Dear Ladies and Gentleman,
I am doing a small project in Prolog that will eventually solve
multiplication cryptarithms. The point is to use something resembling
human methods - "human-derived heuristics", I call them. I have a
rather bizarre idea to use a genetic algorithm that starts with random
permutations for the letters that participate in the cryptarithm,
calculates for each some heuristic function of the distance from the
right solution and so on, the usual GA way. But I am not so sure this
approach is pertinent to my problem at all. I would be obliged if you
proposed some comments on my predicament.
Sincerely yours,
Felix Goldberg (a newcomer to the list).
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|