However you wish to phrase it, biological fitness is a function that depends on offspring.
many approaches with GAs use the approach of the fittest “parents” get extra children, while the worst parents get no children.
With a moment of reflection, I'm sure you will see that if one were to use the biological notion of fitness here, one would be trying to determine the offspring of a parent in terms of a function that depends on the offspring of the parent. So the program would simply (and obviously) not work.
ECO:”However you wish to phrase it, biological fitness is a function that depends on offspring.”
The interesting thing in this mathematical analysis is that outcomes and second generational success is based on “intelligence”. How is the result determined to be a “good” outcome?
PugetSoundSoldier is forgetting that the programmer/mathematician is making an intelligent determination of what is a successful outcome - there is inherent “intelligence” in the simulation because of this.
This is akin to saying that monkeys/apes can communicate through sign language. Sure, we can create a laboratory experiment using a reward system that can produce a desired outcome. But the bigger question is, without outside “intelligence”, whether these primates would choose to communicate in this manner. In the same way, could these GA simulations randomly result in a good outcome without an intelligent measurement of success and reward.