guesstimation

Right now Im running/going to rerun 20 iterations of GA with mutation rate of .6, .7, and .8. There was only one successful run of these so not much data will have to do to waste. If  these run dont snag on anything I dont really know what to expect from the data. From what I have seen I would think that the .6 would have one of the higher averages and maximum finesses. Of the runs that completed, it looks like around 50% mutation rate is a point where fitness peaks and then falls.

This could be speculation, I have not done as many iterations as would have been appropriate, It would be quite interesting to see that final charts for this experiment. Some papers that I have read said that a large population in relation to mutation leaves a generation of better configuration, so much fascination...( /\ ./\) on a roll...anyway, some papers said that a smaller mutation rate would be good with a high population.. high meaning 16 , but this also meant a smaller chance for finding a good solution. With the parameters like 100 population and generation then it could also be that the mutation rate might level off at some point.