source: src/main/java/negotiator/boaframework/opponentmodel/IAMHagglerOpponentConcessionModel.java@ 316

package negotiator.boaframework.opponentmodel;

import java.util.ArrayList;

import agents.Jama.Matrix;
import agents.uk.ac.soton.ecs.gp4j.bmc.BasicPrior;
import agents.uk.ac.soton.ecs.gp4j.bmc.GaussianProcessMixture;
import agents.uk.ac.soton.ecs.gp4j.bmc.GaussianProcessMixturePrediction;
import agents.uk.ac.soton.ecs.gp4j.bmc.GaussianProcessRegressionBMC;
import agents.uk.ac.soton.ecs.gp4j.gp.covariancefunctions.CovarianceFunction;
import agents.uk.ac.soton.ecs.gp4j.gp.covariancefunctions.Matern3CovarianceFunction;
import agents.uk.ac.soton.ecs.gp4j.gp.covariancefunctions.NoiseCovarianceFunction;
import agents.uk.ac.soton.ecs.gp4j.gp.covariancefunctions.SumCovarianceFunction;
import genius.core.utility.AdditiveUtilitySpace;

/**
 * This class is a component of the ANAC 2011 agent IAMHaggler where it estimates
 * the concession rate of the opponent.  For more information on how it works see
 * "Using Gaussian Processes to Optimise Concession in Complex Negotiations against Unknown Opponents"
 * by Colins et al.
 * @author Alex Dirkzwager
 *
 */

public class IAMHagglerOpponentConcessionModel {
        
        private Matrix timeSamples;
        private Matrix means;
        private Matrix variances;
        private GaussianProcessRegressionBMC regression;
        private int lastTimeSlot = -1;
        private ArrayList<Double> opponentTimes = new ArrayList<Double>();
        private ArrayList<Double> opponentUtilities = new ArrayList<Double>();
        private double intercept;
        private double maxUtilityInTimeSlot;
        private Matrix matrixTimeSamplesAdjust;
        private int slots;



        /**
         * 
         * @param numberOfSlots (within the total negotiation
         * @param utilitySpace
         */
        public IAMHagglerOpponentConcessionModel(int numberOfSlots, AdditiveUtilitySpace utilitySpace, int amountOfSamples){
                double discountingFactor = 0.5;
                try
                {
                        discountingFactor = utilitySpace.getDiscountFactor();
                }
                catch(Exception ex)
                {
                        ex.printStackTrace();
                }
                if(discountingFactor == 0)
                        discountingFactor = 1;
                makeTimeSamples(amountOfSamples);
                this.slots = numberOfSlots;
                BasicPrior[] bps = { new BasicPrior(11, 0.252, 0.5),
                                new BasicPrior(11, 0.166, 0.5), new BasicPrior(1, .01, 1.0) };
                CovarianceFunction cf = new SumCovarianceFunction(
                                Matern3CovarianceFunction.getInstance(),
                                NoiseCovarianceFunction.getInstance());
                
                regression = new GaussianProcessRegressionBMC();
                regression.setCovarianceFunction(cf);
                regression.setPriors(bps);
        }
        
        /**
         * updates the model with the most recent opponent bid
         * @param opponentUtility
         * @param time
         */
        public void updateModel(double opponentUtility, double time) {
                // Calculate the current time slot
                int timeSlot = (int) Math.floor(time * slots);
                
                boolean regressionUpdateRequired = false;
                if (lastTimeSlot == -1) {
                        regressionUpdateRequired = true;
                }

                // If the time slot has changed
                if (timeSlot != lastTimeSlot) {
                        if (lastTimeSlot != -1) {
                                // Store the data from the time slot
                                opponentTimes.add((lastTimeSlot + 0.5) / slots);
                                if(opponentUtilities.size() == 0)
                                {
                                        intercept = Math.max(0.5, maxUtilityInTimeSlot);
                                        double[] timeSamplesAdjust = new double[timeSamples.getColumnDimension()];
                                        System.out.println("timeSampleAdjusted[15]: " + timeSamplesAdjust.length);
                                        
                                        
                                        int i = 0;
                                        double gradient = 0.9 - intercept;
                                        for (double d : timeSamples.getRowPackedCopy()) {
                                                timeSamplesAdjust[i++] = intercept + (gradient * d);
                                        }
                                        matrixTimeSamplesAdjust = new Matrix(timeSamplesAdjust, timeSamplesAdjust.length);
                                }
                                opponentUtilities.add(maxUtilityInTimeSlot);
                                // Flag regression receiveMessage required
                                regressionUpdateRequired = true;
                        }
                        // Update the time slot
                        lastTimeSlot = timeSlot;
                        // Reset the max utility
                        maxUtilityInTimeSlot = 0;
                }
                
                // Calculate the maximum utility observed in the current time slot
                maxUtilityInTimeSlot = Math.max(maxUtilityInTimeSlot, opponentUtility);

                if (timeSlot == 0) {
                        return;
                }

                if (regressionUpdateRequired && opponentTimes.size() > 0) {
                        double gradient = 0.9 - intercept;

                        GaussianProcessMixture predictor;

                        if(lastTimeSlot == -1)
                        {
                                predictor = regression.calculateRegression(new double[] {}, new double[] {});
                        }
                        else
                        {
                                double x;
                                double y;
                                try {
                                x = opponentTimes.get(opponentTimes.size() - 1);
                                y = opponentUtilities.get(opponentUtilities.size() - 1);

                                } catch(Exception ex) {
                                        System.out.println("Error getting x or y");
                                        throw new Error(ex);
                                }
                        
                                predictor = regression.updateRegression(
                                                new Matrix(new double[] {x}, 1),
                                                new Matrix(new double[] {y - intercept - (gradient * x)}, 1));
                        }

                        GaussianProcessMixturePrediction prediction = predictor
                                        .calculatePrediction(timeSamples.transpose());

                        // Store the means and variances
                        means = prediction.getMean().plus(matrixTimeSamplesAdjust);
                        
                        variances = prediction.getVariance();
                }
        }
        
        /**
         * Gets all means in a n-by-1 Matrix
         * @return
         */
        public Matrix getMeans(){
                return means;
        }
        
        /**
         * Gets a specific mean point corresponding to the timeSlot
         * @param timeSlot
         * @return
         */
        public double getMeanAt(int timeSlot){
                return means.get(timeSlot, 0);
        }
        
        /**
         * Gets all Variances in a n-by-1 Matrix
         * @return
         */
        public Matrix getVariance(){
                return variances;
        }
        
        /**
         * Gets a specific variance point corresponding to the timeSlot
         * @param timeSlot
         * @return
         */
        public double getVarianceAt(int timeSlot){
                return variances.get(timeSlot, 0);
        }
        
        /**
         * Create a 1-by-n matrix of time samples.
         * 
         * @param n
         *            The sample size.
         */
        private void makeTimeSamples(int n) {
                double[] timeSamplesArray = new double[n + 1];
                {
                        for (int i = 0; i < timeSamplesArray.length; i++) {
                                timeSamplesArray[i] = ((double) i) / ((double) n);
                        }
                }
                System.out.println("timeSampleArray[15]: " + timeSamplesArray.length);
                timeSamples = new Matrix(timeSamplesArray, 1);
        }
}