Comments on: Toward routine billion-variable optimization using genetic algorithms

By: Kumara Sastry » Blog Archives » Towards billion bit optimization via efficient genetic algorithms

Wed, 27 Jun 2007 02:13:06 +0000

[...] Sastry, K., Goldberg, D. E., Llorà, X. (2007). IlliGAL Report No. 2007007. University of Illinois at Urbana-Champaign, Urbana IL. [Full paper - PDF] [Full paper - PS]. [Also see the following paper in the journal complexity]. [...]

By: Kathirvel

Kathirvel — Thu, 08 Feb 2007 12:54:08 +0000

First of all Congrats to all the three researchers. Great Job Kumar.

As a genetic engineer am really excited to see that billion variable handling.
Thanks and regards
KATHIRVEL M.

By: Marcelo

Marcelo — Tue, 06 Feb 2007 19:05:51 +0000

Hello!

Congratulations for all three researchers involved with that over one billion variable problem.

If you permit me, I have some questions about that:

1 – Is the problem decomposable?

2 – Why use binary decision variables which, without noise, can assume only two values/states?

3 – Is the cGA rotationally invariant?

4 – Would be possible to rotate the coordinate system of the variables?

5 – Is that problem some modification of the OneMax?

6 – What kind of random variable that added noise is?

7 – How does that noise increase the problem difficulty?

Thanks!

Marcelo