Corewar - the challenge
- From: sayembara@xxxxxxxxx
- Date: 18 Feb 2006 08:13:36 -0800
Hi,
I'm looking at Corewar and the effective application of GA, example:
http://www.infionline.net/~wtnewton/corewar/evol/index.html
http://www.infionline.net/~wtnewton/corewar/evol/evolving.txt
www.koth.org
Despite many attempts to tackle the standard 'hill' (8000 cores), the
GA evolvers still haven't produce a challenging solution aka a champion
warrior. There are (17*7*8*8000*8*8000) possible combo for a line and
the code may go up to 100 lines.
It is a big universe and this should be a good testing ground for GA/GP
application.
What is the best approach to tackle this challenge, knowing current
attempts are 'too slow to achieve results' or simply said, no good
result yet? New and effective method or algorithm discovered /
developed recently? Want to test your solutions?
(Note: Programs proven effective at smaller scale [800 cores] but the
focus now is standard scale [8000 cores]).
/Zul Nadzri
.
- Prev by Date: MSc in Evolutionary and Adaptive Systems, at Sussex
- Next by Date: VRP solution for Eilon 30 nodes problem
- Previous by thread: MSc in Evolutionary and Adaptive Systems, at Sussex
- Next by thread: VRP solution for Eilon 30 nodes problem
- Index(es):
Relevant Pages
|
