so I think I’ve got that last problem solved. But there’s a problem in my fitness function, such that if there’s an error in the creation of the file needed for analysis (in this case, caused by evaluating a beast with no genes at all), the fitness is inordinately high. (it gets a fitness of 0, which for this test, is the best possible).
So, evolution being the dirty, cheating fink it is, jumps all over that.
It sorta makes me happy, but it’s still a low-down filthy trick.
14 Apr 2006