Re: [jgap-users] Question about applyMutation

[Klaus M. <jg...@kl...>]

HC,

you dont' have to consider all input parameters in applyMutation. However,
if your gene's architecture allows to consider a parameter, you should do.
Otherwise the evolution might not be as good as it could be. Finally, it
depends on your case, a general advice is not possible.

I haven't looked it up, but I think that setToRandomValue() is only called
internally under normal circumstances whereas applyMutation() is called
regularly.

Best

Klaus
www.klaus-meffert.com

 

> -----Original Message-----
> From: jga...@li... 
> [mailto:jga...@li...] On Behalf 
> Of H. Canterburry
> Sent: Tuesday, February 19, 2008 11:53 PM
> To: jga...@li...
> Subject: [jgap-users] Question about applyMutation
> 
> 
> I am new to GA and have implemented a simple gene that I am 
> testing. I have been looking at the JGAP source code to 
> figure out how to code my genes.
> 
> I am right now left with my applyMutation method. Is it a 
> problem if I simply again randomize the gene's state in this 
> method or do I need to use the index and percent for a 
> "proper" GA implementation? What is the risk of simply 
> randomizing here again? When is applyMutation called compared 
> to setToRandomValue()?
> 
> Thanks
> HC