Page 1 of 1

GA and Random Seed

PostPosted: Tue Dec 13, 2011 3:09 pm
by grpesc
Hi,

first of all, thank you for that great piece of software.

I would like to try out the Genetic Algorithm for my work and I am wondering if there is the possibility to set a random seed in order to reproduce an optimization result (given that the other parameters stay the same).

Thank you for the answer.

Re: GA and Random Seed

PostPosted: Tue Dec 13, 2011 3:40 pm
by andrew.kirillov
Hello,

You can always check documentation and see if there is any API which allows doing that.

From the code I can see that most classes create Random object internally without exposing it. So it means you cannot affect seed much unless you change the code.

Re: GA and Random Seed

PostPosted: Wed Dec 14, 2011 7:25 am
by grpesc
Hello,

thank you for the answer. Is there a chance to provide an optional setter in order to change the random seed in one of the the next versions? I think, this would be interesting for many.

Best,

Re: GA and Random Seed

PostPosted: Wed Dec 14, 2011 9:54 am
by andrew.kirillov
You can always put an enhancement request into issue tracking system. Maybe someday it will be resolved.

You can also do the fix yourself and contribute it ...

Re: GA and Random Seed

PostPosted: Wed Nov 21, 2012 3:28 pm
by Eliza
Dear firend,
I have some doubts:
1. The aim of the symbolic regression problem is to find the best approximation function for data input but I do not know how to interpret the solution of each test (used sample 2), you help me?
2. With the same input data the solution can vary greatly, I think it is the generation of random numbers. truth?

Re: GA and Random Seed

PostPosted: Fri Nov 23, 2012 5:38 pm
by Eliza
Now I know that the solution is in notation polanesa. The solutions can vary greatly being generated with the same restrictions and input data and it is due by the processes of generating random numbers.