Really? Explain to me with your own words what shun tried to say
That is just a random combination of words intended to avoid a direct answer……..
I'll give an example from something closer to my expertise as a software engineer: genetic algorithms, which I've used in the past to some extent.
In a GA, one designs the structure of what we call a "chromosome". This is the genotypic description of the phenotypic system that we try to evolve. In a GA, it is the "chromosome" that will mutate. This "chromosome" is then translated into the phenotypic systems which is subjected to a fitness test.
To keep it simple, let's say that the structure of the chromosome is just a fixed string of 10 letters and numbers.
The mutation function, will be a randomized method, which will for example first randomize how many mutations will take place. We'll call that X. Next, it will take X random characters from the string. Next, it will take those characters and replace it by another random character.
This would be the equivalent of a fully randomized mutation process - both in terms of occurance AND in terms of fitness.
Now, we could introduce the equivalent of a strong "mutagen". In the randomized function, we could say that this mutagen will have a 50% chance of changing all characters with the value "A" into the value "B".
This obviously changes the probabilities of which specific mutations will occur, especially when it comes to the "genes" with value "A".
==> This change in probability of
occurance has no impact at all on the probability in terms of
fitness.
Only a fitness test will determine how this mutation affects fitness.
Changing A to B in position 1, might have a completely different effect as opposed to changing A to B in position 7.
It might destroy fitness, it might increase fitness, or it might have no effect whatsover.
See?
Random in terms of
occurance versus random in terms of
fitness.
They are NOT the same thing.