generative modeling for Rhino
Hi Yi Lu,
nope. What you could do is add penalty clauses to poor performing individuals, thus vastly reducing the chances they have of making it into the next generation. Basically add a little function before your fitness that states:
If(fitness < 400, 0, fitness)
Which means "If the fitness is less than 400, then force it to zero, otherwise leave it intact." I'm assuming here that a fitness of 400 is a bad threshold and zero is the lowest possible fitness. You'll need to substitute your own numbers.
Of course the immediate problem is that it's not known whether an individual has a low or high fitness until after you actually compute it, and at that stage it's already too late to optimize.