Added support for NSGA2.

src/core/de.evoal.core.ea/src/main/java/de/evoal/core/ea/main/comparator/ParetoOptimisationValue.java

 package de.evoal.core.ea.main.comparator;
 
 import java.util.Arrays;
+import java.util.Comparator;
+import java.util.function.ToIntFunction;
 
 import de.evoal.core.api.optimisation.OptimisationValue;
+import io.jenetics.ext.moea.ElementComparator;
+import io.jenetics.ext.moea.ElementDistance;
 import io.jenetics.ext.moea.Vec;
 import lombok.NonNull;
 
@@ -50,4 +54,20 @@ public class ParetoOptimisationValue implements OptimisationValue {
 
         return result;
     }
+
+    public static ElementComparator<ParetoOptimisationValue> compare() {
+        return (u, v, i) -> u.fitnessValues.compare(v.fitnessValues, i);
+    }
+
+    public static Comparator<ParetoOptimisationValue> dominance() {
+        return (u, v) -> Vec.dominance(u.fitnessValues.data(), v.fitnessValues.data());
+    }
+
+    public static ElementDistance<ParetoOptimisationValue> distance() {
+        return (u, v, i) -> u.fitnessValues.distance(v.fitnessValues, i);
+    }
+
+    public static ToIntFunction<ParetoOptimisationValue> dimension() {
+        return u -> u.fitnessValues.length();
+    }
 }

src/core/de.evoal.core.ea/src/main/java/de/evoal/core/ea/main/producer/SelectorFactory.java

@@ -3,8 +3,10 @@ package de.evoal.core.ea.main.producer;
 import de.evoal.core.api.board.CoreBlackboardEntries;
 import de.evoal.core.api.cdi.ConfigurationValue;
 import de.evoal.core.api.utils.LanguageHelper;
+import de.evoal.core.ea.main.comparator.ParetoOptimisationValue;
 import de.evoal.languages.model.base.Instance;
 import io.jenetics.*;
+import io.jenetics.ext.moea.NSGA2Selector;
 import lombok.extern.slf4j.Slf4j;
 
 import javax.enterprise.context.ApplicationScoped;
@@ -45,6 +47,8 @@ public class SelectorFactory {
 			case "tournament-selector": return createTournamentSelector(config);
 			case "truncation-selector": return createTruncationSelector(config);
 
+			case "nsga2-selector": return createNSGA2Selector(config);
+
 			/* probability-based selectors */
 			case "boltzmann-selector": return createBoltzmannSelector(config);
 			case "exponential-rank-selector": return createExponentialRankSelector(config);
@@ -64,6 +68,14 @@ public class SelectorFactory {
 		throw new IllegalStateException("Selector '" + name + "' is unknown.");
 	}
 
+	private <G extends Gene<?, G>, C extends Comparable<? super C>> Selector<G, C> createNSGA2Selector(final Instance config) {
+		return (Selector<G, C>) new NSGA2Selector<>(ParetoOptimisationValue.dominance(),
+				ParetoOptimisationValue.compare(),
+				ParetoOptimisationValue.distance(),
+				ParetoOptimisationValue.dimension()
+				);
+	}
+
 	private <G extends Gene<?, G>, C extends Comparable<? super C>> Selector<G,C> createRouletteWheelSelector(final Instance config) {
 		return (Selector<G,C>) new RouletteWheelSelector<>();
 	}

src/core/de.evoal.core.ea/src/main/resources/de/evoal/core/ea/optimisation.dl

@@ -72,6 +72,19 @@ module de.evoal.core.ea.optimisation {
 	
 	abstract type selector {
 	}
+
+    /**
+     * This NSGA2 selector chooses <i>count</i> elements of the population. Therefore, the population is sorted using
+     *   the Crowded-Comparison Operator first. For details, please read the original publication:
+     *
+     * K. Deb, A. Pratap, S. Agarwal, and T. Meyarivan. <i>A fast and elitist multiobjective genetic algorithm:
+     * NSGA-II</i>. Transactions on Evolutionary Computation, 6, 2, 182-197, April 2002. DOI: 10.1109/4235.996017.
+     *
+     * When using this selector, you have to ensure that you are using the ParetoComparator. Otherwise, the selector
+     *  breaks.
+     */
+	type 'nsga2-selector' extends selector {
+	}
 	
 	type 'roulette-wheel-selector' extends selector {
 	}