Would or could any game algorithm ever be applied to real world warfare scenario? I suppose i mean in the case of art imitates reality, the spybird is borne of game logic etc
Printable View
Would or could any game algorithm ever be applied to real world warfare scenario? I suppose i mean in the case of art imitates reality, the spybird is borne of game logic etc
What? And what is martial AI? Please explain.
ignore, i was on one mate,... :-|
The collision avoidance system in most modern aircraft uses an alpha-beta type of algorithm.Quote:
Originally Posted by rogster001
The idea is that you assume the other plane is actively TRYING to collide with you, and make changes to heading/speed in such a way that you avoid the collision even if the other party is malicious.
No, most game AI isn't considered 'true' AI, its just scripted behavior. The Terminator is still at least 3 years off (if I get my funding).
But i think it would be interesting to test pathfinding with a real tank say, given enough terrain data is it not possible to mobilise a group of units and give them a (fixed) target etc, how would that pan out in real scenarioQuote:
No, most game AI isn't considered 'true' AI, its just scripted behavior.
You would have to have a real time satellite uplink for it to be effective, terrain changes too quickly in real life to use stored maps, especially in desert and mountainous regions like Iraq/Afghanistan. There are a lot of tasks for which AI is just not ready. Airborne applications will come first because its much simpler terrain navigation, i.e. avoid all terrain.
That's funny, when I flipped open my old Artificial Intelligence book, alpha-beta was in there.Quote:
Originally Posted by abachler
Given the stuff you focus on, I guess you're one of those who thinks it isn't intelligence unless it's neural, and I suppose that's a valid if somehow arbitrary viewpoint...
A good deal of true AI involves much more than just NN's, but ultimately, the only example we have of true self aware intelligence is NN based, so it seems that is probably the best place to find AI. I'm open to the possibility that some other method will be found that will produce AI, but not really hopeful. My early vision system (early as in early in the processing, not as in early in the development process) uses no NN's at all, but ultimately it is intended to feed into a NN which will do the higher order processing.Quote:
Originally Posted by brewbuck