source: src/main/java/agents/anac/y2018/fullagent/OpponentModel_lgsmi.java@ 341

package agents.anac.y2018.fullagent;

import java.util.HashMap;
import java.util.HashSet;
import java.util.Map;
import java.util.Map.Entry;
import java.util.Set;


import com.sun.xml.internal.bind.api.impl.NameConverter;
import flanagan.math.Matrix;

import negotiator.Bid;
import negotiator.bidding.BidDetails;
import negotiator.boaframework.BOAparameter;
import negotiator.boaframework.NegotiationSession;
import negotiator.boaframework.OpponentModel;
import negotiator.issue.Issue;
import negotiator.issue.IssueDiscrete;
import negotiator.issue.Objective;
import negotiator.issue.ValueDiscrete;


import negotiator.persistent.DefaultStandardInfo;
import negotiator.persistent.StandardInfo;

import negotiator.utility.AdditiveUtilitySpace;
import negotiator.utility.Evaluator;
import negotiator.utility.EvaluatorDiscrete;

/**
 * BOA framework implementation of the HardHeaded Frequecy Model. My main
 * contribution to this model is that I fixed a bug in the mainbranch which
 * resulted in an equal preference of each bid in the ANAC 2011 competition.
 * Effectively, the corrupt model resulted in the offering of a random bid in
 * the ANAC 2011.
 * 
 * Default: learning coef l = 0.2; learnValueAddition v = 1.0
 * 
 * Adapted by Mark Hendrikx to be compatible with the BOA framework.
 * 
 * Tim Baarslag, Koen Hindriks, Mark Hendrikx, Alex Dirkzwager and Catholijn M.
 * Jonker. Decoupling Negotiating Agents to Explore the Space of Negotiation
 * Strategies
 * 
 * 
 */
public class OpponentModel_lgsmi extends OpponentModel {


        // the learning coefficient is the weight that is added each turn to the
        // issue weights
        // which changed. It's a trade-off between concession speed and accuracy.

    /*********** can be reduced over time for giving less importance to later bids *******/
        private double learnCoef;
        // value which is added to a value if it is found. Determines how fast
        // the value weights converge.
    /*********************** can be reduced over time for giving less importance to later bids  *********************/
    private int learnValueAddition;

        private int amountOfIssues;

        /**
         * Initializes the utility space of the opponent such that all value issue
         * weights are equal.
         */
        @Override
        public void init(NegotiationSession negotiationSession, Map<String, Double> parameters) {
                super.init(negotiationSession, parameters);
                this.negotiationSession = negotiationSession;
                if (parameters != null && parameters.get("l") != null) {
                        learnCoef = parameters.get("l");
                } else {
                        learnCoef = 0.2;
                }
                learnValueAddition = 1;
                initializeModel();
        }

        private void initializeModel() {
                opponentUtilitySpace = new AdditiveUtilitySpace(negotiationSession.getDomain());
                amountOfIssues = opponentUtilitySpace.getDomain().getIssues().size();
                double commonWeight = 1D / (double) amountOfIssues;

                // initialize the weights
                for (Entry<Objective, Evaluator> e : opponentUtilitySpace.getEvaluators()) {
                        // set the issue weights
                        opponentUtilitySpace.unlock(e.getKey());
                        e.getValue().setWeight(commonWeight);
                        try {
                                // set all value weights to one (they are normalized when
                                // calculating the utility)
                                for (ValueDiscrete vd : ((IssueDiscrete) e.getKey()).getValues())
                                        ((EvaluatorDiscrete) e.getValue()).setEvaluation(vd, 1);
                        } catch (Exception ex) {
                                ex.printStackTrace();
                        }
                }
        } 

        /**
         * Determines the difference between bids. For each issue, it is determined
         * if the value changed. If this is the case, a 1 is stored in a hashmap for
         * that issue, else a 0.
         * 
         * @param first
         *            bid of the opponent
         * @param second
         *            bid
         * @return
         */
        private HashMap<Integer, Integer> determineDifference(BidDetails first, BidDetails second) {

                HashMap<Integer, Integer> diff = new HashMap<Integer, Integer>();
                try {
                        for (Issue i : opponentUtilitySpace.getDomain().getIssues()) {
                                diff.put(i.getNumber(), (((ValueDiscrete) first.getBid().getValue(i.getNumber()))
                                                .equals((ValueDiscrete) second.getBid().getValue(i.getNumber()))) ? 0 : 1);
                        }
                } catch (Exception ex) {
                        ex.printStackTrace();
                }

                return diff;
        }

        /**
         * Updates the opponent model given a bid.
         */
        @Override
        public void updateModel(Bid opponentBid, double time) {
                if (negotiationSession.getOpponentBidHistory().size() < 2) {
                        return;
                }
                int numberOfUnchanged = 0;
                BidDetails oppBid = negotiationSession.getOpponentBidHistory().getHistory()
                                .get(negotiationSession.getOpponentBidHistory().size() - 1);
                BidDetails prevOppBid = negotiationSession.getOpponentBidHistory().getHistory()
                                .get(negotiationSession.getOpponentBidHistory().size() - 2);
                HashMap<Integer, Integer> lastDiffSet = determineDifference(prevOppBid, oppBid);

                // count the number of changes in value
                for (Integer i : lastDiffSet.keySet()) {
                        if (lastDiffSet.get(i) == 0)
                                numberOfUnchanged++;
                }

                // This is the value to be added to weights of unchanged issues before
                // normalization.
                // Also the value that is taken as the minimum possible weight,
                // (therefore defining the maximum possible also).

        // the proportion given to last bid
        double goldenValue = learnCoef / (double) amountOfIssues;
                // The total sum of weights before normalization.
                double totalSum = 1D + goldenValue * (double) numberOfUnchanged;
                // The maximum possible weight
                double maximumWeight = 1D - ((double) amountOfIssues) * goldenValue / totalSum;

                // re-weighing issues while making sure that the sum remains 1
                for (Integer i : lastDiffSet.keySet()) {

                    //if issue's value unchanged and the weight of the issue is smaller then maximumWeight
                        if (lastDiffSet.get(i) == 0 && opponentUtilitySpace.getWeight(i) < maximumWeight)
                            //if the new weight is legal, set the weight for this issue
                                opponentUtilitySpace.setWeight(opponentUtilitySpace.getDomain().getObjectives().get(i),
                                                (opponentUtilitySpace.getWeight(i) + goldenValue) / totalSum);
                        else
                            // the assumption is that values that have been changed are values that the
                // opponent is willing to compromise on them, so we reduce their weight
                opponentUtilitySpace.setWeight(opponentUtilitySpace.getDomain().getObjectives().get(i),
                                                opponentUtilitySpace.getWeight(i) / totalSum);
                }

                // Then for each issue's value that has been offered last time, a constant

                // value is added to its corresponding ValueDiscrete.
                try {
                        for (Entry<Objective, Evaluator> e : opponentUtilitySpace.getEvaluators()) {
                                // cast issue to discrete and retrieve value. Next, add constant
                                // learnValueAddition to the current preference of the value to

                                // make it more important

                                ((EvaluatorDiscrete) e.getValue()).setEvaluation(
                                                oppBid.getBid().getValue(((IssueDiscrete) e.getKey()).getNumber()),
                                                (learnValueAddition + ((EvaluatorDiscrete) e.getValue()).getEvaluationNotNormalized(
                                                                ((ValueDiscrete) oppBid.getBid().getValue(((IssueDiscrete) e.getKey()).getNumber())))));
                        }
                } catch (Exception ex) {
                        ex.printStackTrace();
                }
        }

        @Override
        public double getBidEvaluation(Bid bid) {
                double result = 0;
                try {
                        result = opponentUtilitySpace.getUtility(bid);
                } catch (Exception e) {
                        e.printStackTrace();
                }
                return result;
        }

        @Override
        public String getName() {

                return "OpponentModel_lgsmi";

        }

        @Override
        public Set<BOAparameter> getParameterSpec() {
                Set<BOAparameter> set = new HashSet<BOAparameter>();
                set.add(new BOAparameter("l", 0.2,
                                "The learning coefficient determines how quickly the issue weights are learned"));
                return set;
        }


    public Map<String, Double> getParameters() {
        Map<String, Double> map = new HashMap<String, Double>();
        //The learning coefficient determines how quickly the issue weights are learned
        map.put("l", 0.2);
        return map;
    }

}