Open
Description
During the adaptive
mutation, the fitness function is called in
In my application, this is a problem because I have to normalize some optimized probabilities before computing the fitness.
So if I have g1
, g2
, g3
, g4
as genes that are optimized with the GA, and I need sum(g1, g2, g3, g4) = 1
for my fitness function, I most likely get an error.
The way I'm doing things with random
mutation is that I normalize this in my on_mutation
method.
Is there a way of using adaptive
mutation with such restrictions?