Re: GA parameters

Hi!

I would consider/try two options:
1) narrowing the optimization space by prohibiting some values in
genome or lowering search interval.
for example you are searching for a minimum of a function f(x,y,z)
where x,y,z are in range [-1000,1000]... then after some iterations you
limit the search space your best genome has values X,Y,Z and you search
within intervals [X-100,X+100] for x...
Of course you can do this multiple times... and every time you lower
your optimizaion(search) space.
you could also do that you keep searcing in space where you have best N
genomes.

2) reducing mutaion/breeding probabilities is also a good idea. with
this you increase local searches and decrease global search but since
you already did this with big mutation probabilities. you can change
mutation rate each generation... maybe using some function like:
1+5*exp(-k*g); where k is a constant, g - generation number, 1 would be
minimum mutation prob, 6(=1+5) maximum mutation prob. exp(-kg) term
reaches value 0.5 in (ln2/k) generations.

enjoy and make sure you let us posted about your progress.

Rok

.