Re: [cilib-devel] Extraction of Initialisation strategies

[Theuns C. <the...@gm...>]

Hi all,

If I understand it correctly, this could be a good idea. I keep
thinking of a pipelined technique. We can even go as generic as having
an initialisation pipeline for all types of algorithms (whether
singular, population based or multi-population based) and each
specific algorithm (PSO, GA, etc) can then hook into the pipeline and
initialises its own entities (or whatever) by initialising its
positions or velocities or whatever.

Does that make sense?

Regards
--
Theuns

On Fri, Jun 19, 2009 at 11:34 AM, Gary Pampara<gpa...@gm...> wrote:
> Hi all,
>
> I'm looking at extracting the initialisation strategies into something more
> generic. Currently, we have initialisation strategies for the velocity vectors
> in Particles.
>
> The thing is that what about phenotypes within individuals? You may want them
> to be initialised in specific manners (randomized / fixed value etc).
>
> Would anyone regard this as a valid change?
>
> Regards,
> Gary