.StopReason = StopReason.MaxPopulation;
}
else if (isRestart)
{
solution.History = new Queue <double>();
solution.StopReason = StopReason.Restart;
int previousLambda = cmaState.p.Lambda;
cmaState = new CMAState(_initialCmaParams, new ColumnVector(this.InitialParameters), this.InitialVarianceEstimate);
cmaState.p.UpdateLambda((int)(previousLambda * this.RestartPopulationMultiple));
}
else if (isGoodEnough)
{
solution.StopReason = StopReason.GoodEnough;
}
else
{
solution.StopReason = StopReason.MaxIterations;
}
solution.LatestCMAState = cmaState;
solution.LatestNumEvals = counteval;
solution.LatestNumRestarts = iteration;
solution.LatestParameters = cmaState.mean.ToArray();
}
