[galib] Passing a point in to the the algorithm

[Patrick Coquillard]

There's something unclear for me. If a point is of interest (i.e. 
provides a good fitness) it should be conserved at next generation. No 
reason to specify this point since it is the job of the GA to select 
best points...

Would you mean that you whish to pass a point outside of the range of 
the gene as defined in the map ?

My best, Patrick



Le 10/10/2012 12:09, Toufik Al Khawli a écrit :
> Dear Galib Users,
>
> I want to ask you if it is possible to pass an interesting point in to 
> the search algorithm.
>
> I am using the library to search for a global minimum in a 5 
> dimensional system. I am using a loop and adapting the function every 
> time.
> Normally galib is always generating a random points between a 
> predefined minimum and maximum values set by the user. the number of 
> the starting points for example is 100 points.
> Is there any method to pass a point from a previous loop, for example 
> generating 99 points plus this additional interesting point in the 
> second loop?
>
> Thanks a lot
> Greeting
> T. Cauli
>
>
> _______________________________________________
> galib mailing list
> galib at mit.edu
> http://mailman.mit.edu/mailman/listinfo/galib

-- 
-- 
------------------------------------------------------------------------

*
Le savoir n'est pas une marchandise,*

*L'Université n'est pas une entreprise.*

  
------------------------------------------------------------------------

*Dr.Patrick Coquillard,*

*Institut Sophia-Agrobiotech**[INRA-CNRS-Université de Nice]**
***Ecologie Théorique etAppliquée aux Systèmes Semi-Confinés

*400, routes desChappes, P.O. box 167,*

*06 903, SophiaAntipolis*

  

Tél/fax :  *#33(0) 492 386 430*

(from  foreign countries dial 33492... instead of 330492...)

  

http://www.unice.fr/coquillard




-----
Message certifiié sans virus (AVG 2013)
Analyse effectuée par AVG - www.avg.fr
Version: 2013.0.2890 / Base de données virale: 2639/6067 - Date: 29/01/2013
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/galib/attachments/20130130/197d1512/attachment.htm