Newsgroups: comp.ai.games
Path: cantaloupe.srv.cs.cmu.edu!rochester!cornellcs!newsfeed.cit.cornell.edu!newsstand.cit.cornell.edu!news.kei.com!wang!news
From: bruck@actcom.co.il (Uri Bruck)
Subject: Re: Cheating by AI in games
Reply-To: bruck@actcom.co.il
Organization: ACTCOM - Internet Services in Israel
Date: Wed, 8 Nov 1995 19:13:27 GMT
Message-ID: <DHqo6t.oB@actcom.co.il>
References: <446muf$i0c@nyx.cs.du.edu> <448rul$asd@ixnews4.ix.netcom.com> <tgooding-2609951122550001@helser196.res.iastate.edu> <44bhsu$r2m@ixnews7.ix.netcom.com> <tgooding-2809950021420001@helser196.res.iastate.edu>  <DG3qDw.9q1@actcom.co.il> <458qk8$8un@ <4701n7$5rs@ixnews3.ix.netcom.com> <pbNkxgR.predictor@delphi.com> <472m0t$f4h@ixnews4.ix.netcom.com>
Sender: news@wang.com
Lines: 40

edybs@ix.netcom.com (Eric Dybsand) wrote:

>Will Dwinnell wrote:
>>Sure thing.  Another idea ocurrs to me: dividing the entire
>>battlefield into regions and establishing a set of priorities for
>>entire regions.

>Actually, this is something I'm looking at now.  With the priorities
>being based on the conditions in each region, reviewed by a higher
>level view of the world.  Right now, I can't tell you how its being
>implemented or how well it works, but ask me again in 6 months.
>
This is closer to the higher level descriptions I mentioned a while
back (I've away from the group for a while). One of the advantages of
this is that you don't need to figure every unit's relations to any
other unit in the 'world' but only in the region. The reason this is
less computationally intensive is that the necessary computer power
rises exponentially with the number of units. 

>(Will Dwinnell:)
>>Another thing which ocurrs to me is that since units would maintain
>>their orders until updated, a rotating use of these different systems
>>over time might play the strengths of on off the weaknesses of the
>>others.

Seperate this into types of updates - 
1.keep following current orders - should be performed constantly, most
often this would consist of nothing more than, advance, check for
visible enemies etc.
2. updating - getting new orders - should not be done continously,
start with high priority regions , work down to low priority regions,
until time is up. 
regions that have not been updated at all, get 1 (or other constant)
more priority point for each turn they are not updated, which ensures
that all regions would be updated.
This way on, say 20 'cycles' the high priority region might be updated
19 out of 20 a very low priority region might be updated 1 out of 20.
The 1 out of 20 'missed' update, sould have very little effect
Uri

