[galib] None

[Liekens, A.M.L.]

Hi again,

Answering your questions would take me a lot of work that has already been
described in several places in the academic literature and on the internet.
The algorithm used in GALib's example number 1, is a simple genetic
algorithm, which is a very common algorithm in the literature. I would
advise you to look up an introductory article or book on the subject in
order to learn more about genetic algorithms and their possible
implementations (of which GALib is a very good one). Maybe Melanie Mitcell's
book "An Introduction to Genetic Algrithms." It may also help to study the
manual of GAlib in order to understand the inner workings of the libraray.

Anthony Liekens,-

-----Original Message-----
From: VINAY VENKATARAGHAVAN
To: galib at mit.edu
Sent: 8/26/2003 6:10 PM
Subject: RE: [galib] None


What I meant when I asked the question as to how is the solution being
obtained is that the objective of the program was to fill a string with
alternating 1's and 0's. 

1. Therefore, how does the evolution process in this
example and case evolve only the fittest  individuals. Since in this
case
I do not see any selection happening. 

2. What is the basis for populating the genomes for subsequent
generations.
Is it as per the genome with the highest score in the current
generation.

3. How is the mutation and combination taking place. Where and when does
it
take place. At which stage of the process in a particular generation.
For
example for a particular generation: are all the individuals first
evaluated and only after that is the mutation and crossover applied? At
what point does mutation and crossover take place? 


Thank you for the previous response. It really helped. 

Vinay


On Fri, 22 Aug 2003, Liekens, A.M.L. wrote:

> 3. In the example in ex1.C how are we evaluating the fitness of the
> individuals in the population belonging to each generation? How
specifically
> is the solution being obtained.
> 
> The first genome that is created in the ex1 main function, is given a
> pointer to the Objective fitness function. As the ga creates more
genomes,
> the objective function is cloned from this first genome. As such, the
> fitness of the individuals can be computed, using this objective
function.
> Evolution now takes care of of evolving individuals with a fitness as
high
> as possible. (I don't really get what you mean with your question "How
> specifically is the solution being obtained.")
> 
> Anthony,-
> 
> 



_______________________________________________
galib mailing list
galib at mit.edu
http://mailman.mit.edu/mailman/listinfo/galib