Re: [Sheflug] AI (not RAID or kernel upgrading)

[Alex Hudson <nospam [at] york.ac.uk>]

On Fri, 9 Jun 2000, Martin P Holland wrote:
> These are the only type of shrink-wrap games that have a nontrivial AI
> (AFAIK) so what you are saying can be paraphrased as "Games which don't
> have a serious AI component don't use the CPU intensively for AI" which is not
> a big shock ;-)

Not quite ;)) What I was saying that there is a class of games which
relies on multiple highly intelligent AI. Given PC's aren't optimized for
this usage (i.e., you don't get AI cards ;), these games loose out a bit
more. Most games require fewer AIs, and have enough processing power to do
a very good job. 

> This is an extremely interesting AI problem, much more more interesting than

Sokoban should be labelled as an additicting substance, though..

> ones). Currently programs are _much worse_ at solving Sokoban problems than
> humans. So there is some hope that we might actually learn something by
> studying this problem. There is some AI literature on it e.g.
> http://www.cs.ualberta.ca/~games/Sokoban

I'm suprised that brute-force methods don't work well with this type of
problem, although I'm sure there's some subtle reason why ;) I guess since
for n blocks there will be a combinatorial solution space, numbers start
getting big quickly... ;)

Cheers,

Alex.


---------------------------------------------------------------------
Sheffield Linux User's Group - http://www.sheflug.co.uk
To unsubscribe from this list send mail to
- <sheflug-request [at] vuw.ac.nz> - with the word 
 "unsubscribe" in the body of the message. 

  GNU the choice of a complete generation.