Newsgroups: comp.ai
Path: cantaloupe.srv.cs.cmu.edu!rochester!cornell!travelers.mail.cornell.edu!news.kei.com!news.mathworks.com!uunet!in1.uu.net!fdn.fr!univ-lyon1.fr!swidir.switch.ch!news.unige.ch!usenet
From: sylvere@divsun.unige.ch (Silvere Martin-Michiellot)
Subject: Re: Games AI...discussion, help.
Message-ID: <1995Feb28.190312.15902@news.unige.ch>
Sender: usenet@news.unige.ch
Reply-To: sylvere@divsun.unige.ch
Organization: University of Geneva, Switzerland
References: <3ilr6a$sts@ixnews3.ix.netcom.com>
Date: Tue, 28 Feb 1995 19:03:12 GMT
Lines: 34

In article sts@ixnews3.ix.netcom.com, mearrin@ix.netcom.com (Michael Arrington) writes:
>My wife and I are working on various games on the IBM PC.  We are 
>interested in discussing and looking for suggestions/help with specific 
>AI algorithms.  Is this group the proper forum for this or would we be 
>better suited to another.  Any help would be appreciated.
>
>Q:  In a "Doom"-type game (oh no! not another one!) we would like our 
>baddies to chase the player(s) if certain conditions are met.  Our model 
>thus far works on a subsumption architecture (ala Brooks at MIT's Mobot 
>lab.)  Switching into "chase" mode is easy, but are there any 
>suggestions for implementing the game logic for that mode (i.e. making 
>the bad guys smart enough to maneuver through the maze)?
>
>Mike A.

There is a very simple algorithm that might help you.

The idea is that players in the maze leave behing them a smell (or footsteps).
If the monsters detect the smell, they just have to follow it (i.e go to the next
cell of the map following the footstep they are on, and again, and again).

This algorithm is so powerful that if the monster gets on your trail, you can't
escape him whatever you do, sorry boys, :-)

If you use a combination with any other algorithm, and your monsters are dangerous,
be sure that your players will remember your name :-)

-----------------

"Is anyone alive down there ?"

Silvere MARTIN-MICHIELLOT


