I'm trying to make a visual representation of the traveling salesman problem using a genetic algorithm. I think I got most of it down but I need to slow the repaint() so i can see what's happening each iteration. Any idea how I should implement the actionPerformed() method while retaining the doGA() logic?
class MyPanel extends JPanel implements ActionListener{
private final static int POINTWIDTH = 8;
private final static Color POINTCOLOR = Color.RED;
private final static Color LINECOLOR = Color.GREEN;
private JLabel label = new JLabel();
private City city,city2,city3,city4,city5,city6,city7,city8,city9,city10,city11,city12,city13,city14,city15,city16,city17,city18,city19,city20;
private Population population;
private ArrayList tempCity = new ArrayList<City>();
private ArrayList tempCity2 = new ArrayList<City>();
int delay = 5000;
Timer timer = new Timer(delay, this);
public void start(){
timer.start();
}
public void doGA(){
city = new City(60, 200);
TourManager.addCity(city);
city2 = new City(180, 200);
TourManager.addCity(city2);
city3 = new City(80, 180);
TourManager.addCity(city3);
city4 = new City(140, 180);
TourManager.addCity(city4);
city5 = new City(20, 160);
TourManager.addCity(city5);
city6 = new City(100, 160);
TourManager.addCity(city6);
city7 = new City(200, 160);
TourManager.addCity(city7);
city8 = new City(140, 140);
TourManager.addCity(city8);
city9 = new City(40, 120);
TourManager.addCity(city9);
city10 = new City(100, 120);
TourManager.addCity(city10);
city11 = new City(180, 100);
TourManager.addCity(city11);
city12 = new City(60, 80);
TourManager.addCity(city12);
city13 = new City(120, 80);
TourManager.addCity(city13);
city14 = new City(180, 60);
TourManager.addCity(city14);
city15 = new City(20, 40);
TourManager.addCity(city15);
city16 = new City(100, 40);
TourManager.addCity(city16);
city17 = new City(200, 40);
TourManager.addCity(city17);
city18 = new City(20, 20);
TourManager.addCity(city18);
city19 = new City(60, 20);
TourManager.addCity(city19);
city20 = new City(160, 20);
TourManager.addCity(city20);
population = new Population(50, true);
for (int i = 0; i < 100; i++) {
population = GA.evolvePopulation(population);
repaint();
System.out.println(population.getFittest().getFitness());
}
}
public void paintComponent(Graphics g){
super.paintComponent(g);
Graphics2D g2 = (Graphics2D) g;
g2.setStroke(new BasicStroke(3));
for(int x = 0; x<19;x++){
g2.setColor(LINECOLOR);
//g2.drawLine(tempCity.get(x).getX(), tempCity.get(x).getY(), tempCity2.get(x).getX(), tempCity2.get(x).getY());
g2.drawLine(population.getFittest().getCity(x).getX(), population.getFittest().getCity(x).getY(), population.getFittest().getCity(x+1).getX(), population.getFittest().getCity(x+1).getY());
g2.setColor(POINTCOLOR);
g2.fillOval(population.getFittest().getCity(x).getX() - POINTWIDTH/2, population.getFittest().getCity(x).getY() - POINTWIDTH/2, POINTWIDTH, POINTWIDTH);
}
g2.setColor(LINECOLOR);
g2.drawLine(population.getFittest().getCity(19).getX(), population.getFittest().getCity(19).getY(), population.getFittest().getCity(0).getX(), population.getFittest().getCity(0).getY());
g2.setColor(POINTCOLOR);
g2.fillOval(population.getFittest().getCity(0).getX() - POINTWIDTH/2, population.getFittest().getCity(0).getY() - POINTWIDTH/2, POINTWIDTH, POINTWIDTH);
}
public void actionPerformed(ActionEvent e){
doGA();
}
}
IF your objective is to call :
every seconds, as example, you could use a thread:
You could also use a timer, as someone mentionned in the comments:
Basically, the
Timer
becomes youfor-loop
, so each time it ticks, you need to increment it's iteration and update the state of the model before it's painted.This means you will HAVE to separate the initialisation and update phases of your code, for example...
The short answer would be, no, you can't keep your
doGA
method as it is, it must be changed, there simply is no other way to achieve it, but the longer answer is, it's not that hard if you simply think of theTimer
as a fancy type of loop...