Re: GA Individual Representation

On Fri, 28 Nov 2003 02:27:45 GMT, Ala Qumsieh <[EMAIL PROTECTED]>
wrote:

>
>Hi all,
>
>In almost all papers that I read (as far as I recall at least), 
>individuals in GAs are encoded as bit vectors. Is this restrictive in 
>any way? What other representations exist?
>
>Also, are there any papers or websites that discuss various 
>selection/crossover/mutation algorithms?
>
>Thanks,
>--Ala

There are GAs that encode individuals as real numbers, integers, and
more-- and for some classes of problems it's advantageous not to use
binary strings to represent individuals.  It's possible to solve any
GA-solvable problem using bit strings, but faster sometimes to use a
representation better tailored to the problem at hand.

A downloadable demo GA that works with reals, integers, binary strings
and permutations is available at:
http://www.nli-ltd.com/products/genetic_algorithms/demos.htm

Steve