Paris Karagiannopoulos 51 2 9. But, you didn't answer my initial question.The method getRouletteWheel returns the selection scheme based on srlection data structure. Also, such a setup encourages the highly fit individuals to be chosen at least once. Then a random se,ection is made similar to how the roulette wheel is rotated. What I mean to say is, I have a fitness function which gives negative values. I'm sure there would be others cases too. This will NOT work for negative fitness numbers, as a negative piece of a pie i. No one talk about replacement of selected item so that selected item didn't get selected again.