I''m a 4th year degree student working on a solo project for a generic AI system that is both non-engine-specific and tactically intricate. The system will be applied to teams of about 5 entities once completed. The pathfinding information I''ve gathered leads me to believe A* is the most common method. Is there anywhere (or anyone who can) that compares this with the multiple other methods? The criteria to differentiate between methods seems to be number of nodes and their complexity, Time and accuracy (optimality). What I''d like is a clear, simple comparison of as many methods and their variations as possible. This is a huge request, and help will be ridiculously well appreciated. My project is a year long, and hopefully I can come up with something pretty good by the end of it. Anyone who helps can read my conference paper, technical report and logbook if they desire. Give us a hand chums, I need you!!

Try here:

http://www.cpcug.org/user/scifair/Preygel/Preygel.html

Nice one ace! Thanks for the response.

However, I need more aid. The community does not seem to understand the importnace and depth of the project I''m undertaking - I really am aiming to produce something previously unseen in the industry, and as a 4th year Cybernetics student, I may have the skills to complete a challenge of this magnitude. If you''re not excited by this prospect, there''s something wrong with your head.

I have a year to produce a tactically realistic, real-time AI engine for generic use in games. If I get it working GameDev loyalists will be cited as a major influence and aid.

Any and all sites relating to (primarily pathfinding) AI systems and methods will be considered.

GIVE ME INFO!

I tend to find that for the theoretical side text books are generally better than anything you will find on the internet. I would suggest investing in some good AI books.

How about:

http://ai-depot.com/BotNavigation/Path.html

There are about 4 different types of algorithms (and all their varients) discussed.


Artificial Intelligence Depot - Maybe it''s not all about graphics...

Thanks Chaps!

Keep ''em coming - I''m like an Information liposuction machine - sucking the AI Fat from the Belly of the Industry, and Forcing it back into the Lips of my Project!

Sexy!

If you want to make an all-around tactical bot, include some text parsing so it can cooperate with the player =-)
If you could learn maclisp and port the text-''understanding'' part of this program to a generic library, it would be awesome:
http://hci.stanford.edu/~winograd/shrdlu/

"The Requested Information Is Unknown Or Classified" -Anonymous

quote: Original post by HypnoBeard
I have a year to produce a tactically realistic, real-time AI engine for generic use in games.

Given the goal of your project, one year is not likely to be enough. However, you could certainly get a framework going in that time that displays some desired behaviours. You should take a look at the work done by John Laird and his team at MIT, working on Quake Bots. You might also find the work of Emma Norling at University of Melbourne intersting. She''s using the BDI architecture for creating human-like adaptive behaviour in simulated environments. Since her PhD research is connected to the defense force - and hence publishing her results based in their simulated environments is not possible - she''s looking into applying her techniques on Bots for Quake II. She has a couple of papers that you might find interesting.

Good luck,

Timkin