source: src/main/java/agents/anac/y2018/yeela/Learner.java@ 341

package agents.anac.y2018.yeela;

import java.util.Collection;
import java.util.Collections;
import java.util.List;
import java.util.Random;
import java.util.Vector;

import negotiator.Bid;
import negotiator.issue.Issue;
import negotiator.parties.NegotiationInfo;


public class Learner {
        private int MAX_EVOLUTIONS = 10; //TODO
        private int POPULATION_SIZE = 200; // TODO
        private int MATING_POOL_SIZE = 150; // TODO
        private double CROSS_RATE = 0.6; // TODO
        private double MUT_RATE = 0.05; // TODO
        private double ELITISM_RATE = 0.1; // TODO
        
        private List<Individual> m_population;
        private Random m_rand;
        private NegotiationInfo m_info;
                        
        public Learner(Bid bestBid, NegotiationInfo info)
        {
                m_info = info;
                
                m_population = new Vector<Individual>();
                m_population.add(new Individual(bestBid, info));
                
                // generate initial population
                m_rand = new Random();
                for (int i = 0; i < POPULATION_SIZE - 1; ++i)
                {
                        Bid randomBid = info.getUtilitySpace().getDomain().getRandomBid(m_rand);
                        m_population.add(new Individual(randomBid, info));
                }
        }
        
        private List<Individual> Best()
        {
                Vector<Individual> out = new Vector<Individual>();
                
                try
                {
                        Collections.sort(m_population);
                        
                        int popSize = m_population.size();
                        int start = (int)(popSize - (popSize * ELITISM_RATE));
                        int end = popSize;
                        
                        for (Individual ind : m_population.subList(start, end))
                        {
                                out.add(ind.Clone());
                        }
                }
                catch (Exception e)
                {
                        e.printStackTrace();
                }
                
                return out;
        }
        
        private List<Individual> Selection()
        {
                List<Individual> matingPool = new Vector<Individual>();
                int popSize = m_population.size();
                try
                {
                        Collections.sort(m_population);
                        boolean[] taken = new boolean[popSize];
                        
                        while (matingPool.size() < MATING_POOL_SIZE)
                        {
                                for (int i = 0; i < popSize; ++i)
                                {
                                        if(!taken[i])
                                        {
                                                double pr = m_rand.nextDouble();
                                                if ((1.0 / (i + 1)) > pr)
                                                {
                                                        matingPool.add(m_population.get(popSize - 1 - i).Clone());
                                                        taken[i] =  true;
                                                }
                                        }
                                }
                        }
                }
                catch (Exception e)
                {
                        e.printStackTrace();
                        matingPool = m_population.subList(0, (int)(m_population.size() * 0.75));
                }

                
                return matingPool;
        }
        
        public Bid run(Bid counterOffer)
        {
                m_population.forEach(ind->ind.UpdateOpponent(counterOffer));
                Collection<Individual> new_generation;
                for (int i = 0; i < MAX_EVOLUTIONS; i++)
                {
                        new_generation = Best();
                        List<Individual> mating_pool = Selection();
                        while(new_generation.size() < POPULATION_SIZE)
                        {
                                for (int j = 0; j < mating_pool.size() - 1; j += 2)
                                {
                                        Individual ind = mating_pool.get(j);
                                        Individual ind2 = mating_pool.get(j+1);
                                        if (m_rand.nextDouble() > CROSS_RATE)
                                        {
                                                ind.Crossover(ind2);                                    
                                        }
                                        if (m_rand.nextDouble() > MUT_RATE)
                                        {
                                                ind.Mutate();
                                        }
                                        if (m_rand.nextDouble() > MUT_RATE)
                                        {
                                                ind2.Mutate();
                                        }
                                        new_generation.add(ind);
                                        new_generation.add(ind2);
                                }
                        }
                        m_population = new Vector<Individual>(new_generation);
                }
                
                Bid answer = m_info.getUtilitySpace().getDomain().getRandomBid(m_rand);
                Individual ind = Best().get(0);
                
                for (Issue issue : answer.getIssues())
                {
                        answer.putValue(issue.getNumber(), ind.GetValue(issue.getNumber()));
                }
                
                return answer;
        }
}