Genetics

/*
 * To change this template, choose Tools | Templates
 * and open the template in the editor.
 */

/**
 *
 * @author Rawan
 */

import java.util.Vector;
import java.io.File; 
import java.util.Scanner; 
import java.io.FileNotFoundException; 
import org.neuroph.core.NeuralNetwork;

public class Genetics {
   public final double mutationprobability = .1; 
   public final double maximumfitness = .6; 
   public Vector[] Genetic_Algorithm(Vector[] population){
       double populationfitness = 0; //average population fitness
       //repeat until population is fit enough
       do{
       Vector[] newpop = new Vector[TestGenetics.populationsize]; //create a new population
       for(int i=0; i<newpop.length; i++)
           newpop[i] = new Vector<Song>(); 
       double fitnessvalue=0; //the new population's fitness value
       //Song[][] newpop = new Song[TestGenetics.populationsize][TestGenetics.playlistsize];//create a new population
       for(int i=0; i<TestGenetics.populationsize; i++)
       {
           
             Vector<Song> x = randomselection(population); //parent1
             Vector<Song> y = randomselection(population); //parent2
             Vector<Song> child = reproduce(x, y); //returns a child (A Vector of song objects)
             double cmprob = Math.random(); //generate a random child mutation probability
             
             if(cmprob <mutationprobability)
                mutate(child); 
             fitnessvalue += fitnessfunction(child);  
             newpop[i].add(child);
       }
       populationfitness = fitnessvalue/TestGenetics.populationsize; //average fitness (0-1)
       population = newpop; 
       }while(populationfitness<maximumfitness); //repeat until population fitness exceeds maximumfitness
       
       return population; 
   }
   //takes two vectors x and y. Produces a child (child) by concating x and y at a random point
    public Vector<Song> reproduce(Vector<Song> x, Vector<Song> y){
        
        int c = (int) (TestGenetics.playlistsize*Math.random()); //select a random concatination index
        Vector<Song> child = new Vector<Song>(); 
        
        //Substring(x, 1 to c)
        for(int i=0; i<c; i++)
        {
            Song temp = x.elementAt(i); 
            child.add(new Song(temp.getTitle(), temp.getArtist(), temp.getPlaycount()));
        }
        
        //Substring(y, c to size)
         for(int i=c; i<y.size(); i++)
        {
            Song temp = y.elementAt(i); 
            child.add(new Song(temp.getTitle(), temp.getArtist(), temp.getPlaycount()));
        }
       return child; 
    }
    //Select a random Vector of songs from a population and return it
    public Vector<Song> randomselection(Vector[] pop)
    {
        Vector<Song> random = new Vector<Song>(); 
        int index = (int)(TestGenetics.populationsize*Math.random()); //creates a random index from 0-population size
        //copy the song objects inside the randomly selected population
        for(int i=0; i<TestGenetics.playlistsize; i++)
        {
        Song temp = (Song)pop[index].elementAt(i);//gets the song at index i
        random.add(new Song(temp.getTitle(), temp.getArtist(), temp.getPlaycount())) ; //copy the song's values into a new object. Then add the object ot the vector
        
                }
        return random; 
    }
    
    public void mutate(Vector<Song> c){
       Song randsong = null; //the random song to be extracted from the file
       try{
          
         Scanner input = new Scanner(new File("musicdata.txt"));
        int filerandom = (int) (Math.random()*TestGenetics.filesize); //generate a random index within the file
        int count = 0; 
        //until we reach our random index or the end of the file
       while(count<filerandom && input.hasNext()){
         String line = input.nextLine(); 
       }
       //read the line at index random
        String line = input.nextLine(); 
              
             String title="" ; 
             int intindex=0; 
             for(int i=0; i<line.length(); i++)
             {
              if(Character.isDigit(line.charAt(i)))
              {intindex = i;  
              break; 
              }
              title+=line.charAt(i); //extract the title
             }
         randsong = new Song(title, line.charAt(intindex), line.charAt(intindex+2)); //the random song extracted from the file
          
        }catch(FileNotFoundException e){
            System.out.println(e.getMessage()); 
        }
       int mutationindex = (int)(Math.random()*TestGenetics.playlistsize);// random number from 0-playlistsize
       
       c.set(mutationindex, randsong); //replace the song at the mutation index with the random song from the file
       
    }
    //calculates the individual's fitness (0-1)
    //using the NN
    public double fitnessfunction(Vector<Song> x){
        
      double fitness=0;//an average of the NN's output
       NeuralNetwork nn = NeuralNetwork.load("MLP11.nnet"); //initilize the NN
       
      for(Song song: x){
        nn.setInput(song.getArtist(), song.getPlaycount());//sets the NN input to song's (artist id, playcount)
        nn.calculate();
      double[] output = nn.getOutput();
      fitness += output[0]; //adds the output(the degree of loving a track)
      }
      return fitness/x.size(); //returns the average
    }
}