Re: [Cilib-users] Calling cilib from existing code

[Edwin P. <ed...@pe...>]

On Monday 19 May 2008 12:42:23 you wrote:

> That's not what I mean. What I'm referring to is the minmove value of a
> particle. What is the default and smallest increment of the parameter space
> that a particle can move from / to?

Ah, okay. That would depend on the position and velocity update strategy as 
well as the PSO parameters being utilised. The StandardVelocityUpdate 
strategy also supports clamping of particle velocities through the VMax 
setting.

> Ikvm in static mode. BTW, how much of a performance overhead do you
> anticipate from this approach? Would you ever implement a .Net version of
> CILib?

Interesting, I have no idea what the performance implications are -- we would 
need to do some tests to determine that. Regarding a .NET version of CILib, I 
doubt we'll be doing that. But, that said, if there is a demand or need for 
such an animal, I'm sure someone will implement it. Time will tell...

> Here it is:
>
> class PSOProblem : OptimisationProblem
>     {
>         public Problem getClone2()
>         {
>             return this.getClone;
>         }
>
>         public OptimisationProblem getClone()
>         {
>             return this.getClone;
>         }
>
>         public Cloneable getClone3()
>         {
>             return this.getClone;
>         }
>
>     }

That does look strange. I don't know if it's an artifact off the Ikvm 
integration or if it's something else. Gary, any ideas?

Regards,
Edwin Peer

-- 
#include http://www.peer.co.za/~espeer/disclaimer.html