Proposal to add binary search to the annealing method

[donRumata03]

The problem is that it's so difficult to find the temperature parameter for the annealing optimization method by hand!

So, it might make sense to do this automatically: there should be a particular percent of accepted mutations achieved.
The more the temperature, the more this percent is. So, this function is monotonous by nature (and is in fact if it't smoothed).
It might make sense to use something like n-ary search to find the optimal temperature with the best-fitting percent.