halfdan/eva2
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Some cosmetics, e.g. to older MOCCO classes

Browse Source
This commit is contained in:
Marcel Kronfeld
2009-03-11 13:30:48 +00:00
parent a347e220ff
commit 08c88a3ed6
15 changed files with 592 additions and 364 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
src/eva2/EvAInfo.java
View File

@@ -4,6 +4,16 @@ package eva2;
* Main product and version information strings.
*
* --- Changelog
* 2.033: There was an interesting problem with the Matlab-Interface, which just hung up after extensive optimization
* loops, yet only if Matlab was started without X-forwarding (which is necessary for qsub, e.g.).
* Debugging was tedious, since the debugging using System.out. itself caused threading deadlocks. The
* problem showed up to be with System.out itself. Whatever Matlab does with that stream, it does it differently
* depending on the X-forwarding option. More specifically, the text written to System.out when X-forwarding
* is not available seems to never show up (as opposed to being printed to the console when X-forwarding is on)
* and silently fill up the JVM-memory. I havent the faintest idea why there havnt been OutOfMemory exceptions
* earlier or whether and how the deadlocks have to do with it.
* The ingeniuos solution was: dont print anything to System.out, which is now done at verbosity 0.
* 2.032: Some cosmetics, e.g. to AbstractEAIndividualComparator and older MOCCO classes.
* 2.031: Some updates to the OptimizerFactory. Review of the MatlabInterface with adding an own options structure.
* Better access to the EvAClient, which now may have a RemoteStateListener added monitoring the optimization run.
* 2.030: Added an EnumEditor to access enums easily through the GUI, which will replace SelectedTags sometimes.

37
src/eva2/gui/BeanInspector.java
View File

@@ -594,6 +594,43 @@ public class BeanInspector {
return value;
}
/**
* Try to get an object member value.
*
* @param obj
* @param mem
* @return the value if successful, else null
*/
public static Object getMem(Object obj, String mem) {
BeanInfo bi;
try {
bi = Introspector.getBeanInfo(obj.getClass());
} catch(IntrospectionException e) {
e.printStackTrace();
return false;
}
PropertyDescriptor[] m_Properties = bi.getPropertyDescriptors();
Method getter = null;
for (int i = 0; i < m_Properties.length; i++) {
if (m_Properties[i].getDisplayName().equals(mem)) {
getter = m_Properties[i].getReadMethod();
break;
}
}
if (getter != null) {
try {
return getter.invoke(obj, (Object[]) null);
} catch (Exception e) {
System.err.println("Exception in invoking setter: "+e.getMessage());
return null;
}
} else {
System.err.println("Getter method for " + mem + " not found!");
return null;
}
}
/**
* Try to set an object member to a given value.
* Returns true if successful, else false. The types are adapted as generally as possible,

14
src/eva2/server/go/individuals/AbstractEAIndividual.java
View File

@@ -567,7 +567,7 @@ public abstract class AbstractEAIndividual implements IndividualInterface, java.
}
/**
* Returns true, if the first fitness vector dominates the second one in every component.
* Returns true if the first fitness vector dominates the second one in every component.
* Symmetric case: if the vectors are equal, true is returned.
*
* @param fit1 first fitness vector to look at
@@ -585,7 +585,7 @@ public abstract class AbstractEAIndividual implements IndividualInterface, java.
}
/**
* Returns true, if the first fitness vector truly dominates the second one in every component.
* Returns true if the first fitness vector truly dominates the second one in every component.
*
* @param fit1 first fitness vector to look at
* @param fit2 second fitness vector to look at
@@ -605,7 +605,7 @@ public abstract class AbstractEAIndividual implements IndividualInterface, java.
* Note this is dominance! If the individuals are not comparable this method will
* return false!
* @param indy The individual to compare to.
* @return True if better false else
* @return True if the own fitness dominates the other indy's
*/
public boolean isDominatingDebConstraints(AbstractEAIndividual indy) {
double[] tmpFitness = indy.getFitness();
@@ -956,10 +956,18 @@ public abstract class AbstractEAIndividual implements IndividualInterface, java.
return isDominatingFitnessNotEqual(m_Fitness, otherFitness);
}
/**
* @return true if the own fitness dominates otherFitness.
* @see #isDominatingFitness(double[], double[])
*/
public boolean isDominant(double[] otherFitness) {
return isDominatingFitness(m_Fitness, otherFitness);
}
/**
* @return true if the individual dominates the given one.
* @see #isDominatingDebConstraints(AbstractEAIndividual)
*/
public boolean isDominant(IndividualInterface indy) {
return isDominatingDebConstraints((AbstractEAIndividual)indy);
}

52
src/eva2/server/go/individuals/AbstractEAIndividualComparator.java
View File

@@ -7,11 +7,15 @@ import java.util.Comparator;
* The default version calls isDominatingDebConstraints() of the AbstractEAIndividual
* class and assigns -1 if first is dominant, 1 if second is dominant, 0 if the two ind.s
* are not comparable.
* By default, the dominance criterion is used on fitness vectors.
* If a specific criterion i is set, the comparison is based only on the i-th
* entry of the fitness vector.
* As an alternative, a data key String may be set which is then used to request a data
* object from the individuals. The objects are interpreted as fitness vectors (double[]) and
* the comparison is based on those. This may be used to access alternative (e.g. older or
* best-so-far fitness values) for individual comparison.
*
* @see #AbstractEAIndividual().isDominatingFitness(double[], double[])
* @author mkron
*
*/
@@ -19,6 +23,7 @@ public class AbstractEAIndividualComparator implements Comparator<Object> {
// flag whether a data field should be used.
String indyDataKey = null;
int fitCriterion = -1;
/**
* Comparator implementation which compares two individuals based on their fitness.
@@ -28,7 +33,7 @@ public class AbstractEAIndividualComparator implements Comparator<Object> {
*
*/
public AbstractEAIndividualComparator() {
indyDataKey = null;
this(null, -1);
}
/**
@@ -41,9 +46,32 @@ public class AbstractEAIndividualComparator implements Comparator<Object> {
* @param indyDataKey
*/
public AbstractEAIndividualComparator(String indyDataKey) {
this.indyDataKey = indyDataKey;
this(indyDataKey, -1);
}
/**
* Constructor for a specific fitness criterion in the multiobjective case.
* For comparison, only the given fitness criterion is used if it is >= 0.
*
* @param fitnessCriterion
*/
public AbstractEAIndividualComparator(int fitnessCriterion) {
this(null, fitnessCriterion);
}
/**
* Generic constructor.
*
* @see #AbstractEAIndividualComparator(int)
* @see #AbstractEAIndividualComparator(String)
* @param indyDataKey
* @param fitnessCriterion
*/
public AbstractEAIndividualComparator(String indDataKey, int fitnessCriterion) {
indyDataKey = indDataKey;
fitCriterion = fitnessCriterion;
}
public AbstractEAIndividualComparator(AbstractEAIndividualComparator other) {
indyDataKey = other.indyDataKey;
}
@@ -58,7 +86,7 @@ public class AbstractEAIndividualComparator implements Comparator<Object> {
*
* @param o1 the first AbstractEAIndividual to compare
* @param o2 the second AbstractEAIndividual to compare
* @return -1 if the first is dominant, 1 if the second is dominant, else 0
* @return -1 if the first is dominant, 1 if the second is dominant, otherwise 0
*/
public int compare(Object o1, Object o2) {
boolean o1domO2, o2domO1;
@@ -66,11 +94,21 @@ public class AbstractEAIndividualComparator implements Comparator<Object> {
if (indyDataKey != null) {
double[] fit1 = (double[])((AbstractEAIndividual)o1).getData(indyDataKey);
double[] fit2 = (double[])((AbstractEAIndividual)o2).getData(indyDataKey);
o1domO2 = AbstractEAIndividual.isDominatingFitness(fit1, fit2);
o2domO1 = AbstractEAIndividual.isDominatingFitness(fit2, fit1);
if (fitCriterion < 0) {
o1domO2 = AbstractEAIndividual.isDominatingFitness(fit1, fit2);
o2domO1 = AbstractEAIndividual.isDominatingFitness(fit2, fit1);
} else {
if (fit1[fitCriterion] == fit2[fitCriterion]) return 0;
else return (fit1[fitCriterion] < fit2[fitCriterion]) ? -1 : 1;
}
} else {
o1domO2 = ((AbstractEAIndividual) o1).isDominatingDebConstraints((AbstractEAIndividual) o2);
o2domO1 = ((AbstractEAIndividual) o2).isDominatingDebConstraints((AbstractEAIndividual) o1);
if (fitCriterion < 0) {
o1domO2 = ((AbstractEAIndividual) o1).isDominatingDebConstraints((AbstractEAIndividual) o2);
o2domO1 = ((AbstractEAIndividual) o2).isDominatingDebConstraints((AbstractEAIndividual) o1);
} else {
if (((AbstractEAIndividual) o1).getFitness()[fitCriterion] == ((AbstractEAIndividual) o2).getFitness()[fitCriterion]) return 0;
return (((AbstractEAIndividual) o1).getFitness()[fitCriterion] < ((AbstractEAIndividual) o2).getFitness()[fitCriterion]) ? -1 : 1;
}
}
if (o1domO2 ^ o2domO1) return (o1domO2 ? -1 : 1);
else return 0; // these are not comparable

2
src/eva2/server/go/mocco/MOCCOProblemInitialization.java
View File

@@ -58,7 +58,7 @@ public class MOCCOProblemInitialization extends MOCCOPhase implements InterfaceP
Class[] altern = null;
try {
altern = ReflectPackage.getAssignableClassesInPackage("eva2.server.oa.go.OptimizationProblems", Class.forName("eva2.server.oa.go.OptimizationProblems.InterfaceMultiObjectiveDeNovoProblem"), true, true);
altern = ReflectPackage.getAssignableClassesInPackage("eva2.server.go.problems", Class.forName("eva2.server.go.problems.InterfaceMultiObjectiveDeNovoProblem"), true, true);
} catch (ClassNotFoundException e) {
// TODO Auto-generated catch block
e.printStackTrace();

232
src/eva2/server/go/mocco/paretofrontviewer/MOCCOViewer.java
View File

@@ -321,122 +321,6 @@ public class MOCCOViewer extends JPanel implements InterfaceRefSolutionListener,
}
};
// private void updateHistory() {
// this.m_History.removeAll();
// this.m_History.setLayout(new GridBagLayout());
// this.m_JLIternations = new JLabel[this.m_MOCCO.m_State.m_PopulationHistory.length];
// this.m_JLIterSize = new JLabel[this.m_MOCCO.m_State.m_PopulationHistory.length];
// this.m_JCUse = new JCheckBox[this.m_MOCCO.m_State.m_PopulationHistory.length];
// this.m_JCShow = new JCheckBox[this.m_MOCCO.m_State.m_PopulationHistory.length];
// this.m_JColor = new JComboBox[this.m_MOCCO.m_State.m_PopulationHistory.length];
// GridBagConstraints gbc = new GridBagConstraints();
// gbc.anchor = GridBagConstraints.WEST;
// gbc.fill = GridBagConstraints.BOTH;
// gbc.gridx = 0;
// gbc.gridy = 0;
// gbc.weightx = 1;
// this.m_History.add(new JLabel("Iteration"), gbc);
// gbc.gridx = 1;
// gbc.gridy = 0;
// this.m_History.add(new JLabel("Size"), gbc);
// gbc.gridx = 2;
// gbc.gridy = 0;
// this.m_History.add(new JLabel("Use"), gbc);
// gbc.gridx = 3;
// gbc.gridy = 0;
// this.m_History.add(new JLabel("Show"), gbc);
// gbc.gridx = 4;
// gbc.gridy = 0;
// this.m_History.add(new JLabel("Color"), gbc);
// for (int i = 0; i < this.m_MOCCO.m_State.m_PopulationHistory.length; i++) {
// gbc.gridx = 0;
// gbc.gridy = i+1;
// this.m_History.add(new JLabel((i+0)+"."), gbc);
// gbc.gridx = 1;
// gbc.gridy = i+1;
// this.m_History.add(new JLabel(this.m_MOCCO.m_State.m_PopulationHistory[i].size()+""), gbc);
// gbc.gridx = 2;
// gbc.gridy = i+1;
// this.m_JCUse[i] = new JCheckBox();
// this.m_JCUse[i].setSelected(this.m_MOCCO.m_State.m_Use[i]);
// this.m_JCUse[i].addActionListener(useModeChanged);
// this.m_History.add(this.m_JCUse[i], gbc);
// gbc.gridx = 3;
// gbc.gridy = i+1;
// this.m_JCShow[i] = new JCheckBox();
// this.m_JCShow[i].setSelected(this.m_MOCCO.m_State.m_Show[i]);
// this.m_JCShow[i].addActionListener(showModeChanged);
// this.m_History.add(this.m_JCShow[i], gbc);
// gbc.gridx = 4;
// gbc.gridy = i+1;
// this.m_JColor[i] = this.getComboBox(i);
// this.m_JColor[i].addActionListener(colorModeChanged);
// this.m_History.add(this.m_JColor[i], gbc);
// }
// }
//
// ActionListener useModeChanged = new ActionListener() {
// public void actionPerformed(ActionEvent event) {
// for (int i = 0; i < m_JCUse.length; i++) {
// m_MOCCO.m_State.m_Use[i] = m_JCUse[i].isSelected();
// }
// }
// };
// ActionListener showModeChanged = new ActionListener() {
// public void actionPerformed(ActionEvent event) {
// for (int i = 0; i < m_JCShow.length; i++) {
// m_MOCCO.m_State.m_Show[i] = m_JCShow[i].isSelected();
// }
// m_View.updateView();
// }
// };
// ActionListener colorModeChanged = new ActionListener() {
// public void actionPerformed(ActionEvent event) {
// for (int i = 0; i < m_JColor.length; i++) {
// m_MOCCO.m_State.m_Color[i] = getColor(m_JColor[i].getSelectedIndex());
// }
// m_View.updateView();
// }
// };
//
// private JComboBox getComboBox(int index) {
// String[] colors = {"RED", "BLUE", "GREEN", "CYAN", "MAGENTA", "ORANGE"};
// JComboBox result;
// result = new JComboBox(colors);
// int color = 0;
// if (this.m_MOCCO.m_State.m_Color[index] == Color.RED) color = 0;
// if (this.m_MOCCO.m_State.m_Color[index] == Color.BLUE) color = 1;
// if (this.m_MOCCO.m_State.m_Color[index] == Color.GREEN) color = 2;
// if (this.m_MOCCO.m_State.m_Color[index] == Color.CYAN) color = 3;
// if (this.m_MOCCO.m_State.m_Color[index] == Color.MAGENTA) color = 4;
// if (this.m_MOCCO.m_State.m_Color[index] == Color.ORANGE) color = 5;
// result.setSelectedIndex(color);
// return result;
// }
// public Color getColor(int i) {
// switch (i) {
// case 0 : {
// return Color.RED;
// }
// case 1 : {
// return Color.BLUE;
// }
// case 2 : {
// return Color.GREEN;
// }
// case 3 : {
// return Color.CYAN;
// }
// case 4 : {
// return Color.MAGENTA;
// }
// case 5 : {
// return Color.ORANGE;
// }
// }
// return Color.BLACK;
// }
public void problemChanged(boolean t) {
this.m_MOCCO.m_State.makeFitnessCache(t);
//this.updateHistory();
@@ -614,4 +498,120 @@ public class MOCCOViewer extends JPanel implements InterfaceRefSolutionListener,
this.m_View.updateView();
}
// private void updateHistory() {
// this.m_History.removeAll();
// this.m_History.setLayout(new GridBagLayout());
// this.m_JLIternations = new JLabel[this.m_MOCCO.m_State.m_PopulationHistory.length];
// this.m_JLIterSize = new JLabel[this.m_MOCCO.m_State.m_PopulationHistory.length];
// this.m_JCUse = new JCheckBox[this.m_MOCCO.m_State.m_PopulationHistory.length];
// this.m_JCShow = new JCheckBox[this.m_MOCCO.m_State.m_PopulationHistory.length];
// this.m_JColor = new JComboBox[this.m_MOCCO.m_State.m_PopulationHistory.length];
// GridBagConstraints gbc = new GridBagConstraints();
// gbc.anchor = GridBagConstraints.WEST;
// gbc.fill = GridBagConstraints.BOTH;
// gbc.gridx = 0;
// gbc.gridy = 0;
// gbc.weightx = 1;
// this.m_History.add(new JLabel("Iteration"), gbc);
// gbc.gridx = 1;
// gbc.gridy = 0;
// this.m_History.add(new JLabel("Size"), gbc);
// gbc.gridx = 2;
// gbc.gridy = 0;
// this.m_History.add(new JLabel("Use"), gbc);
// gbc.gridx = 3;
// gbc.gridy = 0;
// this.m_History.add(new JLabel("Show"), gbc);
// gbc.gridx = 4;
// gbc.gridy = 0;
// this.m_History.add(new JLabel("Color"), gbc);
// for (int i = 0; i < this.m_MOCCO.m_State.m_PopulationHistory.length; i++) {
// gbc.gridx = 0;
// gbc.gridy = i+1;
// this.m_History.add(new JLabel((i+0)+"."), gbc);
// gbc.gridx = 1;
// gbc.gridy = i+1;
// this.m_History.add(new JLabel(this.m_MOCCO.m_State.m_PopulationHistory[i].size()+""), gbc);
// gbc.gridx = 2;
// gbc.gridy = i+1;
// this.m_JCUse[i] = new JCheckBox();
// this.m_JCUse[i].setSelected(this.m_MOCCO.m_State.m_Use[i]);
// this.m_JCUse[i].addActionListener(useModeChanged);
// this.m_History.add(this.m_JCUse[i], gbc);
// gbc.gridx = 3;
// gbc.gridy = i+1;
// this.m_JCShow[i] = new JCheckBox();
// this.m_JCShow[i].setSelected(this.m_MOCCO.m_State.m_Show[i]);
// this.m_JCShow[i].addActionListener(showModeChanged);
// this.m_History.add(this.m_JCShow[i], gbc);
// gbc.gridx = 4;
// gbc.gridy = i+1;
// this.m_JColor[i] = this.getComboBox(i);
// this.m_JColor[i].addActionListener(colorModeChanged);
// this.m_History.add(this.m_JColor[i], gbc);
// }
// }
//
// ActionListener useModeChanged = new ActionListener() {
// public void actionPerformed(ActionEvent event) {
// for (int i = 0; i < m_JCUse.length; i++) {
// m_MOCCO.m_State.m_Use[i] = m_JCUse[i].isSelected();
// }
// }
// };
// ActionListener showModeChanged = new ActionListener() {
// public void actionPerformed(ActionEvent event) {
// for (int i = 0; i < m_JCShow.length; i++) {
// m_MOCCO.m_State.m_Show[i] = m_JCShow[i].isSelected();
// }
// m_View.updateView();
// }
// };
// ActionListener colorModeChanged = new ActionListener() {
// public void actionPerformed(ActionEvent event) {
// for (int i = 0; i < m_JColor.length; i++) {
// m_MOCCO.m_State.m_Color[i] = getColor(m_JColor[i].getSelectedIndex());
// }
// m_View.updateView();
// }
// };
//
// private JComboBox getComboBox(int index) {
// String[] colors = {"RED", "BLUE", "GREEN", "CYAN", "MAGENTA", "ORANGE"};
// JComboBox result;
// result = new JComboBox(colors);
// int color = 0;
// if (this.m_MOCCO.m_State.m_Color[index] == Color.RED) color = 0;
// if (this.m_MOCCO.m_State.m_Color[index] == Color.BLUE) color = 1;
// if (this.m_MOCCO.m_State.m_Color[index] == Color.GREEN) color = 2;
// if (this.m_MOCCO.m_State.m_Color[index] == Color.CYAN) color = 3;
// if (this.m_MOCCO.m_State.m_Color[index] == Color.MAGENTA) color = 4;
// if (this.m_MOCCO.m_State.m_Color[index] == Color.ORANGE) color = 5;
// result.setSelectedIndex(color);
// return result;
// }
// public Color getColor(int i) {
// switch (i) {
// case 0 : {
// return Color.RED;
// }
// case 1 : {
// return Color.BLUE;
// }
// case 2 : {
// return Color.GREEN;
// }
// case 3 : {
// return Color.CYAN;
// }
// case 4 : {
// return Color.MAGENTA;
// }
// case 5 : {
// return Color.ORANGE;
// }
// }
// return Color.BLACK;
// }
}

6
src/eva2/server/go/operators/postprocess/PostProcess.java
View File

@@ -612,7 +612,7 @@ public class PostProcess {
prob.evaluate(subPop);
stepsPerf += subPop.size();
subPop.add(candidates.getEAIndividual(i));
subPop.add(candidates.getEAIndividual(i).clone());
nmPops.add(subPop);
}
@@ -636,9 +636,11 @@ public class PostProcess {
case cmaES: stepsPerf += PostProcess.processWithCMA(subPop, prob, term, subPop.size()-1);
break;
}
// if (TRACE) System.out.println("*** after: " + subPop.getBestEAIndividual().getStringRepresentation());
if (checkRange(subPop.getBestEAIndividual())) {
// and replace corresponding individual (should usually be better)
if (subPop.getBestEAIndividual().isDominant(candidates.getEAIndividual(i))) {
// if (subPop.getBestEAIndividual().isDominant(candidates.getEAIndividual(i))) { // TODO Multiobjective???
if (subPop.getBestEAIndividual().getFitness(0)<candidates.getEAIndividual(i).getFitness(0)) {
// System.out.println("moved by "+ PhenotypeMetric.dist(candidates.getEAIndividual(i), subPop.getBestEAIndividual()));
subPop.getBestEAIndividual().putData("PostProcessingMovedBy", new Double(PhenotypeMetric.dist(candidates.getEAIndividual(i), subPop.getBestEAIndividual())));
candidates.set(i, subPop.getBestEAIndividual());

373
src/eva2/server/go/problems/AbstractMultiObjectiveOptimizationProblem.java
View File

@@ -1,10 +1,17 @@
package eva2.server.go.problems;
import javax.swing.*;
import java.awt.Color;
import java.util.ArrayList;
import javax.swing.JFrame;
import wsi.ra.chart2d.DPoint;
import eva2.gui.Chart2DDPointIconCircle;
import eva2.gui.Chart2DDPointIconText;
import eva2.gui.GraphPointSet;
import eva2.gui.Plot;
import eva2.server.go.individuals.AbstractEAIndividual;
import eva2.server.go.individuals.ESIndividualDoubleData;
import eva2.server.go.operators.archiving.ArchivingAllDominating;
import eva2.server.go.operators.archiving.ArchivingNSGA;
import eva2.server.go.operators.moso.InterfaceMOSOConverter;
@@ -13,12 +20,6 @@ import eva2.server.go.operators.paretofrontmetrics.InterfaceParetoFrontMetric;
import eva2.server.go.operators.paretofrontmetrics.MetricS;
import eva2.server.go.populations.Population;
import java.util.ArrayList;
import java.awt.*;
import wsi.ra.chart2d.DPointIcon;
import wsi.ra.chart2d.DPoint;
/**
* Created by IntelliJ IDEA.
* User: streiche
@@ -33,15 +34,46 @@ public abstract class AbstractMultiObjectiveOptimizationProblem extends Abstract
transient protected Population m_ParetoFront = new Population();
public ArrayList m_AreaConst4Parallelization = new ArrayList();
protected int m_OutputDimension = 2;
double m_borderLow = 0;
double m_borderHigh = 5;
transient protected double[][] m_Border;
transient protected Plot m_Plot;
transient protected JFrame m_Result;
private transient boolean m_Show = false;
/**
* TODO
public AbstractMultiObjectiveOptimizationProblem(double borderHigh) {
super();
m_borderHigh=borderHigh;
this.m_Template = new ESIndividualDoubleData();
initBorder();
if (this.m_Show) this.initProblemFrame();
}
public AbstractMultiObjectiveOptimizationProblem() {
this(5.);
}
/** This method allows you to toggle pareto-front visualisation on and off.
* @param b True if the pareto-front is to be shown.
*/
public double[][] m_Border;
transient protected eva2.gui.Plot m_Plot;
transient protected JFrame m_Result;
transient boolean m_Show = false;
public void setShowParetoFront(boolean b) {
this.m_Show = b;
if (this.m_Show) this.initProblemFrame();
else if (this.m_Plot != null) {
this.m_Plot.dispose();
this.m_Plot = null;
}
}
public boolean isShowParetoFront() {
return this.m_Show;
}
public String showParetoFrontTipText() {
return "Toggles the pareto-front visualisation.";
}
/** This method returns a deep clone of the problem.
* @return the clone
*/
@@ -53,30 +85,38 @@ public abstract class AbstractMultiObjectiveOptimizationProblem extends Abstract
* problem frame (i'll provide a default implementation here.
*/
public void initProblem() {
this.m_Border = new double[2][2];
for (int i = 0; i < this.m_Border.length; i++) {
this.m_Border[i][0] = 0;
this.m_Border[i][1] = 5;
}
initBorder();
this.m_ParetoFront = new Population();
if (this.m_Show) this.initProblemFrame();
}
protected void initBorder() {
initBorder(m_borderLow, m_borderHigh);
}
protected void initBorder(double lower, double upper) {
if (this.m_Border == null) this.m_Border = new double[2][2];
for (int i = 0; i < this.m_Border.length; i++) {
this.m_Border[i][0] = lower;
this.m_Border[i][1] = upper;
}
}
/** This method checks whether the problem has truely evaluated
* to a multiobjective problem
* @return true if all individuals are multiobjective
*/
public boolean isPopulationMultiObjective(Population pop) {
public static boolean isPopulationMultiObjective(Population pop) {
if (pop == null) return false;
if (pop.size() == 0) return false;
int best = pop.getBestFitness().length, tmp;
int bestFitLen = pop.getBestFitness().length, tmpFitLen;
for (int i = 0; i < pop.size(); i++) {
tmp = ((AbstractEAIndividual)pop.get(i)).getFitness().length;
if (tmp <= 1) {
tmpFitLen = ((AbstractEAIndividual)pop.get(i)).getFitness().length;
if (tmpFitLen <= 1) {
//System.out.println("There is a single objective individual in the population!");
return false;
}
if (tmp != best) {
if (tmpFitLen != bestFitLen) {
//System.out.println("Not all individual have equal length fitness!");
return false;
}
@@ -129,7 +169,11 @@ public abstract class AbstractMultiObjectiveOptimizationProblem extends Abstract
evaluatePopulationEnd(population); // refactored by MK
}
public void evaluatePopulationStart(Population population) {
if (this.m_Show && (this.m_Plot==null)) this.initProblemFrame();
}
public void evaluatePopulationEnd(Population population) {
// So what is the problem:
// on the one hand i want to log the pareto-front in the
@@ -140,35 +184,48 @@ public abstract class AbstractMultiObjectiveOptimizationProblem extends Abstract
// could be pretty many
// currently the problem should be multi-criteria
// log the pareto-front
if (this.isPopulationMultiObjective(population)) {
if (this.m_ParetoFront == null) this.m_ParetoFront = new Population();
if (this.m_ParetoFront.getArchive() == null) {
Population archive = new Population();
archive.setPopulationSize(100);
this.m_ParetoFront.SetArchive(archive);
}
this.m_ParetoFront.addPopulation((Population) population.getClone());
ArchivingNSGA archiving = new ArchivingNSGA();
archiving.addElementsToArchive(this.m_ParetoFront);
this.m_ParetoFront = this.m_ParetoFront.getArchive();
}
logPopToParetoFront(m_ParetoFront, population);
// Sometimes you want to transform a multiobjective optimization problem
// into a single objective one, this way single objective optimization
// algorithms can be applied more easily
this.m_MOSOConverter.convertMultiObjective2SingleObjective(population);
if (this.m_Show) this.drawProblem(population);
if (this.m_Show) {
if (m_Plot.isValid()) AbstractMultiObjectiveOptimizationProblem.drawProblem(population, m_Plot, this);
}
}
/**
* Unite the given population with the given pareto front and replace pFront with the new pareto front.
* This variant uses ArchivingNSGA
* @param pFront
* @param pop
*/
public static void logPopToParetoFront(Population pFront, Population pop) {
// log the pareto-front
if (AbstractMultiObjectiveOptimizationProblem.isPopulationMultiObjective(pop)) {
if (pFront == null) System.err.println("Error, give at least an empty population as initial pareto front");
if (pFront.getArchive() == null) {
pFront.SetArchive(new Population(100));
}
Population tmpPop=new Population(pop.size());
tmpPop.addPopulation(pop);
tmpPop.addPopulation(pFront);
ArchivingNSGA archiving = new ArchivingNSGA();
archiving.addElementsToArchive(tmpPop);
pFront.clear();
pFront.addPopulation(tmpPop.getArchive());
}
}
/** This method will init the problem specific visualisation of the problem
*/
public void initProblemFrame() {
double[] tmpD = new double[2];
tmpD[0] = 0;
tmpD[1] = 0;
if (this.m_Plot == null) m_Plot = new eva2.gui.Plot("Multiobjective Optimization", "Y1", "Y2", tmpD, tmpD);
if (this.m_Plot == null) m_Plot = new Plot("Multiobjective Optimization", "Y1", "Y2", tmpD, tmpD);
// plot init stuff
this.initAdditionalData(this.m_Plot, 10);
@@ -178,17 +235,56 @@ public abstract class AbstractMultiObjectiveOptimizationProblem extends Abstract
* @param plot The plot where you can draw your stuff.
* @param index The first index where you can draw your stuff
*/
public void initAdditionalData(eva2.gui.Plot plot, int index) {
public void initAdditionalData(Plot plot, int index) {
// for example plot the current population
plot.clearGraph(index);
plot.setUnconnectedPoint(0, 0, index);
}
/** This method will draw the current state of the optimization process
* @param pFront The current population
*/
public static void drawProblem(Population pFront, Population archive, Plot plot) {
ArchivingAllDominating tmpArch = new ArchivingAllDominating();
// Population tmpPop = null;
// i want to plot the pareto front for MOEA and other strategies
// but i have to differentiate between the case where
// there is a true MOEA at work and where the
// MOOpt was converted into a SOOpt
if (pFront != null && (plot != null)) {
// i got either a multiobjective population or a multiobjective local population
plot.clearAll();
tmpArch.plotParetoFront(pFront, plot);
if (archive != null) {
GraphPointSet mySet = new GraphPointSet(10, plot.getFunctionArea());
DPoint myPoint;
Chart2DDPointIconCircle icon;
double[] tmpD;
mySet.setConnectedMode(false);
for (int i = 0; i < archive.size(); i++) {
icon = new Chart2DDPointIconCircle();
tmpD = ((AbstractEAIndividual)archive.get(i)).getFitness();
myPoint = new DPoint(tmpD[0], tmpD[1]);
if (((AbstractEAIndividual)archive.get(i)).getConstraintViolation() > 0) {
icon.setBorderColor(Color.RED);
icon.setFillColor(Color.RED);
} else {
icon.setBorderColor(Color.BLACK);
icon.setFillColor(Color.BLACK);
}
myPoint.setIcon(icon);
mySet.addDPoint(myPoint);
}
}
}
}
/**
* This method will draw the current state of the optimization process
* @param p The current population
*/
public void drawProblem(Population p) {
public static void drawProblem(Population p, Plot plot, AbstractMultiObjectiveOptimizationProblem moProblem) {
ArchivingAllDominating tmpArch = new ArchivingAllDominating();
Population tmpPop = null;
@@ -198,9 +294,9 @@ public abstract class AbstractMultiObjectiveOptimizationProblem extends Abstract
// but i have to differentiate between the case where
// there is a true MOEA at work and where the
// MOOpt was converted into a SOOpt
if (this.isPopulationMultiObjective(p)) {
if (AbstractMultiObjectiveOptimizationProblem.isPopulationMultiObjective(p)) {
// in this case i have to use my local archive
tmpPop = this.m_ParetoFront;
tmpPop = moProblem.m_ParetoFront;
} else {
// in this case i use the population of the optimizer
// and eventually the pop.archive if there is one
@@ -210,54 +306,79 @@ public abstract class AbstractMultiObjectiveOptimizationProblem extends Abstract
tmpArch.addElementsToArchive(tmpPop);
tmpPop = tmpPop.getArchive();
}
if (tmpPop != null) {
// i got either a multiobjective population or a multiobjective local population
this.m_Plot.clearAll();
tmpArch.plotParetoFront(tmpPop, this.m_Plot);
if ((true) && (p.getArchive() != null)) {
GraphPointSet mySet = new GraphPointSet(10, this.m_Plot.getFunctionArea());
DPoint myPoint;
Chart2DDPointIconCircle icon;
double[] tmpD;
mySet.setConnectedMode(false);
tmpPop = p.getArchive();
for (int i = 0; i < tmpPop.size(); i++) {
icon = new Chart2DDPointIconCircle();
tmpD = ((AbstractEAIndividual)tmpPop.get(i)).getFitness();
myPoint = new DPoint(tmpD[0], tmpD[1]);
if (((AbstractEAIndividual)tmpPop.get(i)).getConstraintViolation() > 0) {
icon.setBorderColor(Color.RED);
icon.setFillColor(Color.RED);
} else {
icon.setBorderColor(Color.BLACK);
icon.setFillColor(Color.BLACK);
}
myPoint.setIcon(icon);
mySet.addDPoint(myPoint);
}
}
} else {
// // in this case i got a single objective optimization problem
// if (this.m_Plot != null) this.m_Plot.dispose();
// if (this.m_Result == null) {
// this.m_Result = new JFrame();
// this.m_Result.addWindowListener(new WindowAdapter() {
// public void windowClosing(WindowEvent ev) {
// System.gc();
// }
// });
// }
}
if (tmpPop != null) drawProblem(tmpPop, p.getArchive(), plot);
// else : in this case i got a single objective optimization problem
// draw additional data
this.drawAdditionalData(this.m_Plot, p, 10);
moProblem.drawAdditionalData(plot, p, 10);
}
}
// /** This method will draw the current state of the optimization process
// * @param p The current population
// */
// public static void drawProblem(Population p, Plot plot, AbstractMultiObjectiveOptimizationProblem moProblem) {
// ArchivingAllDominating tmpArch = new ArchivingAllDominating();
// Population tmpPop = null;
//
// if (p.getGeneration() > 2) {
//// m_Plot = new eva2.gui.Plot("Multiobjective Optimization", "Y1", "Y2");
// // i want to plot the pareto front for MOEA and other strategies
// // but i have to differentiate between the case where
// // there is a true MOEA at work and where the
// // MOOpt was converted into a SOOpt
// if (AbstractMultiObjectiveOptimizationProblem.isPopulationMultiObjective(p)) {
// // in this case i have to use my local archive
// tmpPop = moProblem.m_ParetoFront;
// } else {
// // in this case i use the population of the optimizer
// // and eventually the pop.archive if there is one
// tmpPop = new Population();
// tmpPop.addPopulation(p);
// if (p.getArchive() != null) tmpPop.addPopulation(p.getArchive());
// tmpArch.addElementsToArchive(tmpPop);
// tmpPop = tmpPop.getArchive();
// }
// if (tmpPop != null) {
// // i got either a multiobjective population or a multiobjective local population
// plot.clearAll();
// tmpArch.plotParetoFront(tmpPop, plot);
// if ((true) && (p.getArchive() != null)) {
// GraphPointSet mySet = new GraphPointSet(10, plot.getFunctionArea());
// DPoint myPoint;
// Chart2DDPointIconCircle icon;
// double[] tmpD;
// mySet.setConnectedMode(false);
// tmpPop = p.getArchive();
// for (int i = 0; i < tmpPop.size(); i++) {
// icon = new Chart2DDPointIconCircle();
// tmpD = ((AbstractEAIndividual)tmpPop.get(i)).getFitness();
// myPoint = new DPoint(tmpD[0], tmpD[1]);
// if (((AbstractEAIndividual)tmpPop.get(i)).getConstraintViolation() > 0) {
// icon.setBorderColor(Color.RED);
// icon.setFillColor(Color.RED);
// } else {
// icon.setBorderColor(Color.BLACK);
// icon.setFillColor(Color.BLACK);
// }
// myPoint.setIcon(icon);
// mySet.addDPoint(myPoint);
// }
// }
// } else {
//// // in this case i got a single objective optimization problem
// }
// // draw additional data
// moProblem.drawAdditionalData(plot, p, 10);
// }
// }
/** This method will plot a reference solutions or something like it
/**
* This method will plot a reference solutions or something like it
* @param plot The plot where you can draw your stuff.
* @param index The first index where you can draw your stuff
*/
public void drawAdditionalData(eva2.gui.Plot plot, Population pop, int index) {
public void drawAdditionalData(Plot plot, Population pop, int index) {
double[] tmpFitness;
// for example plot the current population
plot.clearGraph(index);
@@ -270,6 +391,86 @@ public abstract class AbstractMultiObjectiveOptimizationProblem extends Abstract
plot.setUnconnectedPoint(this.m_Border[0][0], this.m_Border[1][0], index);
}
/**
* Plot the given population to a MO-plot with red dots for constraint violations and blue otherwise.
* The given border is used to set plot limits but may be null.
*
* @param plot
* @param pop
* @param border
* @param index
*/
public static void drawWithConstraints(Plot plot, Population pop, double[][] border, int index) {
// for example plot the current population
double[][] trueFitness, moFitness;
double[] constraint;
GraphPointSet mySet = new GraphPointSet(index, plot.getFunctionArea());
DPoint myPoint;
double tmp1;
trueFitness = new double[pop.size()][];
constraint = new double[pop.size()];
if (((AbstractEAIndividual)pop.get(0)).hasData("MOFitness")) {
moFitness = new double[pop.size()][];
} else moFitness = null;
for (int i = 0; i < pop.size(); i++) {
constraint[i] = ((AbstractEAIndividual)pop.get(i)).getConstraintViolation();
trueFitness[i] = ((AbstractEAIndividual)pop.get(i)).getFitness();
if (moFitness != null) moFitness[i] = (double[])((AbstractEAIndividual)pop.get(i)).getData("MOFitness");
}
mySet.setConnectedMode(false);
for (int i = 0; i < trueFitness.length; i++) {
if (moFitness != null) {
// moso is active
if (checkValidAt(moFitness, i)) {
myPoint = new DPoint(moFitness[i][0], moFitness[i][1]);
tmp1 = Math.round(trueFitness[i][0] *100)/100.0;
addPoint(constraint, mySet, myPoint, i, ""+tmp1);
}
} else {
// no moso is active
if (checkValidAt(trueFitness, i)) {
myPoint = new DPoint(trueFitness[i][0], trueFitness[i][1]);
addPoint(constraint, mySet, myPoint, i, "");
}
}
}
if (border!= null) {
plot.setUnconnectedPoint(border[0][1], border[1][1], index+1);
plot.setUnconnectedPoint(border[0][0], border[1][0], index+1);
}
}
/**
* Check whether the given fitness array contains valid double values at position i.
*
* @param fitArray
* @param i
* @return
*/
private static boolean checkValidAt(double[][] fitArray, int i) {
return !(new Double(fitArray[i][0]).isNaN()) && !(new Double(fitArray[i][1]).isNaN()) &&
!(new Double(fitArray[i][0]).isInfinite()) && !(new Double(fitArray[i][1]).isInfinite());
}
private static void addPoint(double[] constraint, GraphPointSet mySet,
DPoint myPoint, int i, String text) {
Chart2DDPointIconCircle icon;
Chart2DDPointIconText tmp = new Chart2DDPointIconText(text);
icon = new Chart2DDPointIconCircle();
if (constraint[i] > 0) {
icon.setBorderColor(Color.RED);
icon.setFillColor(Color.RED);
} else {
icon.setBorderColor(Color.BLUE);
icon.setFillColor(Color.BLUE);
}
tmp.setIcon(icon);
myPoint.setIcon(tmp);
mySet.addDPoint(myPoint);
}
/** This method returns a double value that will be displayed in a fitness
* plot. A fitness that is to be minimized with a global min of zero
* would be best, since log y can be used. But the value can depend on the problem.

4
src/eva2/server/go/problems/AbstractOptimizationProblem.java
View File

@@ -203,6 +203,10 @@ public abstract class AbstractOptimizationProblem implements InterfaceOptimizati
return m_Template;
}
public String individualTemplateTipText() {
return "Choose the individual representation to use.";
}
/**
* This method extracts the individuals from a given population that are assumed to correspond to local or global optima.
* Similar individuals are clustered together with a density based clustering method

2
src/eva2/server/go/problems/AbstractProblemDouble.java
View File

@@ -144,7 +144,7 @@ public abstract class AbstractProblemDouble extends AbstractOptimizationProblem
/**
* Get the lower bound of the double range in the given dimension. Override
* this to implement non-symmetric ranges. User setDefaultRange for symmetric ranges.
* this to implement non-symmetric ranges. Use setDefaultRange for symmetric ranges.
*
* @see makeRange()
* @see getRangeUpperBound(int dim)

162
src/eva2/server/go/problems/TF1Problem.java
View File

@@ -1,24 +1,24 @@
package eva2.server.go.problems;
import wsi.ra.chart2d.DPoint;
import java.util.ArrayList;
import java.awt.Color;
import java.io.BufferedWriter;
import java.io.FileWriter;
import java.util.ArrayList;
import wsi.ra.chart2d.DPoint;
import wsi.ra.math.RNG;
import eva2.gui.Chart2DDPointIconCircle;
import eva2.gui.Chart2DDPointIconText;
import eva2.gui.GraphPointSet;
import eva2.gui.Plot;
import eva2.gui.PropertyFilePath;
import eva2.server.go.individuals.AbstractEAIndividual;
import eva2.server.go.individuals.ESIndividualDoubleData;
import eva2.server.go.individuals.InterfaceDataTypeDouble;
import eva2.server.go.operators.constraint.InterfaceConstraint;
import eva2.server.go.operators.moso.InterfaceMOSOConverter;
import eva2.server.go.operators.paretofrontmetrics.InterfaceParetoFrontMetric;
import eva2.server.go.populations.Population;
import eva2.server.go.strategies.InterfaceOptimizer;
import wsi.ra.math.RNG;
/**
* Created by IntelliJ IDEA.
@@ -36,18 +36,16 @@ public class TF1Problem extends AbstractMultiObjectiveOptimizationProblem implem
protected double m_YOffSet = 0.0;
protected boolean m_ApplyConstraints = false;
transient private GraphPointSet mySet;
// transient private GraphPointSet mySet;
public TF1Problem() {
this.m_Template = new ESIndividualDoubleData();
this.m_Border = new double[2][2];
for (int i = 0; i < this.m_Border.length; i++) {
this.m_Border[i][0] = 0;
this.m_Border[i][1] = 1;
}
if (this.m_Show) this.initProblemFrame();
super(1.);
}
public TF1Problem(double borderHigh) {
super(borderHigh);
}
public TF1Problem(TF1Problem b) {
//AbstractOptimizationProblem
if (b.m_Template != null)
@@ -88,18 +86,6 @@ public class TF1Problem extends AbstractMultiObjectiveOptimizationProblem implem
return (Object) new TF1Problem(this);
}
/** This method inits the Problem to log multiruns
*/
public void initProblem() {
this.m_Border = new double[2][2];
for (int i = 0; i < this.m_Border.length; i++) {
this.m_Border[i][0] = 0;
this.m_Border[i][1] = 1;
}
this.m_ParetoFront = new Population();
if (this.m_Show) this.initProblemFrame();
}
/** This method inits a given population
* @param population The populations that is to be inited
*/
@@ -107,11 +93,7 @@ public class TF1Problem extends AbstractMultiObjectiveOptimizationProblem implem
AbstractEAIndividual tmpIndy;
this.m_ParetoFront = new Population();
double[][] newRange = new double[this.m_ProblemDimension][2];
for (int i = 0; i < this.m_ProblemDimension; i++) {
newRange[i][0] = 0;
newRange[i][1] = 1;
}
double[][] newRange = makeRange();
population.clear();
@@ -125,8 +107,20 @@ public class TF1Problem extends AbstractMultiObjectiveOptimizationProblem implem
// population init must be last
// it set's fitcalls and generation to zero
population.init();
if (this.m_Show) this.initProblemFrame();
}
protected double[][] makeRange() {
return makeRange(0, 1);
}
protected double[][] makeRange(double lower, double upper) {
double[][] newRange = new double[this.m_ProblemDimension][2];
for (int i = 0; i < this.m_ProblemDimension; i++) {
newRange[i][0] = lower;
newRange[i][1] = upper;
}
return newRange;
}
/** This method evaluate a single individual and sets the fitness values
* @param individual The individual that is to be evalutated
@@ -194,89 +188,9 @@ public class TF1Problem extends AbstractMultiObjectiveOptimizationProblem implem
return result;
}
/** This method will plot a reference solutions or something like it
* @param plot The plot where you can draw your stuff.
* @param index The first index where you can draw your stuff
*/
public void drawAdditionalData(eva2.gui.Plot plot, Population pop, int index) {
double[] tmpFitness;
// for example plot the current population
//plot.clearGraph(index);
if (true) {
double[][] trueFitness, moFitness;
double[] constraint;
if (mySet == null) mySet = new GraphPointSet(index, this.m_Plot.getFunctionArea());
mySet.removeAllPoints();
mySet = new GraphPointSet(index, this.m_Plot.getFunctionArea());
DPoint myPoint;
double tmp1, tmp2;
Chart2DDPointIconText tmp;
Chart2DDPointIconCircle icon;
trueFitness = new double[pop.size()][];
constraint = new double[pop.size()];
if (((AbstractEAIndividual)pop.get(0)).getData("MOFitness") != null) {
moFitness = new double[pop.size()][];
}
else moFitness = null;
for (int i = 0; i < pop.size(); i++) {
constraint[i] = ((AbstractEAIndividual)pop.get(i)).getConstraintViolation();
trueFitness[i] = ((AbstractEAIndividual)pop.get(i)).getFitness();
if (moFitness != null) moFitness[i] = (double[])((AbstractEAIndividual)pop.get(i)).getData("MOFitness");
}
mySet.setConnectedMode(false);
for (int i = 0; i < trueFitness.length; i++) {
if (moFitness != null) {
// moso is active
if (!(new Double(moFitness[i][0]).isNaN()) && !(new Double(moFitness[i][1]).isNaN()) &&
!(new Double(moFitness[i][0]).isInfinite()) && !(new Double(moFitness[i][1]).isInfinite())) {
myPoint = new DPoint(moFitness[i][0], moFitness[i][1]);
tmp1 = Math.round(trueFitness[i][0] *100)/100.0;
tmp = new Chart2DDPointIconText(""+tmp1);
icon = new Chart2DDPointIconCircle();
if (constraint[i] > 0) {
icon.setBorderColor(Color.RED);
icon.setFillColor(Color.RED);
} else {
icon.setBorderColor(Color.BLACK);
icon.setFillColor(Color.BLACK);
}
tmp.setIcon(icon);
myPoint.setIcon(tmp);
mySet.addDPoint(myPoint);
}
} else {
// no moso is active
if (!(new Double(trueFitness[i][0]).isNaN()) && !(new Double(trueFitness[i][1]).isNaN()) &&
!(new Double(trueFitness[i][0]).isInfinite()) && !(new Double(trueFitness[i][1]).isInfinite())) {
myPoint = new DPoint(trueFitness[i][0], trueFitness[i][1]);
tmp = new Chart2DDPointIconText("");
icon = new Chart2DDPointIconCircle();
if (constraint[i] > 0) {
icon.setBorderColor(Color.RED);
icon.setFillColor(Color.RED);
} else {
icon.setBorderColor(Color.BLACK);
icon.setFillColor(Color.BLACK);
}
tmp.setIcon(icon);
myPoint.setIcon(tmp);
mySet.addDPoint(myPoint);
}
}
}
}
// for (int i = 0; i < pop.size(); i++) {
// tmpFitness = ((AbstractEAIndividual)pop.get(i)).getFitness();
// if (tmpFitness.length <= 1) tmpFitness = (double[])((AbstractEAIndividual)pop.get(i)).getData("MOFitness");
// if (!(new Double(tmpFitness[0]).isNaN()) && !(new Double(tmpFitness[1]).isNaN()) &&
// !(new Double(tmpFitness[0]).isInfinite()) && !(new Double(tmpFitness[1]).isInfinite()))
// plot.setUnconnectedPoint(tmpFitness[0], tmpFitness[1], index);
// else System.out.println("Index "+i+":("+tmpFitness[0]+"/"+tmpFitness[1]+")");
// }
// System.out.println(".");
plot.setUnconnectedPoint(this.m_Border[0][1], this.m_Border[1][1], index+1);
plot.setUnconnectedPoint(this.m_Border[0][0], this.m_Border[1][0], index+1);
}
public void drawAdditionalData(Plot plot, Population pop, int index) {
AbstractMultiObjectiveOptimizationProblem.drawWithConstraints(plot, pop, m_Border, index);
}
/** This method returns a string describing the optimization problem.
* @param opt The Optimizer that is used or had been used.
@@ -426,24 +340,6 @@ public class TF1Problem extends AbstractMultiObjectiveOptimizationProblem implem
return (InterfaceDataTypeDouble)this.m_Template;
}
/** This method allows you to toggle pareto-front visualisation on and off.
* @param b True if the pareto-front is to be shown.
*/
public void setShowParetoFront(boolean b) {
this.m_Show = b;
if (this.m_Show) this.initProblemFrame();
else if (this.m_Plot != null) {
this.m_Plot.dispose();
this.m_Plot = null;
}
}
public boolean getShowParetoFront() {
return this.m_Show;
}
public String showParetoFrontTipText() {
return "Toggles the pareto-front visualisation.";
}
/** This method allows you to set a Multiobjective to
* Singleobjective converter if you choose to.
* @param b The new MO2SO converter.

45
src/eva2/server/stat/AbstractStatistics.java
View File

@@ -46,9 +46,10 @@ public abstract class AbstractStatistics implements InterfaceTextListener, Inter
protected double[] currentBestFit;
protected double[] meanFitness;
protected double[] currentWorstFit;
protected double[] meanBestOfRunFitness;
protected double avgPopDist;
protected double maxPopDist;
protected IndividualInterface bestCurrentIndividual, bestIndividualAllover;
protected IndividualInterface bestCurrentIndividual, bestRunIndividual, bestIndividualAllover;
private ArrayList<InterfaceTextListener> textListeners;
@@ -118,11 +119,13 @@ public abstract class AbstractStatistics implements InterfaceTextListener, Inter
convergenceCnt = 0;
if (saveParams) m_StatsParams.saveInstance();
initOutput(infoString);
bestCurrentIndividual = null;
bestIndividualAllover = null;
meanBestOfRunFitness = null;
if (refineMultiRuns) meanCollection = new ArrayList<double[][]>();
else meanCollection = null;
}
bestCurrentIndividual = null;
bestRunIndividual = null;
runIterCnt = 0;
if (printRunIntroVerbosity()) printToTextListener("\n****** Multirun "+runNumber);
if (params != null) {
@@ -150,13 +153,21 @@ public abstract class AbstractStatistics implements InterfaceTextListener, Inter
if (Mathematics.norm(bestCurrentIndividual.getFitness()) < this.m_StatsParams.getConvergenceRateThreshold()) {
convergenceCnt++;
}
if (printRunStoppedVerbosity()) printToTextListener(" Run best individual : " + BeanInspector.toString(bestCurrentIndividual) + "\n");
if (printRunStoppedVerbosity()) printToTextListener(" run solution data : " + AbstractEAIndividual.getDefaultDataString(bestCurrentIndividual) + "\n");
if (printRunStoppedVerbosity()) printToTextListener(" run solution fit : " + BeanInspector.toString(bestCurrentIndividual.getFitness()) + "\n");
}
if (currentBestFit!= null) {
if (printRunStoppedVerbosity()) printToTextListener(" Best Fitness: " + BeanInspector.toString(currentBestFit) + "\n");
if (printRunStoppedVerbosity()) printToTextListener(" Last best individual : " + BeanInspector.toString(bestCurrentIndividual) + "\n");
if (printRunStoppedVerbosity()) printToTextListener(" Last solution data : " + AbstractEAIndividual.getDefaultDataString(bestCurrentIndividual) + "\n");
if (printRunStoppedVerbosity()) printToTextListener(" Last solution fit : " + BeanInspector.toString(bestCurrentIndividual.getFitness()) + "\n");
}
if (bestRunIndividual != null) {
if (printRunStoppedVerbosity()) printToTextListener(" Run best individual : " + BeanInspector.toString(bestRunIndividual) + "\n");
if (printRunStoppedVerbosity()) printToTextListener(" Run solution data : " + AbstractEAIndividual.getDefaultDataString(bestRunIndividual) + "\n");
if (printRunStoppedVerbosity()) printToTextListener(" Run solution fit : " + BeanInspector.toString(bestRunIndividual.getFitness()) + "\n");
if (meanBestOfRunFitness==null) {
meanBestOfRunFitness=bestRunIndividual.getFitness().clone();
} else addMean(meanBestOfRunFitness, bestRunIndividual.getFitness());
}
// if (currentBestFit!= null) {
// if (printRunStoppedVerbosity()) printToTextListener(" Best Fitness: " + BeanInspector.toString(currentBestFit) + "\n");
// }
if (optRunsPerformed == m_StatsParams.getMultiRuns()) finalizeOutput();
}
@@ -171,6 +182,12 @@ public abstract class AbstractStatistics implements InterfaceTextListener, Inter
for (int i=0; i<meanCollection.size(); i++) divideMean(meanCollection.get(i), optRunsPerformed);
if (printFinalVerbosity()) printToTextListener(refineToText(meanCollection, showAvgIntervals));
}
if ((meanBestOfRunFitness!=null) && (optRunsPerformed>1)) {
Mathematics.svDiv((double)optRunsPerformed, meanBestOfRunFitness, meanBestOfRunFitness);
if (printFinalVerbosity()) {
printToTextListener(" Averaged best run fitness: " + BeanInspector.toString(meanBestOfRunFitness)+"\n");
}
}
if (TRACE)
System.out.println("End of run");
if (resultOut != null) {
@@ -327,9 +344,11 @@ public abstract class AbstractStatistics implements InterfaceTextListener, Inter
bestCurrentIndividual = pop.getBestIndividual().getClone();
if ((bestIndividualAllover == null) || (secondIsBetter(bestIndividualAllover, bestCurrentIndividual))) {
bestIndividualAllover = bestCurrentIndividual;
// printToTextListener("new best found!, last was " + BeanInspector.toString(bestIndivdualAllover) + "\n");
// printToTextListener("new best found!, last was " + BeanInspector.toString(bestIndividualAllover) + "\n");
}
if ((bestRunIndividual==null) || (secondIsBetter(bestRunIndividual, bestCurrentIndividual))) {
bestRunIndividual=bestCurrentIndividual;
}
// IndividualInterface WorstInd = Pop.getWorstIndividual();
if (bestCurrentIndividual == null) {
System.err.println("createNextGenerationPerformed BestInd==null");
@@ -429,7 +448,7 @@ public abstract class AbstractStatistics implements InterfaceTextListener, Inter
if (indy1 == null) return true;
if (indy2 == null) return false;
if (indy1 instanceof AbstractEAIndividual) return ((AbstractEAIndividual)indy2).isDominatingDebConstraints((AbstractEAIndividual)indy1);
return (indy1.isDominant(indy2));
return (indy2.isDominant(indy1));
}
public double[] getBestFitness() {
@@ -440,6 +459,10 @@ public abstract class AbstractStatistics implements InterfaceTextListener, Inter
return bestIndividualAllover;
}
public IndividualInterface getRunBestSolution() {
return bestRunIndividual;
}
public int getFitnessCalls() {
return functionCalls;
}

1
src/eva2/server/stat/InterfaceStatistics.java
View File

@@ -38,6 +38,7 @@ public interface InterfaceStatistics {
public void createNextGenerationPerformed(PopulationInterface Pop, List<InterfaceAdditionalPopulationInformer> informerList);
public void createNextGenerationPerformed(double[] bestfit,double[] worstfit,int calls);
public InterfaceStatisticsParameter getStatisticsParameter(); // called from moduleadapter
public IndividualInterface getRunBestSolution(); // return the best fitness of the last run (may not be equal to the last population)
public IndividualInterface getBestSolution(); // returns the best overall solution
public double[] getBestFitness(); // returns the best overall fitness
}

14
src/eva2/server/stat/StatisticsDummy.java
View File

@@ -17,7 +17,7 @@ import eva2.server.go.problems.InterfaceAdditionalPopulationInformer;
public class StatisticsDummy implements InterfaceStatistics, InterfaceTextListener {
boolean consoleOut = false;
StatsParameter sParams = null;
AbstractEAIndividual bestCurrentIndividual, bestIndividualAllover;
AbstractEAIndividual bestCurrentIndividual, bestRunIndy, bestIndividualAllover;
public StatisticsDummy() {
bestIndividualAllover = null;
@@ -39,6 +39,9 @@ public class StatisticsDummy implements InterfaceStatistics, InterfaceTextListen
public void createNextGenerationPerformed(PopulationInterface pop,
List<InterfaceAdditionalPopulationInformer> informerList) {
bestCurrentIndividual = (AbstractEAIndividual)pop.getBestIndividual();
if ((bestIndividualAllover == null) || (AbstractStatistics.secondIsBetter(bestIndividualAllover, bestCurrentIndividual))) {
bestIndividualAllover = bestCurrentIndividual;
}
if ((bestIndividualAllover == null) || (AbstractStatistics.secondIsBetter(bestIndividualAllover, bestCurrentIndividual))) {
bestIndividualAllover = bestCurrentIndividual;
}
@@ -56,7 +59,11 @@ public class StatisticsDummy implements InterfaceStatistics, InterfaceTextListen
public IndividualInterface getBestSolution() {
return bestIndividualAllover;
}
public IndividualInterface getRunBestSolution() {
return bestRunIndy;
}
public InterfaceStatisticsParameter getStatisticsParameter() {
return sParams;
}
@@ -72,7 +79,8 @@ public class StatisticsDummy implements InterfaceStatistics, InterfaceTextListen
public void startOptPerformed(String InfoString, int runnumber,
Object params) {
bestIndividualAllover = null;
if (runnumber==0) bestIndividualAllover = null;
bestRunIndy = null;
}
public void stopOptPerformed(boolean normal, String stopMessage) {}

2
src/eva2/tools/MultirunRefiner.java
View File

@@ -220,7 +220,7 @@ public class MultirunRefiner {
int numExp = 0, iteration = 0, lineCnt = 0;
ArrayList<double[]> result = new ArrayList<double[]>();
String line;
String runHeader = "Fit.-calls Best Mean Worst Solution";
String runHeader = "Fun.calls Best Mean Worst Solution";
String runFinalizer = " Best solution: ";
boolean readRun = false;