Like crossover,
mutation is a
genetic operator. It randomly flips the bits of a an offspring's genotype. For example,
one mutation of the CA |
 |
can give rise to the CA |
 |
The first produces this from a random initial distribution. |
 |
The second produces this from a random initial distribution. |
 |
|
Why include mutation in the genetic algorithm? |
Mutation prevents the population from homogenizing in a
particular gene. If every proposed solution has a "yes" answer to the first
classifier system question, then crossover will never change this answer. |