public void init() {
// System.out.println("TRIBES.init()");
// Generate a swarm
swarm =
new TribesSwarm(
this, range, initRange); // TODO initRange is hard coded equal to problem range
// swarm.generateSwarm(initExplorerNb, initType, m_problem);
// swarm.displaySwarm(swarm,out);
// print("\n Best after init: "+swarm.Best.position.fitness,out);
iter = 0;
adapt = 0;
informOption = -1;
// Hard coded option
// -1 = absolute best informant
// 1 = relative (pseudo-gradient) best informant. For "niching"
// See also moveExplorer, which can be modified in order to avoid this parameter
population.clear();
population.addAll(swarm.toPopulation());
population.init(); // necessary to allow for multi-runs
if (m_Show) show();
}
public void incEvalCnt() {
population.incrFunctionCalls();
if (notifyAfter(population.getFunctionCalls())) {
// System.out.println("Notifying after " + population.getFunctionCalls());
firePropertyChangedEvent(Population.nextGenerationPerformed);
}
}
private void plotAll(Population pop) {
// TODO
double pos[], vel[];
for (int i = 0; i < pop.size(); i++) {
pos = ((TribesExplorer) pop.getEAIndividual(i)).getDoubleData();
vel = ((TribesExplorer) pop.getEAIndividual(i)).getVelocity();
plotIndy(pos, vel, i);
// hier weiter!
}
}
/**
* As TRIBES manages an own structured set of particles (the list of Tribes containing explorers
* and memories), the setPopulation method is only telling Tribes the range of the indiviuals in
* the beginning of the run, the individuals will be discarded.
*/
public void setPopulation(Population pop) {
if (pop == null) return;
population = pop;
if (population.get(0) instanceof InterfaceDataTypeDouble) {
range = ((InterfaceDataTypeDouble) population.get(0)).getDoubleRange();
setDimension(range.length);
} else {
System.err.println(
"warning, TRIBES requires InterfaceESIndidivual instead of "
+ population.get(0).getClass()
+ ". Couldnt correctly init the problem range.");
}
}
/** Expects newPop to have correct number of generations set. */
public void adaptGenerational(
Population oldPop, Population selectedPop, Population newPop, boolean updateSelected) {
// nothing to do? Oh yes, we can easily transfer the cma-params from the old to the new
// population.
if (!newPop.hasData(cmaParamsKey)) {
if (!oldPop.hasData(cmaParamsKey))
System.err.println("warning: no cma param set found (MutateESRankMuCMA!");
else newPop.putData(cmaParamsKey, oldPop.getData(cmaParamsKey));
}
// newPop.addData(cmaParamsKey, oldPop.getData(cmaParamsKey));
// newPop.addPopulationChangedEventListener(this); // listen to every population to be informed
// about reinit events
}
/**
* Requires selected population to be sorted by fitness.
*
* @param params refering parameter set
* @param iterations number of iterations performed
* @param selected selected population
* @return true if the parameters seem ok or were corrected, false if new parameters must be
* produced
*/
private boolean testAndCorrectNumerics(
CMAParamSet params, int iterations, Population selected) { // not much left here
boolean corrected = true;
/* Flat Fitness, Test if function values are identical */
if (iterations > 1 && (selected.size() > 1)) {
// selected pop is sorted
if (nearlySame(
selected.getEAIndividual(0).getFitness(),
selected.getEAIndividual(selected.size() - 1).getFitness())) {
if (TRACE_1)
System.err.println(
"flat fitness landscape, consider reformulation of fitness, step-size increased");
params.sigma *= Math.exp(0.2 + params.c_sig / params.d_sig);
// sigma=0.1;
}
}
if (!checkValidDouble(params.sigma)) {
System.err.println("Error, unstable sigma!");
corrected = false;
// params.sigma=params.firstSigma; // MK TODO
// System.err.println(
}
/* Align (renormalize) scale C (and consequently sigma) */
/* e.g. for infinite stationary state simulations (noise
* handling needs to be introduced for that) */
double fac = 1.;
double minEig = 1e-12;
double maxEig = 1e8;
if (Mathematics.max(params.eigenvalues) < minEig)
fac = 1. / Math.sqrt(Mathematics.max(params.eigenvalues));
else if (Mathematics.min(params.eigenvalues) > maxEig)
fac = 1. / Math.sqrt(Mathematics.min(params.eigenvalues));
if (fac != 1.) {
// System.err.println("Scaling by " + fac);
params.sigma /= fac;
for (int i = 0; i < params.meanX.length; ++i) {
params.pathC[i] *= fac;
params.eigenvalues[i] *= fac * fac;
for (int j = 0; j <= i; ++j) {
params.mC.set(i, j, params.mC.get(i, j) * fac * fac);
if (i != j) params.mC.set(j, i, params.mC.get(i, j));
}
}
}
return corrected;
} // Test...
public void initByPopulation(Population pop, boolean reset) {
this.m_Population = (Population) pop.clone();
if (reset) {
this.m_Population.init();
this.m_Problem.evaluate(this.m_Population);
this.firePropertyChangedEvent(Population.nextGenerationPerformed);
}
}
protected TribesExplorer positionToExplorer(TribesPosition pos) {
TribesExplorer tmp = (TribesExplorer) population.get(0);
if (tmp == null) System.err.println("Error in Tribes::positionToExplorer!");
TribesExplorer indy = tmp.clone();
indy.clearPosVel();
indy.SetDoubleGenotype(pos.getPos());
indy.SetFitness(pos.getFitness());
return indy;
}
public AbstractEAIndividual getBestInd() {
TribesPosition bestMemPos = swarm.getBestMemory().getPos();
AbstractEAIndividual bestExp = population.getBestEAIndividual();
if (bestMemPos.firstIsBetter(bestMemPos.getFitness(), bestExp.getFitness())) {
AbstractEAIndividual indy = (AbstractEAIndividual) bestExp.clone();
indy.SetFitness(bestMemPos.getFitness());
((InterfaceDataTypeDouble) indy).SetDoubleGenotype(bestMemPos.getPos());
return indy;
} else return bestExp;
}
/**
* Test condition of C (Auger&Hansen, CEC '05, stopping criterion conditioncov). Return true, if a
* diagonal entry is <= 0 or >= d.
*
* @param d
* @return true, if a diagonal entry is <= 0 or >= d, else false
*/
public boolean testCCondition(Population pop, double d) {
// if (min(diag(D)) <= 0) || (max(diag(D)) > 1e14*min(diag(D)))
// stopflag(end+1) = {'warnconditioncov'};
CMAParamSet params = (CMAParamSet) pop.getData(cmaParamsKey);
Pair<Double, Double> minMax = params.mC.getMinMaxDiag();
if ((minMax.head <= 0) || (minMax.tail >= d)) {
if (TRACE_TEST) System.out.println("testCCondition hit");
return true;
} else return false;
}
/** This method will optimize */
public void optimize() {
AbstractEAIndividual indy;
Population original = (Population) this.m_Population.clone();
double tmpD;
InterfaceMutation tmpMut;
for (int i = 0; i < this.m_Population.size(); i++) {
indy = ((AbstractEAIndividual) this.m_Population.get(i));
tmpD = indy.getMutationProbability();
indy.setMutationProbability(1.0);
if (mutator == null) indy.mutate();
else mutator.mutate(indy);
indy.setMutationProbability(tmpD);
}
this.m_Problem.evaluate(this.m_Population);
for (int i = 0; i < this.m_Population.size(); i++) {
if (((AbstractEAIndividual) original.get(i))
.isDominatingDebConstraints(((AbstractEAIndividual) this.m_Population.get(i)))) {
this.m_Population.remove(i);
this.m_Population.add(i, original.get(i));
} else {
// else: mutation improved the individual
}
}
this.m_Population.incrGeneration();
// for (int i = 0; i < this.m_Population.size(); i++) {
// indy1 = (AbstractEAIndividual) this.m_Population.get(i);
// indy2 = (AbstractEAIndividual)(indy1).clone();
// indy2.mutate();
// this.m_Problem.evaluate((AbstractEAIndividual) indy2);
// //indy2.SetFitness(0, indy2.evaulateAsMiniBits());
// this.m_Population.incrFunctionCalls();
// //if (indy2.getFitness(0) < indy1.getFitness(0)) {
// if (indy2.isDominating(indy1)) {
// this.m_Population.remove(i);
// this.m_Population.add(i, indy2);
// }
// }
// this.m_Population.incrGeneration();
this.firePropertyChangedEvent(Population.nextGenerationPerformed);
}
/** This method is simply for debugging. */
protected void show() {
if (this.m_Plot == null) {
// InterfaceDataTypeDouble indy = (InterfaceDataTypeDouble)this.population.get(0);
// double[][] range = indy.getDoubleRange();
// double[] tmpD = new double[2];
// tmpD[0] = 0;
// tmpD[1] = 0;
this.m_Plot =
new eva2.gui.Plot("TRIBES " + population.getGeneration(), "x1", "x2", range[0], range[1]);
// this.m_Plot.setCornerPoints(range, 0);
}
}
/**
* Return a SolutionSet of TribesExplorers (AbstractEAIndividuals) of which some where memory
* particles, thus the returned population is larger than the current population.
*
* @return a population of possible solutions.
*/
public InterfaceSolutionSet getAllSolutions() {
// return population and memories?
Population all = (Population) population.clone();
List<TribesPosition> mems = swarm.collectMem();
for (Iterator<TribesPosition> iterator = mems.iterator(); iterator.hasNext(); ) {
TribesPosition tp = iterator.next();
all.add(positionToExplorer(tp));
}
all.SetFunctionCalls(population.getFunctionCalls());
all.setGenerationTo(population.getGeneration());
// all.addPopulation(pop);
return new SolutionSet(population, all);
}
/**
* From Auger&Hansen, CEC '05, stopping criterion noeffectcoord.
*
* @param d
* @return
*/
public boolean testNoEffectCoord(Population pop, double d) {
// if any(xmean == xmean + 0.2*sigma*sqrt(diag(C)))
// stopflag(end+1) = {'warnnoeffectcoord'};
boolean ret = false;
CMAParamSet params = (CMAParamSet) pop.getData(cmaParamsKey);
int i = 0;
while ((i < params.meanX.length) && !ret) {
ret =
ret
|| (params.meanX[i]
== (params.meanX[i] + d * getSigma(params, i) * Math.sqrt(params.mC.get(i, i))));
i++;
}
if (TRACE_TEST) if (ret) System.out.println("testNoEffectCoord hit");
return ret;
}
/**
* From Auger&Hansen, CEC '05, stopping criterion TolX.
*
* @param tolX
* @return
*/
public boolean testAllDistBelow(Population pop, double tolX) {
// if all(sigma*(max(abs(pc), sqrt(diag(C)))) < stopTolX)
boolean res = true;
CMAParamSet params = (CMAParamSet) pop.getData(cmaParamsKey);
int i = 0;
while (res && i < params.meanX.length) {
res =
res
&& (getSigma(params, i)
* Math.max(Math.abs(params.pathC[i]), Math.sqrt(params.mC.get(i, i)))
< tolX);
i++;
}
if (TRACE_TEST) if (res) System.out.println("testAllDistBelow hit");
return res;
}
/**
* This method allows you to merge to populations into an archive. This method will add elements
* from pop to the archive but will also remove elements from the archive if the archive target
* size is exceeded.
*
* @param pop The population that may add Individuals to the archive.
*/
public void addElementsToArchive(Population pop) {
if (pop.getArchive() == null) pop.SetArchive(new Population());
// test for each element in population if it
// is dominating a element in the archive
for (int i = 0; i < pop.size(); i++) {
if (this.isDominant((AbstractEAIndividual) pop.get(i), pop.getArchive())) {
this.addIndividualToArchive(
(AbstractEAIndividual) ((AbstractEAIndividual) pop.get(i)).clone(), pop.getArchive());
}
}
// Now clear the archive of surplus individuals
Population archive = pop.getArchive();
this.m_Cleaner.removeSurplusIndividuals(archive);
}
/**
* Retrieve the initial sigma for the given population and the user defined method. For the
* halfRange case, -1 is returned, as the range is not available here but set in initializing the
* CMAParams.
*
* @param initGen
* @return
*/
private double getInitSigma(Population initGen) {
switch (initializeSig) {
case avgInitialDistance:
// scaled by average range as the measures are normed
// return initGen.getPopulationMeasures(null)[0]*getAvgRange();
// use euclidian measures without normation and scaling
return initGen
.getPopulationMeasures(new EuclideanMetric())[
0]; // use euclidean metric which is not normed by range instead of phenotype metric
// case halfRange: return getAvgRange(range)/2.;
case userDefined:
return userDefInitSig;
case halfRange:
return -0.5;
case quarterRange:
return -0.25;
}
throw new RuntimeException("Unknown initial sigma type!");
}
/**
* From Auger&Hansen, CEC '05, stopping criterion noeffectaxis.
*
* @param d
* @param gen
* @return
*/
public boolean testNoChangeAddingDevAxis(Population pop, double d, int gen) {
// if all(xmean == xmean + 0.1*sigma*BD(:,1+floor(mod(countiter,N))))
// i = 1+floor(mod(countiter,N));
// stopflag(end+1) = {'warnnoeffectaxis'};
CMAParamSet params = (CMAParamSet) pop.getData(cmaParamsKey);
int dim = params.meanX.length;
int k = gen % dim;
double[] ev_k = params.mB.getColumn(k);
Mathematics.svMult(
Math.sqrt(params.eigenvalues[k]), ev_k, ev_k); // this is now e_k*v_k = BD(:,...)
int i = 0;
boolean res = true;
while (res && (i < dim)) {
res = res && (params.meanX[i] == (params.meanX[i] + d * getSigma(params, i) * ev_k[i]));
i++;
}
if (TRACE_TEST) if (res) System.out.println("testNoChangeAddingDevAxis hit");
return res;
}
/** Override population evaluation to do some data output. */
public void evaluatePopulationEnd(Population population) {
// the translation on the first two dimensions
// double delta = Math.sqrt(Math.pow(getTranslation(0, getCurrentProblemTime()), 2.) +
// Math.pow(getTranslation(1, getCurrentProblemTime()), 2.));
double delta =
getTranslation(0, getCurrentProblemTime()) + getTranslation(1, getCurrentProblemTime());
if (isExtraPlot() == true) {
if (myplot != null) {
myplot.jump();
} else {
if (TRACE) System.out.println("creating myplot instance");
double[] tmpD = new double[2];
tmpD[0] = 0;
tmpD[1] = 0;
// im not really certain about what tmpD is required for
this.myplot = new Plot("population measures", "x1", "x2", tmpD, tmpD);
}
myplot.setConnectedPoint(population.getFunctionCalls(), delta, 0);
// myplot.setUnconnectedPoint(population.getFunctionCalls(),
// population.getPopulationMeasures()[2], 2);
} else myplot = null;
}
/* update C */
private void updateCov(
CMAParamSet params,
double[] newPathC,
double[] newMeanX,
double hsig,
int mu,
Population selected) {
double newVal = 0;
int dim = newMeanX.length;
double ccv = getCCov(params.weights, mu, dim);
if (ccv > 0) {
double mcv = CMAParamSet.getMuCov(params.weights, mu);
/* (only upper triangle!) */
/* update covariance matrix */
// System.out.println("CCov " + getCCov(selected) + " Cc " + getCc() + " muCov " +
// getMuCov(selected));
for (int i = 0; i < dim; ++i)
for (int j = 0; j <= i; ++j) {
// oldVal = mC.get(i,j);
newVal =
(1 - ccv) * params.mC.get(i, j)
+ ccv
* (1. / mcv)
* (newPathC[i] * newPathC[j]
+ (1 - hsig) * getCc() * (2. - getCc()) * params.mC.get(i, j));
checkValidDouble(newVal);
params.mC.set(i, j, newVal);
if (isRankMu()) {
for (int k = 0; k < mu; ++k) {
/*
* additional rank mu
* update
*/
// double[] x_k =
// ((InterfaceDataTypeDouble)selected.getEAIndividual(k)).getDoubleData();
double[] x_k =
AbstractEAIndividual.getDoublePositionShallow(selected.getEAIndividual(k));
newVal =
params.mC.get(i, j)
+ ccv
* (1 - 1. / mcv)
* params.weights[k]
* (x_k[i] - params.meanX[i])
* (x_k[j] - params.meanX[j])
/ (getSigma(params, i) * getSigma(params, j)); // TODO right sigmas?
checkValidDouble(newVal);
params.mC.set(i, j, newVal);
}
}
}
// fill rest of C
for (int i = 0; i < dim; ++i) {
for (int j = i + 1; j < dim; ++j) {
params.mC.set(i, j, params.mC.get(j, i));
}
}
if (params.mC.get(0, 1) != params.mC.get(1, 0)) {
System.err.println("WARNING, C is not symmetric!");
}
// maxsqrtdiagC = Math.sqrt(math.max(math.diag(C)));
// minsqrtdiagC = Math.sqrt(math.min(math.diag(C)));
} // update of C
}
public void optimize() {
int initOption = 0;
if (iter == 0) { // first iteration!
if (initRange == null) {
rangeInitType = 0;
} else {
rangeInitType = 1; // 1 means in initRange for a start
}
// * initOption Options: 0 - random, 1 - on the bounds, 2 - sunny spell, 3 - around a center
// * rangeInitType for options 0,1: 1 means use initRange, 0 use default range
swarm.generateSwarm(initExplorerNb, initOption, rangeInitType, m_problem);
}
iter++;
m_problem.evaluatePopulationStart(population);
swarm.setSwarmSize();
// swarm.Best.positionPrev = swarm.Best.position;
swarm.moveSwarm(
range, new TribesParam(), informOption, m_problem); // *** HERE IT MOVES and EVALUATES
// public void moveSwarm(double[][] range, int fitnessSize, TribesParam pb, TribesSwarm swarm,
// int informOption, AbstractOptimizationProblem prob) {
if (Tribes.adaptOption != 0) { // perform adaption
if (Tribes.adaptOption == 1) { // Just reinitialize the swarm
adaptThreshold = iter - adapt;
// adaptMax=swarmSize;
adaptMax = swarm.linkNb(swarm);
if (adaptThreshold >= adaptMax) {
if (swarm.getBestMemory().getPrevPos().getTotalError()
<= swarm.getBestMemory().getPos().getTotalError()) {
adapt = iter; // Memorize at which iteration adaptation occurs
for (int i = 0; i < swarm.getTribeCnt(); i++) {
swarm.reinitTribe(i, rangeInitType, m_problem);
}
}
}
} else // if(swarm.Best.positionPrev.getTotalError()<=swarm.Best.position.getTotalError())
{
// Structural adaptations
adaptThreshold = iter - adapt;
// adaptMax=swarmSize;
adaptMax = swarm.linkNb(swarm);
if (adaptThreshold >= adaptMax) {
adapt = iter; // Memorize at which iteration adaptation occurs
swarm.adaptSwarm(rangeInitType, m_problem); // Re´alise l'adaptation
}
}
}
population.clear();
population.addAll(swarm.toPopulation());
if (m_Show) plotAll(population);
m_problem.evaluatePopulationEnd(population);
// this.population.incrFunctionCallsby(evals);
this.population.incrGeneration();
// this.firePropertyChangedEvent("NextGenerationPerformed"); // This is now done implicitely, as
// after every evaluation, addEvals is called
if (TRACE) {
System.out.println("loop finished after " + population.getFunctionCalls() + " evaluations");
// for (int i=0; i<population.size(); i++) System.out.println(" *
// "+((TribesExplorer)population.get(i)).getStringRepresentation());
System.out.println(
" best: "
+ population.getBestEAIndividual().getStringRepresentation()
+ " - "
+ population.getBestEAIndividual().getFitness(0));
System.out.println(
"swarm contains " + swarm.numParticles() + " particles in iteration " + iter);
}
}
/**
* After optimization start, this returns the initial sigma value actually employed.
*
* @return the initial sigma value actually employed
*/
public double getFirstSigma(Population pop) {
return ((CMAParamSet) pop.getData(cmaParamsKey)).firstSigma;
}
public static double[] getMeanXOfPop(Population pop) {
CMAParamSet params = (CMAParamSet) pop.getData(MutateESRankMuCMA.cmaParamsKey);
if (params == null) return null;
else return params.meanX;
}
/**
* calculate weighted mean of the selected population
*
* @param selectedPop
* @return
*/
private double[] calcMeanX(double[] weights, Population selectedPop) {
return selectedPop.getCenterWeighted(weights);
}
/**
* Perform the main adaption of sigma and C using evolution paths. The evolution path is deduced
* from the center of the selected population compared to the old mean value. See Hansen&Kern 04
* for further information.
*
* @param oldGen
* @param selectedP
*/
public void adaptAfterSelection(Population oldGen, Population selectedP) {
Population selectedSorted =
selectedP.getSortedBestFirst(new AbstractEAIndividualComparator(-1));
int mu, lambda;
mu = selectedP.size();
lambda = oldGen.size();
int generation = oldGen.getGeneration();
if (mu >= lambda) {
// try to override by oldGen additional data:
if (oldGen.hasData(EvolutionStrategies.esMuParam))
mu = (Integer) oldGen.getData(EvolutionStrategies.esMuParam);
if (oldGen.hasData(EvolutionStrategies.esLambdaParam))
lambda = (Integer) oldGen.getData(EvolutionStrategies.esLambdaParam);
}
if (mu >= lambda) {
mu = Math.max(1, lambda / 2);
EVAERROR.errorMsgOnce(
"Warning: invalid mu/lambda ratio! Setting mu to lambda/2 = "
+ mu
+ ", lambda = "
+ lambda);
}
CMAParamSet params;
if (oldGen.getGeneration()
<= 1) { // init new param set. At gen < 1 we shouldnt be called, but better do it once too
// often
if (oldGen.hasData(cmaParamsKey))
params =
CMAParamSet.initCMAParams(
(CMAParamSet) oldGen.getData(cmaParamsKey),
mu,
lambda,
oldGen,
getInitSigma(oldGen));
else params = CMAParamSet.initCMAParams(mu, lambda, oldGen, getInitSigma(oldGen));
} else {
if (!oldGen.hasData(cmaParamsKey)) {
if (oldGen.getGeneration() > 1)
EVAERROR.errorMsgOnce("Error: population lost cma parameters. Incompatible optimizer?");
params = CMAParamSet.initCMAParams(mu, lambda, oldGen, getInitSigma(oldGen));
} else params = (CMAParamSet) oldGen.getData(cmaParamsKey);
}
if (lambda == 1
&& (oldGen.size() == 1)
&& (selectedP.size() == 1)
&& (oldGen.getEAIndividual(0).equals(selectedP.getEAIndividual(0)))) {
// nothing really happened, so do not adapt and just store default params
lastParams = (CMAParamSet) params.clone();
oldGen.putData(cmaParamsKey, params);
selectedP.putData(cmaParamsKey, params);
return;
}
if (TRACE_1) {
System.out.println("WCMA adaptGenerational **********");
// System.out.println("newPop measures: " +
// BeanInspector.toString(newPop.getPopulationMeasures()));
System.out.println("mu_eff: " + CMAParamSet.getMuEff(params.weights, mu));
System.out.println(params.toString());
System.out.println("*********************************");
}
double[] newMeanX = calcMeanX(params.weights, selectedSorted);
if (TRACE_1) System.out.println("newMeanX: " + BeanInspector.toString(newMeanX));
int dim = params.meanX.length;
double[] BDz = new double[dim];
for (int i = 0; i < dim; i++) {
/* calculate xmean and BDz~N(0,C) */
// Eq. 4 from HK04, most right term
BDz[i] =
Math.sqrt(CMAParamSet.getMuEff(params.weights, mu))
* (newMeanX[i] - params.meanX[i])
/ getSigma(params, i);
}
// if (TRACE_2) System.out.println("BDz is " + BeanInspector.toString(BDz));
double[] newPathS = params.pathS.clone();
double[] newPathC = params.pathC.clone();
double[] zVect = new double[dim];
/* calculate z := D^(-1) * B^(-1) * BDz into artmp, we could have stored z instead */
for (int i = 0; i < dim; ++i) {
double sum = 0.;
for (int j = 0; j < dim; ++j) {
sum += params.mB.get(j, i) * BDz[j]; // times B transposed, (Eq 4) in HK04
}
if (params.eigenvalues[i] < 0) {
EVAERROR.errorMsgOnce(
"Warning: negative eigenvalue in MutateESRankMuCMA! (possibly multiple cases)");
zVect[i] = 0;
} else {
zVect[i] = sum / Math.sqrt(params.eigenvalues[i]);
if (!checkValidDouble(zVect[i])) {
System.err.println("Error, infinite zVect entry!");
zVect[i] = 0; // TODO MK
}
}
}
/* cumulation for sigma (ps) using B*z */
for (int i = 0; i < dim; ++i) {
double sum = 0.;
for (int j = 0; j < dim; ++j) sum += params.mB.get(i, j) * zVect[j];
newPathS[i] =
(1. - params.c_sig) * params.pathS[i]
+ Math.sqrt(params.c_sig * (2. - params.c_sig)) * sum;
if (!checkValidDouble(newPathS[i])) {
System.err.println("Error, infinite pathS!");
}
}
// System.out.println("pathS diff: " + BeanInspector.toString(Mathematics.vvSub(newPathS,
// pathS)));
// System.out.println("newPathS is " + BeanInspector.toString(newPathS));
double psNorm = Mathematics.norm(newPathS);
double hsig = 0;
if (psNorm / Math.sqrt(1. - Math.pow(1. - params.c_sig, 2. * generation)) / expRandStepLen
< 1.4 + 2. / (dim + 1.)) {
hsig = 1;
}
for (int i = 0; i < dim; ++i) {
newPathC[i] =
(1. - getCc()) * params.pathC[i] + hsig * Math.sqrt(getCc() * (2. - getCc())) * BDz[i];
checkValidDouble(newPathC[i]);
}
// TODO missing: "remove momentum in ps"
if (TRACE_1) {
System.out.println("newPathC: " + BeanInspector.toString(newPathC));
System.out.println("newPathS: " + BeanInspector.toString(newPathS));
}
if (TRACE_1) System.out.println("Bef: C is \n" + params.mC.toString());
if (params.meanX == null) params.meanX = newMeanX;
updateCov(params, newPathC, newMeanX, hsig, mu, selectedSorted);
updateBD(params);
if (TRACE_2) System.out.println("Aft: C is " + params.mC.toString());
/* update of sigma */
double sigFact = Math.exp(((psNorm / expRandStepLen) - 1) * params.c_sig / params.d_sig);
if (Double.isInfinite(sigFact))
params.sigma *= 10.; // in larger search spaces sigma tends to explode after init.
else params.sigma *= sigFact;
if (!testAndCorrectNumerics(params, generation, selectedSorted)) {
// parameter seemingly exploded...
params =
CMAParamSet.initCMAParams(
params,
mu,
lambda,
params.meanX,
((InterfaceDataTypeDouble) oldGen.getEAIndividual(0)).getDoubleRange(),
params.firstSigma);
}
if (TRACE_1) {
System.out.println("sigma=" + params.sigma);
System.out.print("psLen=" + (psNorm) + " ");
outputParams(params, mu);
}
// take over data
params.meanX = newMeanX;
params.pathC = newPathC;
params.pathS = newPathS;
params.firstAdaptionDone = true;
lastParams = (CMAParamSet) params.clone();
oldGen.putData(cmaParamsKey, params);
selectedP.putData(cmaParamsKey, params);
// if (TRACE_2) System.out.println("sampling around " + BeanInspector.toString(meanX));
}
