Joined: Oct. 2009
Yes, I see the differences.
One describes actual things that are actually happening in the real world and was developed using a significant amount of evidence from observation and experiment.
The other does not describe real things. It has no evidential support.
And I would argue that the step in the middle "If desired fitness reached, then stop" is wrong. If we knew the desired fitness level then we could design a system to do what we needed.
Genetic algorithms are used when it is very difficult to design a system, but easy to evaluate the results of a system. For example, in optics. It is very easy to do a ray trace on an optical lens system, but it is difficult (nearly impossible) to design a lens system to maximize some things, minimize some things, while keeping others constant (or within an approved range).
Now, the reason, I submit that the "if fitness reached, then stop" is an incorrect statement is because genetic algorithms often produce results that are surprising to researchers and engineers. If those runs were stopped as soon as minimum fitness requirement was reached, then the maximum benefit would not have been reached.
For example, in diesel engine management systems, researchers used genetic algorithms to vary the many components and inputs for a diesel engine in order to improve the efficiency. Let's say the researchers wanted a 5% reduction in soot, a 5% reduction in NOx emmisons, and a 5% increase in fuel efficiency. If the researchers had stopped there, then they would not have found the solutions that resulted in a 50% reduction in NOx, AND a 50% reduction in soot, AND a 10% reduction in fuel consumption.
In my experience, researchers allow a GA to run until it reaches an optimum which is not improved by the GA itself OR they run out of time or money.
BTW: Still have some issues you need to address. Why won't you even talk about these things?
Ignored by those who can't provide evidence for their claims.