Archived

This topic is now archived and is closed to further replies.

Ketchaval

Uncertainty and AI

Recommended Posts

I know very little about AI, but here is an interesting (if possibly very elementary) question that might be worth considering. Uncertainty and AI. For discussion of the creation of an enemy AI system, which would use facts about its environment / what it perceives, combined with some logic to act in a reasonable manner.. in conflict situations. Ie. It takes precautions relative to the likelihood of risk that it perceives. It (probably) won''t run around a corner with lots of Players round it, if there is a corner that it GUESSES may have less enemies hiding round it. Of course the likelihood of things that the AI perceives is constantly updated, and it will calculate a new plan of action when the situation changes sufficiently. Ie. It peeks round the (supposedly) empty corner, and sees a sleeping Basilik. It then considers what to do, does it run past the players or tiptoe past the potentially more dangerous basilisk?

Share this post


Link to post
Share on other sites
quote:
Original post by Ketchaval
I know very little about AI, but here is an interesting (if possibly very elementary) question that might be worth considering.

Uncertainty and AI.

For discussion of the creation of an enemy AI system, which would use facts about its environment / what it perceives, combined with some logic to act in a reasonable manner.. in conflict situations.

Ie. It takes precautions relative to the likelihood of risk that it perceives. It (probably) won''t run around a corner with lots of Players round it, if there is a corner that it GUESSES may have less enemies hiding round it.

Of course the likelihood of things that the AI perceives is constantly updated, and it will calculate a new plan of action when the situation changes sufficiently.

Ie. It peeks round the (supposedly) empty corner, and sees a sleeping Basilik. It then considers what to do, does it run past the players or tiptoe past the potentially more dangerous basilisk?


What exactly is the question? Is sneaking a good thing to do, or how should I model a belief/perception system for the scenario above? If the latter there''s an excellent discussion going on about Fuzzy Logic right now, in fact.




Ferretman

ferretman@gameai.com
www.gameai.com

From the High Mountains of Colorado

Share this post


Link to post
Share on other sites