We start with the assumption that we have a working simulation, and go through the questions we are trying to address with the simulation.
The main benefit this approach is that it might help in building models by starting with simpler skills used by species earlier in the chain of evolution. We can then move to more complex skills through normal accrual methods of evolution, i.e. software engineering can follow in the footpath of evolution by incrementing the functionality of skills.
Using this kind of model we can focus on changes in the skilled behaviour due to evolution-in-progress. For instance, we now have a lot of the world's knowledge on the internet. We can search for bits of knowledge with Google or other search engines. We now can access this knowledge via our smartphones. Soon we will get access via Google glasses. We conjecture that this will take us past speech and writing to a new level of knowledge processing. We can see a similar progression where we are shedding the limits of mathematics to go to more general data-structures and algorithms.
We can add to the list of questions.
There is very little work on such models and, to my knowledge, no experience in using them for skill engineering.