Introduction | Assumptions | Design | Implementation | Results

Here are the results obtained. (Note that you can click these images to see full-size versions.)

The two orange cells represent combinations where not all experiments resulted in an eventual success, so there is likely to be less certainty about their correctness.

Here’s a graph of those results, showing that we’ve identified the mutation rates which are most beneficial to eventual population success.

It is clear that the range of helpful mutation likelihoods is between about 0.2% and 1.5%. Outside of this, mutation has an increasingly harmful effect.

Notice that as the breeding pool size increases in size the sensitivity to mutation outside this region increases much more dramatically. Notice also that there are small local minima either side of the optimal region. I leave it to the reader to investigate further if they wish.