public code v1

plugins/afryca.consensusmodel.fen2015/src/afryca/consensusmodel/fen2015/Fen2015.java

@@ -0,0 +1,319 @@
+package afryca.consensusmodel.fen2015;
+
+import java.util.ArrayList;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+
+import afryca.ase.RunnableScript;
+import afryca.consensumodel.addon.E4DIAddon;
+import afryca.consensusmodel.ConsensusModelMulticriteria;
+import afryca.consensusmodel.definition.EResultElements;
+import afryca.decisionmatrix.DecisionMatrix;
+import afryca.domain.fuzzyset.FuzzySet;
+import afryca.structure.Structure;
+import afryca.structure.pair.Pair;
+import afryca.twotuple.TwoTuple;
+
+public class Fen2015 extends ConsensusModelMulticriteria {
+
+	private static final String CONSENSUS_MODEL_NAME = "F. Zou et al. (2015)"; //$NON-NLS-1$
+
+	private static final String ZITA = "zita"; //$NON-NLS-1$ //
+	private static final String LAMBDA = "lambda"; //$NON-NLS-1$
+	private static final String MAX_ROUNDS = "h_max"; //$NON-NLS-1$
+	private static final String ALPHA = "alpha"; //$NON-NLS-1$
+
+	private Integer zita;
+	private Float lambda;
+	private Float alpha;
+	private Integer h_max;
+	private FuzzySet blts;
+
+	private Integer round;
+	private Integer currentConsensus;
+	private DecisionMatrix collective;
+	private ArrayList<Structure> twoTuplePreferencesDM;
+	private Map<Integer, List<Map<Pair<Integer, Integer>, Integer>>> betas;
+
+	private Structure[] preferencesArray;
+	
+	@Override
+	protected void setModelConfiguration() {
+	}
+
+	@Override
+	protected void obtainConfigurationValues() {
+		zita = Math.round((Float) configuration.getValue(ZITA));
+		lambda = (Float) configuration.getValue(LAMBDA);
+		alpha = (Float) configuration.getValue(ALPHA);
+		h_max = (Integer) configuration.getValue(MAX_ROUNDS);
+
+		betas = new HashMap<>();
+		round = 0;
+
+		transformValuationToTwoTuple();
+	}
+
+	@SuppressWarnings("unchecked")
+	private void transformValuationToTwoTuple() {
+		List<Structure> preferencesDM = new ArrayList<Structure>();
+		for (int i = 0; i < numberOfCriteria * numberOfExperts; i++) {
+			preferencesDM.add(preferences[i]);
+		}
+		
+		ArrayList<Structure> unifiedValuations = (ArrayList<Structure>) ((RunnableScript) E4DIAddon.aseService.createExecutionBuilder().setFunction("Unification")
+				.putVariable("preferences", preferencesDM).putVariable("blts", getBLTS(preferencesDM)).execute()).getResult();
+
+		twoTuplePreferencesDM = (ArrayList<Structure>) ((RunnableScript) E4DIAddon.aseService.createExecutionBuilder().setFunction("Disunification").putVariable("preferences", unifiedValuations)
+				.execute()).getResult();
+	}
+
+	private FuzzySet getBLTS(List<Structure> preferencesDM) {
+		return (FuzzySet) E4DIAddon.aseService
+				.createFragmentBuilder()
+				.putVariable("preferences", preferencesDM.toArray()) //$NON-NLS-1$
+				.setCode("unification.getBLTS(preferences)") //$NON-NLS-1$
+				.setOutputType(FuzzySet.class.getName())
+				.eval()
+				.getResult();
+	}
+
+	@Override
+	protected void preFirstSaveRoundResults() {
+		computeDistanceDegrees();
+		Integer currentConsensus = getMinimunDegreeOfSimilarity();
+		preSaveRoundResult(1, preferencesArray, obtainVisualizeValues(), (float) currentConsensus);
+
+		result.put(EResultElements.initial_consensus_degree, currentConsensus);
+		result.put(EResultElements.maxround, h_max);
+		result.put(EResultElements.consensus_threshold, zita);
+		result.put(EResultElements.consensus_model, CONSENSUS_MODEL_NAME);
+	}
+
+
+	@Override
+	protected void preSaveRoundResults() {
+		Integer currentConsensus = getMinimunDegreeOfSimilarity();
+		preSaveRoundResult(round + 1, preferencesArray, obtainVisualizeValues(), (float) currentConsensus);
+	}
+
+	@SuppressWarnings({ "rawtypes", "unchecked" })
+	private void computeDistanceDegrees() {
+		DecisionMatrix dm1, dm2;
+		Integer beta;
+		Map<Pair<Integer, Integer>, Integer> bValue;
+		List<Map<Pair<Integer, Integer>, Integer>> b;
+
+		betas.clear();
+		for (int q = 0; q < numberOfExperts; ++q) {
+			for (int i = 0; i < numberOfCriteria; ++i) {
+				dm1 = (DecisionMatrix) twoTuplePreferencesDM.get(q * numberOfCriteria + i);
+				for (int j = 0; j < numberOfAlternatives; ++j) {
+					beta = 0;
+					for (int k = 0; k < numberOfExperts; ++k) {
+						dm2 = (DecisionMatrix) twoTuplePreferencesDM.get(k * numberOfCriteria + i);
+						beta += computeLambda((TwoTuple) dm1.getValue(0, j), (TwoTuple) dm2.getValue(0, j));
+					}
+
+					if (betas.get(q) != null) {
+						bValue = new HashMap<>();
+						bValue.put(new Pair(i, j), beta);
+						b = betas.get(q);
+						b.add(bValue);
+					} else {
+						bValue = new HashMap<>();
+						bValue.put(new Pair(i, j), beta);
+						b = new ArrayList<>();
+						b.add(bValue);
+						betas.put(q, b);
+					}
+				}
+			}
+		}
+	}
+
+	private Integer computeLambda(TwoTuple valuation1, TwoTuple valuation2) {
+		if (computeSimilarityDegrees(valuation1, valuation2) >= alpha) {
+			return 0;
+		} else {
+			return 1;
+		}
+	}
+
+	private Integer getMinimunDegreeOfSimilarity() {
+		Integer beta, betaMax = -1;
+
+		for (Integer expert : betas.keySet()) {
+			List<Map<Pair<Integer, Integer>, Integer>> bValues = betas.get(expert);
+			for (Map<Pair<Integer, Integer>, Integer> bValue : bValues) {
+				for (Pair<Integer, Integer> pair : bValue.keySet()) {
+					beta = bValue.get(pair);
+					if (beta > zita) {
+						if (beta > betaMax) {
+							betaMax = beta;
+						}
+					}
+				}
+			}
+		}
+
+		if (betaMax == -1) {
+			currentConsensus = zita.intValue();
+			betaMax = zita.intValue();
+		} else {
+			currentConsensus = betaMax;
+		}
+
+		return betaMax;
+	}
+
+	private DecisionMatrix computeAggregation() {
+		
+		
+		Structure[] preferecesAux = new Structure[preferences.length - 1];
+		List<Structure> preferencesDM = new ArrayList<Structure>();
+	
+		for (int i = 0; i < preferecesAux.length; i++) {
+			try {
+				preferecesAux[i] = (Structure) preferences[i].clone();
+				preferencesDM.add((Structure) preferences[i].clone());
+			} catch (CloneNotSupportedException e) {
+				e.printStackTrace();
+			}
+		}
+		
+		blts = getBLTS(preferencesDM);
+		
+		Object[][] collective = (Object[][]) ((RunnableScript) E4DIAddon.aseService.createExecutionBuilder().setFunction("arithmeticAggregationLPR").putVariable("preferences", preferecesAux)
+				.putVariable("alternatives", numberOfAlternatives).putVariable("criteria", numberOfCriteria).putVariable("experts", numberOfExperts).execute()).getResult();
+
+		if ((collective != null)) {
+
+			DecisionMatrix result = new DecisionMatrix(numberOfCriteria, numberOfAlternatives);
+			for (int i = 0; i < numberOfCriteria; i++) {
+				for (int j = 0; j < numberOfAlternatives; j++) {
+					result.setDomain(i, j, blts);
+				}
+			}
+			result.prepareStructureForPreferences(collective);
+			return result;
+		}
+		return new DecisionMatrix(numberOfCriteria, numberOfAlternatives);
+	}
+
+	private Float computeSimilarityDegrees(TwoTuple valuation1, TwoTuple valuation2) {
+		return (float) valuation1.similarityMeasure(valuation2);
+	}
+
+	@Override
+	protected void consensusRound() {
+		computeDistanceDegrees();
+		Integer betaMax = getMinimunDegreeOfSimilarity();
+		if (betaMax > zita && round < h_max) {
+			Map<Integer, List<Pair<Integer, Integer>>> expertsToRecommend = getExpertsToRecommend(betaMax);
+			collective = computeAggregation();
+			
+			Float delta1, delta2;
+			for (Integer expert : expertsToRecommend.keySet()) {
+				List<Pair<Integer, Integer>> pairsToChange = expertsToRecommend.get(expert);
+				for (Pair<Integer, Integer> pairToChange : pairsToChange) {
+					delta1 = (float) ((TwoTuple) twoTuplePreferencesDM.get(expert * numberOfCriteria + pairToChange.getLeft()).getValue(0, pairToChange.getRight())).calculateInverseDelta();
+					delta2 = (float) ((TwoTuple) collective.getValue(pairToChange.getLeft(), pairToChange.getRight())).calculateInverseDelta();
+
+					TwoTuple twoTupleModified = new TwoTuple(blts);
+					twoTupleModified.calculateDelta((float) (Math.pow(delta1, lambda) * Math.pow(delta2, 1 - lambda)));
+
+					((DecisionMatrix) twoTuplePreferencesDM.get(expert * numberOfCriteria + pairToChange.getLeft())).setValue(0, pairToChange.getRight(), twoTupleModified);
+				}
+			}
+			round++;
+		}
+	}
+
+	private Map<Integer, List<Pair<Integer, Integer>>> getExpertsToRecommend(Integer betaMax) {
+		Integer beta;
+
+		Map<Integer, List<Pair<Integer, Integer>>> result = new HashMap<>();
+
+		for (Integer expert : betas.keySet()) {
+			List<Map<Pair<Integer, Integer>, Integer>> bValues = betas.get(expert);
+			for (Map<Pair<Integer, Integer>, Integer> bValue : bValues) {
+				for (Pair<Integer, Integer> pair : bValue.keySet()) {
+					beta = bValue.get(pair);
+					if (beta == betaMax) {
+						if (result.get(expert) == null) {
+							List<Pair<Integer, Integer>> expertP = new ArrayList<>();
+							expertP.add(pair);
+							result.put(expert, expertP);
+						} else {
+							List<Pair<Integer, Integer>> expertP = result.get(expert);
+							expertP.add(pair);
+						}
+					}
+				}
+			}
+		}
+		return result;
+	}
+
+	@Override
+	protected void posSaveRoundResults() {
+		posSaveRoundResult(preferencesArray, obtainVisualizeValues(), (float) currentConsensus, null);
+	}
+
+	@Override
+	protected boolean mustBeCarriedOutAnotherRound() {
+		return (round < h_max) && (currentConsensus > zita);
+	}
+
+	@Override
+	protected void saveExecutionResults() {
+
+		this.configuration.setValue(PREFERENCES, preferencesArray);
+		result.put(EResultElements.number_of_rounds_required, round);
+		result.put(EResultElements.consensus_degree_achieved, (float) currentConsensus);
+	}
+
+	@Override
+	protected void groupPreferencesByCriteria() {
+		computeDistanceDegrees();
+		getMinimunDegreeOfSimilarity();
+		collective = computeAggregation();
+		preferencesArray = new Structure[numberOfExperts+1];
+
+		DecisionMatrix dmExpert, dmByCriterion;
+		for (int k = 0; k < numberOfExperts; ++k) {
+			dmExpert = new DecisionMatrix(numberOfCriteria, numberOfAlternatives);
+			for (int i = 0; i < numberOfCriteria; ++i) {
+				dmByCriterion = (DecisionMatrix) twoTuplePreferencesDM.get(k * numberOfCriteria + i);
+				for (int col = 0; col < numberOfAlternatives; ++col) {
+					dmExpert.setValue(i, col, dmByCriterion.getValue(0, col));
+					dmExpert.setDomain(i, col, dmByCriterion.getDomain(0, col));
+				}
+			}
+			preferencesArray[k]=dmExpert;
+		}
+		
+		preferencesArray[numberOfExperts]=collective;
+		
+		
+
+	}
+
+	@Override
+	protected Float[][][] obtainVisualizeValues() {
+
+		Float[][][] preferencesGroupVisualization=new Float[preferencesArray.length][numberOfCriteria][numberOfAlternatives];
+		
+		for (int k = 0; k < preferencesArray.length; k++) {
+			for (int i = 0; i < numberOfCriteria; i++) {
+				for (int j = 0; j < numberOfAlternatives; j++) {
+					preferencesGroupVisualization[k][i][j] = ((TwoTuple) preferencesArray[k].getValue(i,j)).calculateInverseDelta();
+				}
+			}
+		}
+		return preferencesGroupVisualization;
+	}
+
+}