One weakness of Galapagos I never managed to overcome

I wanted to share with you a problem with Galapagos I didn't find solution to yet.

Borrowing from David Rutten his elegant analogy to fitness level, I will demonstrate it using topography. In this example, X and Y axes represent two genes of the sheep. The higher the sheep stands, the fitter it is. The problem is that it can't see nor move. Its only way of moving is by giving birth to lambs, which have different sets of X/Y genes. Naturaly, lambs that are born on a high terrain have more chance of surviving (thus having offsprings in the next generation).

Now, what is my problem? I found that on many times the solver get stuck on solution A. If the "valley" is too wide and deep, the "herd" will never cross it towards solution B. This is not just a theoretical problem, it is actually a weakness of evolution wherever you use it.

A good way of overcoming this problem is by increasing the mutations, so a sheep on one side of the valley can give birth to a lamb on the other side of it.

Galapagos, however, doesn't give you direct control over mutation rate (as far as I understand). You can control inbreeding ratios, but no matter what value you use, the mutations get smaller and smaller with the generations. It means that many times, once an initial solution was found, the "sheep" rarely migrate to the other side of the "valley".

I would love to hear your opinion and suggestions.

Gilad.